Giving the Federal Circuit a Run for Its Money: Challenging Patents in the PTAB

upon issuance. 4 The increased importance of knowledge production throughout the global economy, coupled with the popularity of the Patent Cooperation Treaty, which eases the burden of filing in multiple countries, 5 had expanded the number of patent applications and swamped the PTO's examination resources. 6 The resulting thickets of patents-and especially the increase in patent assertions by nonpracticing entities (NPEs)-were thought to be imposing a high tax on innovation. At the urging of the National Academies of Science and with substantial support from the American Intellectual Property Law Association (AIPLA), 7 Congress adopted these procedures to improve the likelihood that invalid patents would be quickly weeded out of the system.

To be sure, there are other ways to tackle the problem of low-quality patents. Increasing PTO resources would give examiners more time to search the literature and ensure that all the requirements for patentability are satisfied before patent rights are granted. Indeed, the AIA did some of this as well: it gave the PTO the power to set fees and retain them. 8 But there are limits to that approach. As Mark Lemley has convincingly argued, many patents have little commercial significance; examining them more thoroughly wastes social resources. 9 Furthermore, there are inevitable distortions in the process. Although examination is intended to protect the integrity of the public domain and the interest of the public in full disclosure of protected inventions, examiners deal only with those aspiring to acquire protection. While the PTO no longer follows an avowed "customer service paradigm" aimed at satisfying applicants, 10 systematic contact with only one constituency can lead to unconscious bias. Furthermore, as Melissa Wasserman, Michael Frakes, and Rob Merges have demonstrated, factors such as the pressure to earn maintenance fees, time constraints, informational asymme-tries, metrics for rating examiners' performance, and high examiner turnover can create other pathologies that lead to over-granting. 11 Better, to economists like Adam Jaffe and Josh Lerner, is to devote additional resources only to those patents that matter. 12 Of course, litigation does exactly that: patents that are asserted or challenged in a declaratory judgment action are, by definition, significant to someone. But litigation has its own problems. The process is extremely expensive. 13 While these heavy costs are borne only by the challenger, the industry as a whole enjoys the benefits of invalidation. 14 Thus, there is a collective action problem: every would-be challenger has an incentive to sit back and wait for a competitor to do the heavy lifting. 15 A patent can thus have an impact on an industry even when the industry questions its validity. In addition, not everyone who is affected by the exclusivity of a patent has standing to challenge it in court. To be sure, the patentee's competitors will often have direct enough interests, but as members of the same industry, they are unlikely to make arguments that could imperil their own patent holdings. Accordingly, there may be patents that impair the public interest that no one has both an incentive and the capacity to challenge. For example, it took decades for patents on isolated genes to be questioned. 17 The parties who finally did so (the ACLU and various doctors and patients) barely managed to convince a court that 11 See generally Michael D. Frakes & Melissa F. Wasserman, Does Agency Funding Affect Decisionmaking?: An Empirical Assessment of the PTO's Granting Patterns, 66 VAND. L. REV. (2013) (analyzing the impact of funding on USPTO decisionmaking); Robert P. Merges, As Many as Six Impossible Patents Before Breakfast: Property Rights for Business Concepts and Patent System Reform, 14 BERKELEY TECH. L.J. 577 (1999) (analyzing how business concept patents overload the patent system); Melissa F. Wasserman, The PTO's Asymmetric Incentives: Pressure to Expand Substantive Patent Law, 72 OHIO ST. L.J. 379 (2011).

12 ADAM B. JAFFE & JOSH LERNER, INNOVATION AND ITS DISCONTENTS: HOW OUR BROKEN PATENT SYSTEM IS ENDANGERING INNOVATION AND PROGRESS, AND WHAT TO DO ABOUT IT (2004). 13 See, e.g., AM. INTELLECTUAL PROP. LAW ASS'N, 2013 REPORT OF THE ECONOMIC SURVEY (2013), http://www.patentinsurance.com/custdocs/2013aipla%20survey.pdf (estimating that patent infringement litigation costs over $2 million through trial for cases worth less than $25 million and close to $6 million for the most valuable cases).

14 See Blonder-Tongue Labs., Inc. v. Univ. of Ill. Found., 402 U.S. 313, 335 (1971) (abolishing the doctrine of mutuality of estoppel).

15 See, e.g., Joseph Farrell & Robert P. Merges, Incentives to Challenge and Defend Patents: Why Litigation Won't Reliably Fix Patent Office Errors and Why Administrative Patent Review Might Help, 19 BERKELEY TECH. L.J. 943 (2004) (explaining the problems in the existing administrative system); John R. Thomas, Collusion and Collective Action in the Patent System: A Proposal for Patent Bounties, 2001 U. ILL. L. REV. 305, 307 ("Not only are opposition regimes ripe with collective action and free rider problems, they do not account for the possibility of collusion between the patentee and holder of patent-defeating prior art.").

16 See Michael J. Burstein, Rethinking Standing in Patent Challenges, 83 GEO. WASH. L. REV. 498, 500 (2015). 17 one of them was affected in a manner that met the standing requirements of federal law. 18 Even after a challenge is brought, there are obstacles to successfully attacking invalid claims. Deference to administrative procedures, along with a provision of patent law providing that "[a] patent shall be presumed valid," has led courts to assign to the challenger the burden of establishing invalidity by clear and convincing evidence. 19 While this burden seemingly makes little sense in situations where the PTO did not find-and thus never reviewedall the evidence relevant to validity, the Supreme Court has insisted on fidelity to that longstanding practice. 20 Prior to 1982, there was also a problem of persistent inconsistencies among the appellate courts hearing patent appeals, which led to forum shopping, uncertainty, and (arguably) a flight from patents to trade secrecy. The Federal Circuit was established in 1982 to remedy that concern; it hears virtually all patent appeals arising from the district courts, the PTO, and the International Trade Commission (ITC). 21 But as I documented in a series of articles, 22 channeling cases to a single appellate court introduced new complications. The court apparently saw in its creation an expression of congressional desire to strengthen protection. Thus, it reduced the standards of patentability; for example, it made it easier to find an invention nonobvious and extended the scope of patentable subject matter (to include, for instance, business methods). 23 This deepened the patent thicket, which led the Supreme Court to grant certiorari in increasing numbers of cases and to reverse or modify the Federal Circuit's resolution in most of them. 24 Not only has the Federal Circuit had difficulty persuading others that it has taken normatively appropriate positions, 25 it has also had problems implementing the Supreme Court's views and guiding district court practice. 26 Some scholars and judges have suggested revising the jurisdictional rules to give other courts some authority over adjudicating patent disputes. This would increase percolation and bring other points of view, including an antitrust perspective, to bear on questions of patentability, infringement, defenses, and remedies. But because practitioners prefer certainty and predictability, Congress is unlikely to make that change. 28 The new procedures instituted by the AIA could remedy many of these problems. Because these challenges are adjudicated within the PTO, the Patent Trial and Appeal Board (PTAB or Board), which hears these cases, applies the same preponderance of the evidence standard used to determine whether to grant a patent in the first place. 29 Since the cost of engaging in these administrative adjudications is lower than litigation, those who believe the patent invalid may be less likely to wait for someone else to come forward. Furthermore, the availability of a less expensive procedure to challenge patents could make the NPE business model less viable. Both IPRs and PGRs can be brought by any interested party. Accordingly, the public can use the procedures to vindicate interests that are not shared by the patent holder's rivals. And because a request for a PGR must be made within nine months of the issuance of the patent and all three procedures are subject to strict time limits, invalid patents can be cleared quickly and more timely feedback provided to the examining corps. 30 As important, these procedures have the potential to change the institutional culture of both the PTO and the federal courts. They require the PTO to focus on the concerns of nonpatentees (that is, to consider challengers to be among its "customers") and to evaluate patents later in their life, after more art pertinent to the question of validity has emerged. Furthermore, the PTAB is likely to be among the first to implement new Supreme Court pronouncements; it will probably also be the first to consider questions raised by the substantive changes made in the AIA. 31 Unlike the district courts, which must also grapple with these issues, the PTAB has expertise to cope with the technical aspects of its cases: it is composed entirely of patent and technology experts and there is an effort to staff each three-member panel with at least one judge who has specific familiarity with the field of the patented invention. 32 With these improvements over its predecessor, the Board of Patent Appeals and Interferences (BPAI), the PTAB could provide the Federal Circuit with a partner in the enterprise of interpreting patent law and implementing Supreme Court decisions. While it is true that the Federal Circuit, as the reviewing court, has the final say, 33 the PTAB breaks its isolation. At the same time, the thorough analysis furnished by the PTAB could give district courts better guidance on how to apply the law to emerging technologies.

The question is whether these salutary effects will, in fact, come to pass. At this point, there have been several statistical studies of the new procedures, including some by the PTO. This Article is intended to look beyond the numbers. To that end, I read all the final written decisions in CBM reviews handed down by the end of November 2014 as well as a set of final written decision in IPRs, starting with the first case filed, more than fifty cases in total. In addition, I read a selection of decisions to institute review, decisions on various motions, all the opinions issued by the Federal Circuit reviewing the PTAB, as well as the considerable outpouring (mostly in the form of blogs) published by those practicing before the PTAB. This is very 30 Cf. Bronwyn H. Hall & Dietmar Harhoff, Post-Grant Reviews in the U.S. Patent System-Design Choices and Expected Impact, 19 BERKELEY TECH. L.J. 989 (2004) (analyzing the efficiency effects of proposals for post-grant opposition procedures). 31 early in the life of these procedures; CBMs and IPRs began on September 16, 2012, and the PGR procedure can only be used to challenge patents issued on applications filed after the AIA went into full effect on March 16, 2013. Only a few PGR petitions have been publicly filed so far; the PTAB has instituted only two reviews, and as of this writing, no case has been decided. 34 While the other two procedures have been in use for two years, the early cases may be clearing a backlog of questionable patents and thus may not be representative of future practice. 35 Furthermore, the Federal Circuit has reviewed very few final decisions of the PTAB under the new procedures. Most of the cases have been decided without written opinion. 36 Nonetheless, it is valuable to consider what has happened so far. The cases I considered were sufficient to provide a sense of how the procedures operate and the opportunities they create to streamline-but also to gamethe system. My main objective is directed at institutional questions, to help the PTO and Congress as each considers changes to the system 37 and to gauge how well the PTAB could function to ameliorate the effect of Federal Circuit isolation and provide a basis for the court to consider new perspectives, write more persuasively, and provide better guidance. A second goal is 34 The instituted cases are American Simmental Ass'n v. Leachman Cattle of Colorado, LLC, No. PGR2015-00003 (P.T.A.B. June 19, 2015), and American Simmental Ass'n v. Leachman Cattle of Colorado, LLC, No. PGR2015-00005 (P.T.A.B June 19, 2015). As of June 29, 2015, eleven PGRs had been filed. USPTO, https://ptabtrials.uspto.gov/prweb/PRWebLDAP2/ HcI5xOSeX_yQRYZAnTXXCg%5B%5B*/!STANDARD?UserIdentifier=searchuser (last visited Oct. 27, 2015). 35 See Colleen Chien & Christian Helmers, Inter Partes Review and the Design of Post-Grant Patent Reviews, STAN. TECH. L. REV. (forthcoming) (manuscript at 14-16), http:// papers.ssrn.com/sol3/papers.cfm?abstract_id=2601562 (noting other selection effects). 36 See FED. CIR. R. 36 (permitting entry of judgment without opinion). The cases decided with written opinion include Versata Development Group, Inc. v. SAP America, Inc., 793 F.3d 1306 (Fed. Cir. 2015), Microsoft Corp. v. Proxyconn, Inc., 789 F.3d 1292 (Fed. Cir. 2015), and In re Cuozzo Speed Technologies, LLC, 793 F.3d 1268 (Fed. Cir. 2015). Only the decision in Proxyconn was reversed. 37 The STRONG Patents Act, S. 632, 114th Cong. (2015), would change many aspects of post-grant review, including the burden of proof on invalidity, the standard for claim construction, and standing to petition for review. The PATENT Act, S. 1137, 114th Cong. (2015), was initially directed at other issues, but Section 11 of the current Manager's Amendment includes a section on PTAB procedures that would change the standard for claim construction, broaden the grounds for amending claims, and alter other rules for conducting post-grant opposition; the Innovation Act, H.R. 9, 114th Cong. (2015), is similar. The changes contemplated by the PTO are outlined in PTO Director Michelle K. Lee's blog. See Michelle K. Lee, PTAB'S Quick-Fixes for AIA Rules Are to Be Implemented Immediately, U.S. PATENT & TRADEMARK OFF. (Mar. 27, 2015, 10:18 AM), http://www.uspto.gov/ blog/director/entry/ptab_s_quick_fixes_for. These include changing page limits, adjusting motion practice (including for amending claims), permitting the introduction of more evidence (including live testimony), allowing more discovery (especially on the question of who is the real party in interest), and adding new rules to deal with multiple challenges to the same patent. See also Amendments to the Rules of Practice for Trials Before the Patent Trial and Appeal Board, 80 Fed. Reg. 50720 (proposed Aug. 20, 2015) (to be codified at 37 C.F.R. pt. 42).

to draw greater academic attention to the potential these procedures have for changing the patent system and to provoke discussion-outside the emerging PTAB bar-on how they ought to operate. In particular, the statute layers inter partes review in a specialized agency under appellate review in a specialized court. I offer some thoughts on how authority over patent jurisprudence should be allocated between these two centers of expertise.

The Article proceeds as follows. Part I describes the three procedures. Part II evaluates use of the new system. Part III discusses the interaction between the PTAB and the Federal Circuit. Part IV looks at the problems these procedures raise for the parties, the adjudicators, and the sound administration of patent law.

The AIA provisions establishing the three new procedures, coupled with the PTO's regulations implementing them, create a complex procedural landscape. This Part describes the provisions and how they interact.

As noted above, Congress had many reasons to create new mechanisms for challenging patents within the PTO. First among them, however, were cost and timing. 38 The procedures Congress mandated and the regulations adopted by the PTO reflect that priority. 39 PGRs, IPRs, and CBMs are loosely referred to by attorneys as "trials," and are presided over by the PTAB using many procedural safeguards. But they are not the demonstrative spectacle of the courtroom. Instead, reviews are conducted in staggered fashion, according to a strict time table, and with page and discovery limitations, preferably on "paper" (in fact, electronic) filings. 40 In theory, the parties are represented by individuals who practice before the PTO, although other attorneys (such as litigators) can be admitted pro hac vice upon a showing of good cause. 41 The fees are relatively low but rise as more claims are challenged. 42 The three proceedings have much in common. Each begins with a petition to institute, 43 to which the patent owner may file a preliminary response. Once a petition is granted in whole or part, 44 (Fed. Cir. 2015). Whether they can be reviewed in conjunction with review of the final written decision is unclear. In In re Cuozzo Speed Technologies, LLC, the Federal Circuit denied review on an IPR challenge. 793 F.3d 1268, 1273-74 (Fed. Cir. 2015). However, in Versata, a CBM case, the Federal Circuit restyled the question as whether the PTAB how the proceedings will be conducted, usually through a telephone conference. The parties then have a limited period in which to conduct discovery. The patent owner can file a response to the petition, which can include a motion to amend any claims found to be unpatentable; the petitioner can then file a reply and an opposition to the motion to amend. The next few months provide the parties with opportunities to challenge evidence, file observations, and request and engage in oral argument. All in all, the procedure, once instituted, should conclude in one year (although it can be extended for six additional months). 45 Given the tight time limits on preinstitution activity, the entire process should, at most, take two years from start to finish.

In considering the challenge to a claim, the statute requires the Board, sitting in panels of three judges, to evaluate whether the petitioner sustained the burden of proving that it is more probable than not that the claim is invalid. 46 This is identical to the burden the PTO bears in the examination process. Similar to examination, the regulations require the Board to give unexpired claims their broadest reasonable interpretation. 47 Because challenges can be brought only against individual patents, but many challenges against a patent (or particular claims in it) can be mounted, the PTAB must also consider whether proceedings should be joined and considered together. 48 Once the PTAB hands down a final written decision, the losing party may appeal to the Federal Circuit. 49 By statute, a disappointed patent holder may also propose a reasonable number of amendments, but may not broaden the scope of the patent. 50 As interpreted by the PTO, each proposed amendment must be a direct response to the successful challenge. 51 After the time to appeal has expired or the appeal is terminated, the PTAB exceeded its authority in calling the invention in question a covered business method and then invalidating it. 793 F.3d at 1319-20. Over a strong partial dissent by Judge Hughes, id. at 1337-42, it then held it could decide whether the PTAB made the institution decision correctly. 45 issues a certificate in accordance with its findings (cancelling, confirming, or amending each challenged claim). 52 Settlements are possible, but any agreement made in connection with the settlement must be filed with the PTO. These can, however, be made confidential. 53 Along the way, the parties must plead with considerable precision. For example, the petition to institute must identify the real parties in interest and all administrative and judicial matters related to the proceedings, 54 and include a certification that the petitioner is not barred from bringing the challenge, the precise relief requested for each claim challenged, the basis for the challenge (including, according to the PTO, how the claim should be construed), and citation of the evidence that supports the challenge. 55 Petitions and responses must also conform to strict page limits. 56 Testimony and cross-examination (which are time-limited) are submitted in the form of exhibits (affidavits, transcripts of depositions); 57 supplemental information can be filed, but only within a narrow time frame or on a good-cause or interests-of-justice standard. 58 While the three new procedures have much in common, there are also important distinctions. Together, PGRs and IPRs are meant to provide a way to challenge patents throughout their term, but the proceedings are separately crafted to deal with right holders' growing reliance interests. CBMs are different again: they are meant to help clear particularly questionable patents from the system.

The PGR is, in essence, the main response to the National Academies' call to "weed out invalid patents or revise and narrow the claims." 59 Nominally, it is available only to challenge patents issued under the first-to-file priority rule, which was also instituted by the AIA. 60 Thus, it applies to patents issued on applications filed on or after March 16, 2013. 61 However, because the statute refers to patents that contain "a claim" with this effective filing date, 62 it is possible that PGRs can be brought to challenge claims in patents with earlier priority dates, so long as one claim that was not supported by the original disclosure was added after that date. 63 A petition for post-grant review may be filed by anyone who is not the owner of the patent, has not (or whose real party in interest has not) already filed a civil action challenging a claim in the patent, and is not estopped by a prior action. 64 It must be instituted within nine months of the grant of a patent or the broadening reissuance of a patent previously granted. 65 In keeping with the notion that patents may be of low quality for many reasons, including lack of novelty, obviousness, indefiniteness, inadequacy of a written description, or failure to enable, the petitioner is permitted to raise any of the grounds of invalidity denominated in the Patent Act as "defenses" to an infringement action. 66 And in keeping with this broad scope of challenge, the scope of discovery is potentially liberal: notwithstanding the explicit limits, 67 additional discovery is available upon a showing of good cause. 68 Discovery is, however, limited to "evidence . . . related to factual assertions advanced by either party." 69 The decision to institute a PGR requires the Board to determine whether the petition presents information that "if [it] is not rebutted, would demonstrate that it is more likely than not that at least [one] of the claims challenged . . . is unpatentable." 70 Importantly, the Board can also institute if the petition raises a "novel or unsettled legal question that is important to other patents or patent applications," 71 for this allows the PTAB to consider questions that will arise as the PTO implements the substantive provisions of the AIA. Once the determination to institute is made, other claims in the patent can also be challenged. While a petitioner cannot institute a PGR if it has already instituted a civil action challenging validity, 72 a civil action can be filed after the PGR is instituted. In that case, the civil action is stayed until the patent owner moves to lift the stay or files an infringement action. 73 Although adverse decisions can be appealed to the Federal Circuit, once a final written decision is handed down, the statute provides that the petitioner, its privies, and real party in interest are estopped from raising before the PTO, in a civil action, or in the ITC, an issue that was raised or reasonably could have been raised in the PGR. 74 Given the narrow window for bringing PGRs and the limitations on discovery and argument, however, it is not clear that Congress intended to extend estoppel to grounds that were not raised. Because the "reasonably could have raised" language appears to have been a drafting error, legislative efforts are being made to strike it. 75 Arguably, however, estoppel plays an important role in protecting patent owners from harassment. 76 As discussed more fully below, the impact of PGRs may depend heavily on the fate of this provision-or on how reasonably "reasonably" is interpreted and who are considered in privity with the challenger. A final written decision also estops the patent holder. It cannot take "action inconsistent with the adverse judgment." 77 This includes obtaining a claim that is the same as a claim that was cancelled or a proposed amendment that was rejected. 78

To some extent, IPRs pick up where PGRs leave off: for first-to-file claims, they can be instituted as soon as the date to file a PGR has passed. 79 IPRs are, however, also available to challenge patents issued under the firstto-invent system. Since these are not subject to PGR review, a challenge can be filed as soon as the patent is granted. 80 As with PGRs, anyone who is not the patent owner and is not already estopped may file one, 81 but as before, petitioners face a choice: no one who has filed a civil action challenging the validity of a claim in the patent can then file an IPR. 82 As with PGRs, there is 73 Id. § 325(a)(2). In addition, PGRs can be instituted after validity is challenged in a counterclaim. Id a provision to stay court proceedings the petitioner files after filing an IPR. In addition, if the petitioner, his privy, or real party in interest has already been served with a complaint alleging infringement, the IPR must be sought within a year of service. 84 IPRs are different from PGRs in one critical respect: claims can be challenged only on the grounds of novelty and nonobviousness, and then only if the prior art constitutes a patent or a printed publication. 85 The standard for finding the claim unpatentable is the same as for PGRs, but the limit on the grounds for invalidation means that the patent holder and its post-issuance investment are exposed to invalidation on this standard only with respect to issues that could not easily be determined soon after issuance (and only with the use of materials that can be easily put into evidence). There are other subtleties. Institution depends only on whether there is "a reasonable likelihood that the petitioner would prevail with respect to at least [one] of the claims challenged in the petition," taking into account the patent owner's preliminary response (if one is filed). 86 There is no provision for additional discovery on a good-cause standard, although it can be provided in the interest of justice. 87 While estoppel for the patent holder is the same as for PGRs, the effect of preclusion is different for the petitioner. As the only grounds that can be raised in the IPR are novelty and nonobviousness based on a narrow range of prior art, there are many grounds on which the identical claims can be challenged in court or in the ITC.

CBMs were added to the AIA at the tail end of legislative deliberations, in the wake of several Supreme Court cases questioning the Federal Circuit's interpretation of the scope of patentable subject matter and as NPE actions emerged as a problem. 89 As its formal name implies, the program is 83 intended to sunset. 90 Presumably, by September 16, 2020, anyone affected by these questionable patents will have had time to challenge them. As with IPRs, claims in both first-to-file and first-to-invent patents can be challenged; as with PGRs, the challenge can be on any of the grounds that are denominated defenses to an infringement action. 91 CBMs can be brought any time in the life of the patent, but the patent holder is arguably protected from harassment in that a petition may be instituted only if the petitioner, its real party in interest, or privy has been sued for, or charged with, infringement and is not estopped by a prior action. 92 Furthermore, only the holders of "covered business method patent[s]" 93 are vulnerable. Presumably the patents whose validity is most questionable are the ones most often asserted by NPEs. These are defined as patents that claim "a method or corresponding apparatus for performing data processing or other operations used in the practice, administration, or management of a financial product or service, except that the term does not include patents for technological inventions." 94 The PTO determines whether a patent is a technological innovation by considering, on a case-by-case basis, "whether the claimed subject matter as a whole recites a technological feature that is novel and unobvious over the prior art; and solves a technical problem using a technical solution." 95 These limitations may appear to narrow the set of challenges that can be considered. Nevertheless, the PTAB has shown itself willing to consider a fairly wide variety of cases. 96 Furthermore, once it finds one claim to be within the definition, the PTAB can consider challenges to other claims in the patent. 97 On the whole, the conduct of proceedings for CBMs mirrors that for IPRs. 98 However, because CBMs can mainly be brought in cases where litigation is pending or threatened, the procedures include a special provision on stays of trial court actions. According to the statute, trial courts must determine (A) whether a stay, or the denial thereof, will simplify the issues in question and streamline the trial; (B) whether discovery is complete and whether a trial date has been set; (C) whether a stay, or the denial thereof, would unduly prejudice the nonmoving party or present a clear tactical advantage for the moving party; and (D) whether a stay, or the denial thereof, will reduce the burden of litigation on the parties and on the court. 99 There is a right to an immediate interlocutory appeal from the trial court's decision and this may be de novo to ensure consistent application. 100 In its first decision involving this provision, VirtualAgility Inc. v. Salesforce.com, Inc., the Federal Circuit reversed the denial of a stay, criticized the trial court for reviewing the PTAB's determination that it was more likely than not that one claim was not patentable, and stressed the need to consider whether the stay would simplify the issues and streamline the trial. 101 Unlike PGRs and IPRs, the estoppel provision for CBMs is narrow. Upon issuance of a final written decision, the petitioner or its real party in interest may not later assert invalidity on any ground raised-not any ground that reasonably could have been raised-during the proceeding. 102 The estoppel applies to district court and ITC proceedings, but the statute makes no mention of subsequent proceedings in the PTO. Nor does it mention the petitioner's privies.

In a sense, the numbers speak volumes. Despite the fears that the estoppel provision or collective-action problems would discourage use of these procedures, IPRs and CBMs have proven extremely popular. The number of petitions is impressive. Near the two-year mark, the statistics were as follows: Even more revealing than the number of cases have been the dispositions. Initially, around 85% of the claims challenged in an instituted IPR were cancelled; that figure was 100% for CBMs. 104 While these rates appear to be dropping for IPRs, 105 a study by Brian Love and Shawn Ambwani shows that over the first two years, petitioners were extremely successful:

• The PTAB grants-or "institutes"-IPR petitions for at least one challenged claim 84 percent of the time; • Among instituted IPRs, all challenged claims are instituted 74 percent of the time; • Among IPRs that reach a final decision on the merits, all instituted claims are invalidated or disclaimed more than 77 percent of the time. 106 103 PATENT TRIAL & APPEAL BD., AIA PROGRESS STATISTICS (AS OF 9/11/14) 1 (2014), http://www.uspto.gov/ip/boards/bpai/stats/aia_statistics_091114.pdf. The PTO did not provide numbers for the two-year mark (September 16, 2014). However, the fiscal year runs from October 1 through September 30 and numbers were published for fiscal year 2014. They show that by September 30, 2014, seventy more IPRs and eighteen more CBMs had been filed. 104 While these statistics speak loudly about the public's eagerness and ability to use these procedures to "weed out" bad patents, it is more difficult to interpret what the numbers mean from a normative standpoint. To some, they suggest that the Board is out of control. As Randall Rader, once chief judge of the Federal Circuit, put it, the judges of the PTAB are "acting as death squads, killing property rights." 109 the PTAB is "where patent claims go to die." 110 More temperately, after comparing cancellation rates in IPRs to invalidation rates in court and considering the number of claims cancelled in IPRs that had previously survived ex parte reexamination, 111 Gregory Dolin concluded that it is "too easy to invalidate a duly issued patent" in an IPR. He called the CBM statistics "even more staggering." 112 But the numbers can be understood in a very different way. Before the PTAB can issue a final written decision cancelling claims, it must decide whether to institute a proceeding. The standard for determining whether to institute, although slightly different for the three procedures, is essentially whether it is more probable than not that at least one challenged claim is unpatentable. Since the preponderance of the evidence standard for determining whether a claim should be cancelled is also, essentially, whether it is more probable than not invalid, the high invalidation rate is basically a reflection of the PTAB's ability to forecast correctly how it will decide on at least one claim. Because the panel that decides whether to institute also decides the case on the merits, a strong correlation is to be expected. 113 This is especially so because the institution decisions are far from pro forma: they are often as long as the merits decision, cover the same issues (claim construction is often central), and are as thoughtful and probing of the arguments as the decisions on the merits. 114 Admittedly, once the PTAB decides one claim may be invalid, it can entertain challenges to other claims as well. But it need not hear every claim the petitioner seeks to cancel. Partial institutions are possible, and in practice, the PTAB screens out claims that appear to be valid at the institution stage. That is, in deciding whether to institute, the 113 The PTO is considering a pilot program in which institution decisions would be made by only one judge; if that judge decides to institute, two new judges would be added to decide the case. See Lee, supra note 37. Such a procedure would benefit from continuity and also give a greater appearance of impartiality; it will be interesting to see if it leads to a lower rate of cancellation. 114 To take two examples, arbitrarily chosen, the institution decision in SAP America, Inc. v. Versata Development Group, Inc. was forty-four pages long. No. CBM2012-00001 (P.T.A.B. Jan. 9, 2013) (institution decision). In the same case, the final written decision was thirty-five pages. PTAB often considers every claim and every ground to determine whether each claim is more likely than not unpatentable on each alleged ground. 115 Along the same lines, CBMs cannot be instituted if the claim as a whole is drawn to a "technological feature that is novel and unobvious over the prior art; and solves a technical problem using a technical solution." 116 Because many business-method claims involve fairly abstract ideas tied only to a general-purpose computer, and, per the Supreme Court, can only be considered patentable subject matter if they contain an "inventive concept," 117 the "technical solution" requirement, although not the same as the "inventive concept" test, will filter out some subject-matter challenges that would be losers on the merits. 118 It is also important to keep in mind that the claims challenged are not always independent of one another. Depending on their relationship and the grounds considered, cancellation of some claims may be highly likely to lead to the cancellation of others. 119 For example, a newly revealed piece of prior art that anticipates a dependent claim is likely to anticipate the independent claim on which it depends and other dependent claims as well. 120 Once a claim or a facet of prior art is interpreted in a particular way, the interpretation can have an adverse impact on an entire family of patents. 121 Additionally, multiple parties can file challenges to the same claims. 122 This practice inflates the cancellation rate. It may also be abusive. That issue is discussed in Part IV. 123 It is also misleading to compare the outcome of litigation with the results of PTAB review. The preponderance standard for determining invalidity in PTAB challenges, which is statutorily mandated, can easily lead to the cancellation of claims that Rader, as a Federal Circuit judge applying a clear and convincing standard, would have upheld and which may, indeed, have been previously upheld by district courts. 124 As important is the difference in the approach to claim interpretation. The PTAB gives unexpired patent claims their broadest reasonable interpretation; in court, they are given their ordinary customary meaning, which is generally narrower. 125 The broader the claim, the more likely the challenger can find prior art to invalidate it on novelty or obviousness grounds, 126 the more abstract it is likely to read, 127 120 Cf. In re Cuozzo Speed Techs., LLC, 793 F. and the less likely it is to be fully supported by the written description, adequately enabled, and distinctly claimed. 128 Most important from a normative perspective is that many claims deserve to die. Dolin makes much of the fact that 15% of the claims cancelled in an IPR had survived a challenge in ex parte reexamination. Because he fails to say which IPRs these are, it is impossible to know for sure what happened, but it is worth noting that ex parte reexaminations, like initial examination, are ex parte-that is, without an opportunity for a challenger to respond to the patent holder's arguments. Inter partes review corrects the asymmetry between the patent holder's and the patent office's knowledge of the field and the invention; the greater number of invalidations in CBMs and IPRs can be read as affirming the suspicion that ex parte procedures are not fully effective at flushing out bad claims.

More to the point, Dolin's reexaminations may have preceded the Supreme Court's articulation of the concern that Federal Circuit jurisprudence was producing patents that "impede rather than 'promote the Progress of Science and useful Arts.' " 130 In the years following this statement by Justice Breyer, the Supreme Court reworked the law on many key issues of patentability. In KSR International Co. v. Teleflex Inc., 131 the question was nonobviousness, which requires the invention to be beyond the grasp of a person with ordinary skill in the art. 132 The Federal Circuit had imposed a requirement that the challenger show a "teaching, suggestion, or motivation" to combine prior art references. 133 But the Supreme Court rejected the "errors" of the Federal Circuit and its "rigid" rule. It stressed that "a person of ordinary skill is also a person of ordinary creativity, not an automaton," that "[t]he obviousness analysis cannot be confined by a formalistic conception of the words teaching, suggestion, and motivation, or by overemphasis 5004949 (Fed. Cir. Aug. 24, 2015); Intellectual Ventures Mgmt., LLC v. Xilinx, Inc., No. IPR2012-00018 (P.T.A.B. Feb. 10, 2014) (noting that the determination of obviousness turned on the definition of "inside"). 127 See, e.g., Fidelity Nat'l Info. Servs., Inc. v. Cashedge, Inc., No. CBM2013-00028 (P.T.A.B. Dec. 23, 2013) (institution decision deciding whether the claims are drawn to patentable subject matter premised on a reading of claim that reads out its nonabstract elements). 128 See, e.g., SAP Am., Inc. v. Lakshmi Arunachalam, No. CBM2013-00013, at (P.T.A.B. Sept. 18, 2014) (construing the term "routed transactional data structure" in a manner that leads to a determination of indefiniteness). The breadth of the claim may also determine whether an institution decision is made. on the importance of published articles and the explicit content of issued patents," 135 that "it often may be the case that market demand, rather than scientific literature, will drive design trends," 136 and that persons of ordinary skill in the art will repurpose solutions to one problem in order to solve a different problem.

By the same token, in a series of four cases, beginning with Bilski v. Kappos and ending with Alice Corp. v. CLS Bank, 137 the Court rejected the Federal Circuit's view that any invention that produces a useful, concrete, and tangible result is drawn to patentable subject matter, as well as its rule that any invention that constitutes a machine or effectuates a physical transformation is statutory subject matter. Instead, the Court emphasized that laws of nature, phenomena of nature, and abstract ideas are not patentable (although abstract claims can be "saved" if they include an "inventive concept"). 138 And in Nautilus, Inc. v. Biosig Instruments, Inc., the Supreme Court rejected the Federal Circuit's rule that a claim is valid so long as it is not "insolubly ambiguous"; under the Court's test, a "patent's claims, [when] viewed in light of the specification and prosecution history, [must] inform those skilled in the art about the scope of the invention with reasonable certainty." 139 Given the fundamental nature of these changes to patent law, it is not surprising that the PTAB would institute review and then invalidate many claims issued, reexamined, or litigated before these cases were handed down. Indeed, the technological distribution of the cases subject to review is suggestive: Ct. 2120, 2129 (2014). There are questions as to whether the standard for indefiniteness used in a court challenge is the same as the one that applies pre-issuance and, if so, which standard applies to post-grant opposition procedures. See Erika H. Arner et al., Section 112(b) and the PTAB: Is the Legal Standard for Indefiniteness Itself Indefinite?, 90 PAT. TRADEMARK & COPYRIGHT J. (BNA) 2335 (2015). As noted in connection with the standard for claim construction, efficiency will be vastly improved if the same standard were applied throughout. See infra text accompanying notes 215-14. Given the notice function of definiteness, it is particularly difficult to see why different standards should be used. The vast majority of the cases are in the electrical/computer sector, where the rejection of the "useful, concrete, and tangible result" and the "machine or transformation" tests are most significant, and in mechanicals, the field of the invention in KSR (and arguably, Nautilus). The decisions themselves rely heavily on the recent Supreme Court cases. It is also noteworthy that in a 2013 study, Shawn Miller showed that as with the PTAB, courts are more likely to invalidate patents in the areas directly affected by KSR and Bilski. 141 Furthermore, the cancellation rate is comparable to the findings of a 2014 German study, where researchers estimated that around 75% of German patents are at least partially invalid. 142 Selection effects must also be considered. Unauthorized users have three shots at retaining freedom to operate: it is possible that the patent holder will fail to sue; that if the patent holder does sue, the patent will be invalidated; or that in the suit, the user's activity will be found noninfringing. At the PTAB, there is only one possibility: invalidation. 143 Accordingly, challengers are likely to use these proceedings only when the case for invalidity is extremely strong. Table 2, which shows a fairly high number of cases in which the patent holder requested an adverse judgment, 144 supports the notion that in the mix of losses are many claims that even the patent holder understood to be defective. Finally, it is not insignificant that, as with the cancellation rate, the percentage of cases in which the PTAB has decided to institute proceedings appears to be declining. 145 This suggests that petition-ers may have saved claims that were clearly invalid for the onset of these new procedures.

If it continues, the eagerness with which the new procedures have been embraced suggests that those who proposed these mechanisms were right: there are many advantages to channeling patent validity challenges back to the PTO. However, the ultimate success of the system depends heavily on the Federal Circuit's approach to reviewing PTAB decisions. This Part elaborates on the benefits and on their viability once the Federal Circuit begins to see more of these cases. But even if all potential benefits are realized, the procedures have significant costs. With three types of review, the system is complex and use of the procedures is not exactly cheap. Their two-year potential duration is less than the length of district court litigation, but there is no right to an expedited appeal. Accordingly, the total period of uncertainty could be almost as long as it is for litigation. These matters are discussed in Part IV.

The numbers reviewed in the previous Part demonstrate how effectively the PTAB is clearing questionable claims from the system. The potential impact is evident: these procedures can promote freedom to operate, facilitate settlement, lower the incidence of litigation, and curb NPE practice. With a workable procedure for issuing stays, they could also streamline litigation and make what goes on in the courtroom more efficient. 146 Moreover, Michael Kasdan, Trends from 2 Years of AIA Post-Grant Proceedings, LAW360 (Sept. 29, 2014), http://www.law360.com/articles/581512/trends-from-2-years-of-aia-post-grant-proceed-ings; cf. Michael Fuller, The PTAB May Be Taking a More Balanced Approach in Biotech and Pharmaceutical IPRs, KNOBBE MARTENS (Mar. 23, 2015) http://knobbe.com/news/2015/ 03/ptab-may-be-taking-more-balanced-approach-biotech-and-pharmaceutical-iprs-bloomberg-bna (noting a 75% institution rate for fiscal year 2015). Still, it was high during the period of examination (82% overall for IPRs as of September 3, 2014, and 80% for CBMs as of that date). See PTO Trials, SIDLEY AUSTIN LLP (Sept. 3, 2014), http://ptotri-als.sidley.com/. It is important to note that the institution rate is hard to calculate because only some claims in a given petition may be subjected to scrutiny and there are sometimes multiple petitions directed at the same claims. Note also that PTAB's categorization of the cases can be unclear. See supra text accompanying note 108; see also Zetec, Inc. v. Westinghouse Elec. Co., No. IPR 2014-00384 (P.T.A.B. July 23, 2014) (institution decision refusing to institute because arguments were undeveloped). 146 While the hard line on stays taken in VirtualAgility, a CBM case, is not directly applicable to IPRs and PGRs, the Federal Circuit stressed the possibility that PTAB consideration would simplify trial. VirtualAgility Inc. v. Salesforce.com, Inc., 759 F.3d 1307, 1314 (Fed. Cir. 2014). As the text below makes clear, however, presently, the multiplicity of these procedures can make it difficult for a trial court to determine the exact contours of a case or know when to dissolve stays and proceed to trial. See infra text accompanying notes 287-288, 306; see also, e.g., Ultratec, Inc. v. Sorenson Commc'ns, Inc., No. 13-cv-346-bbc, 2015 WL 2248437 (W.D. Wis. May 13, 2015) (initially refusing to stay litigation for an IPR, as Joseph Casino and Michael Kasdan have noted, "[t]he PTAB runs a tight ship." 147 Discovery has been kept to a minimum and the PTAB has managed to keep within the time limits mandated. Indeed, Love and Ambwani found that the average pendency of IPR petitions within the PTAB is fifteen months. 148 Given the three years it can take to try a patent case in popular jurisdictions, speed can be an important benefit to using these proceedings. 149 As proponents hoped, the Board has taken its duty to protect the public interest seriously. Thus, it has adopted a broad understanding of what constitutes a "financial product" within the meaning of the transitional business method program 150 and what it means to be the object of an assertion in a CBM. 151 At the same time, it has interpreted the limits on filing IPRs narrowly, 152 taken a dim view of theories of assignor estoppel, 153 and adopted a skeptical approach to arguments that a claim is within the "technological innovation" exception to CBMs. 154 on that practice and decrease the tax it is said to impose on innovation. Furthermore, the PTAB has allowed interested parties who may lack standing in court-suppliers are an example-to bring IPRs. 156 Interestingly, although the settlement rate has increased, the PTAB does not always terminate proceedings after the parties settle. Instead, it sometimes retains authority to issue a final written decision cancelling challenged claims. 157 Indeed, for strategic reasons (such as to avoid other invalidations), some patent holders have simply requested adverse judgment once review was instituted. 158 Most important, because the PTAB applies patentability criteria using the preponderance of the evidence standard for invalidation and the broadest reasonable interpretation of the claims, its decisions do what the examiners would have done had they had the PTAB's extra time, resources, and expertise, and had they understood the law as enunciated in recent Supreme Court interventions. Application of the law nunc pro tunc goes a longer way than litigation can towards safeguarding the public's interest in the free availability of technology that should not be privately controlled. 159 npes-fleeing-ptab-patent-challenge-proceedings (depicting that 75% of all 2013 IPR settlements involved NPEs); see also, e.g., Bank of Am., N However, the statutory requirement that the PTAB apply the preponderance of the evidence rule suggests that providing a substitute, cheaper, route to litigation was not Congress's sole concern.

In the future, fear of the "death squads" (or, an efficient PTAB) may discourage patent prosecutors from drafting abstract or overly broad claims. 160 My reading of the cases suggests many more subtle ways in which the invigoration of patent review within the PTO could contribute meaningfully to patent jurisprudence and the administration of patent law. Like the BPAI it replaced, the PTAB is not bound by the Manual of Patent Examining Procedure or PTO Guidelines and thus it brings an independent analysis to bear on patentability questions. 161 As noted earlier, the PTAB is staffed with experienced patent lawyers and each panel is meant to include someone with close knowledge of the field of the invention. 162 The institution and final decisions reflect this expertise. The Board's opinions are extremely well written and closely reasoned; their fluidity suggests deep immersion in both the technical facts of the cases and the law to be applied. In short, the Board is well-structured to add an important new voice to the development of patent jurisprudence.

Most obviously, the PTAB's expertise can be exploited in a "preview" capacity. Many of the cases coming before the Board raise questions about implementing recent Supreme Court decisions. Consider, for example, definiteness. When the Supreme Court rejected the Federal Circuit's "insolubly ambiguous" standard, it did not spell out exactly what it meant for a claim to be "precise enough to afford clear notice of what is claimed" while "tak[ing] into account the inherent limitations of language." 163 The Federal Circuit's attempt to anticipate Nautilus was a failure 164 and in the immediate aftermath, it considered the issue only twice, rather briefly. 165 Notably, oral argument in the remand of Nautilus suggested the court was having a very difficult time coming up with a workable approach; 166 ultimately, it basically reinstated its earlier opinion. 167 In contrast, the PTAB started citing the decision three weeks after it issued. 168 Repeated exposure to definiteness problems in CBMs (and eventually in PGRs) is leading to the development of nuanced and context-specific standards. For example, in SAP America, Inc. v. Lakshmi Arunachalam, decided three months after Nautilus, the claims covered a method and apparatus for providing real-time, two-way transactions on the web. 169 To determine whether the term "routed transactional data structure" was indefinite, the Board engaged in a technical analysis of the specification's teachings concerning how objects were transmitted over the network and focused considerable attention to the question of what a person of ordinary skill would understand, and in particular, whether a proprietary protocol referenced in the patent disclosed sufficient information to inform such a person of the scope of the claim. 170 While there is some uncertainty as to whether the PTAB should apply the same standard on this issue as courts, PTAB cases can still furnish something of a blueprint for deciding definiteness issues, particularly in the Internet arena, where sharper claims could help alleviate the trolling problem. 171 Similarly, the law laid out in the four Supreme Court cases on patentable subject matter is hard to understand, especially with regard to the meaning of "conventional steps" and "inventive concept." Yet the PTAB rendered decisions based on Alice within a few months of the decision. 172 In these cases, the Board set out an analytical framework and did so in specific fields, such as the interactive web applications mentioned above. 173 Several decisions offer a methodology for thinking about what counts as a general purpose computer, 174 what constitutes more than a "conventional step" in the use of a general-purpose computer, and a way to consider abstractness in the context of computer implementation, including (as some commentators have suggested) by reference to what humans could accomplish on their own. 175 The PTAB tends to begin with a careful construction of the challenged claims; in contrast, the Federal Circuit has shown a willingness to dis- pense with that step, which, in some cases, may make patents more vulnerable in court than in the PTAB. 176 In some instances, the PTAB considers interstitial issues of fact or law that could have arisen in earlier cases, but did not. Examples include whether an SEC filing constitutes prior art. 177 Some arise out of the new procedures themselves, such as determining the relationships that give rise to privity 178 or whether challenges based on subject matter limitations can be heard by the PTAB. 179 Others arise from the procedures, but could have relevance in the courtroom as well. An example is determining whether a claim is within the technical solution exception to CBMs, which (as suggested earlier) may sometimes help to decide certain subject matter cases in the wake of Alice. 180 The PTAB's preview capacity should also be extremely helpful as examiners apply the new rules on priority, prior art, and the grace period. For example, the AIA substantially changed the language of the novelty provision of the Patent Act, adding a new phrase-"or otherwise available to the public"-to describe art that invalidates a patent. 181 The addition raises the question whether prior practice, which made some confidential uses and secret sales patent-barring, has been legislatively overruled. 182 The PTO Guidelines suggests this is so, 183 but the PTO received numerous comments arguing that Congress did not mean to change the standard. 184 The issue is quite likely to arise first in a PGR, where the grant of a patent despite evidence of a secret sale could be challenged on the basis that the petition raises a "novel or unsettled legal question that is important to other patents or patent applications." 185 As noted, the PTAB is not bound by examination guidelines and it would be of considerable value to have its independent judgment on these issues, even though the Federal Circuit will ultimately review the work.

Even regarding issues that arise repeatedly, there are many benefits to be derived from the PTAB's closer contact with the field of the invention. Nonobviousness provides an example. As we saw, KSR stressed the creative powers of the person of ordinary skill in the art to take what is known and adapt or extend it to meet other needs. The issue of what this means in terms of what the prior art teaches and what the person of ordinary skill can do with it comes up in the Internet and software context in CBMs and on a particularly regular basis in IPRs, where nonobviousness is the main basis for challenge. Even the earliest cases show how effectively PTAB panels grapple with questions of nonobviousness law and fact. For example, in Liberty Mutual Insurance Co. v. Progressive Casualty Insurance Co., 186 which involved Internet transactions, one question was whether the prior art on which the challenge relied must be enabling. Citing one Federal Circuit case saying it need not be enabling and a second case saying prior art is presumed to be enabled, the PTAB cut through to the real question: whether the art in fact taught an artisan how to practice the patented claim. 187 In another part of the opinion, it determined what people could do on the Internet, and what the design community was already able to accomplish at the time of the invention. 188 In Garmin International, Inc. v. Cuozzo Speed Technologies LLC, 189 a case about GPS devices, the PTAB demonstrated its approach to the questions of how a person of ordinary skill who was "not an automaton" would combine references when there is no explicit statement in the prior art suggesting combination and how such a person would deal with disclosures in patents that did not claim an invention dealing with the same problem facing the inventor. 190 The Board also considered (and rejected) a series of argu-ments on when an art discourages ("teaches away" from) the claimed invention, including arguments about the absence of affirmative statements on feasibility and cost. 191 In Microsoft Corp. v. Proxyconn, Inc., it discussed what constitutes "analogous art" to a system for data access in a packet switched network. 192 In addition, the PTAB has had to reconcile the parts of Federal Circuit law on nonobviousness that the Supreme Court has not directly reversed with the Court's current approach to the issue. Thus, in Kyocera Corp. v. Softview LLC, 193 a case on Internet displays (such as HTML), the question was how to use the objective criterion of nonobviousness, which the Federal Circuit has mandated be considered in every nonobviousness case. 194 Reliance on these so-called "secondary considerations" (such as commercial success) to support a finding of patentability has been highly controversial. 195 It is also somewhat difficult to square with KSR's rejection of rigid rules and its notion that commercial demand can undermine patentability by providing sufficient motivation to advance the art. In Kyocera, the PTAB lessened the tension by assigning to the patent holder a heavy burden of showing a nexus between commercial success and the patented feature. Further, it elaborated on what must be shown: "[P]roof that the sales were a direct result of the unique characteristics of the claimed invention." 196 In many of these cases, the basic rules derive from Federal Circuit case law. However, the contextual use of the case law, the systematic nature of the inquiry, and the showing required of the party bearing the burden of proof will provide useful guidance to examiners. 197 As Federal Circuit Judge Raymond Chen put it: I foresee an opportunity for these board decisions to assist in a forward-looking way to improve patent quality. The patent board will be developing a large body of data that can perhaps yield patterns or insights for what went right, or what went wrong during the initial examination process. And the agency can use those lessons learned to improve patent examination. In the next few years, the patent board will have created a rich source of in-house generated material the agency can potentially use to further improve the quality of patent examination. 198 PTAB decisions could be similarly helpful to district court judges. Indeed, the availability of detailed instruction from the PTAB could change the nature of Federal Circuit jurisprudence. Thus, one reason why the Federal Circuit tends to create rules that the Supreme Court regards as overly "rigid" may be that it is drawing bright lines that nontechnical trial judges can apply with ease, thereby effectuating its perceived mandate to ensure the uniform application of patent law. 199 The Supreme Court dislikes this approach because-like all rules-it will not lead to the right result in every case and perhaps also because, as in KSR and Mayo, errors tend to favor the patent holder. Furthermore, although the Supreme Court has never done so directly, one can question the Federal Circuit's categorization of what constitutes a question of law and what constitutes a question of fact: Why, for example, is enablement a question of law and written description a question of fact when both are meant to guarantee the adequacy of disclosure? 200 The answer may again be uniformity: by classifying heavily technical issues as legal questions, the court can review the district court's resolution de novo. Now that the PTAB is in the picture, the rigid rules and distorted classifications of questions as law or fact may not be needed, for the PTAB's decisions spell out more clearly how factual issues should be determined and lay out detailed analyses for particularly difficult technologies. If parties begin to cite what Judge Chen calls the PTAB's "patterns or insights," 201 bright line rules and de novo review may become less necessary. As a result, the number gesting that the PTAB will provide guidance on implementing Federal Circuit cases and giving the example of PTAB decisions applying the then-recent Federal Circuit opinion in Alice). 198 Tamlin Bason, Judge Chen: Board Could Develop Rich Data Source that Will Help Improve Patent Quality, 88 PAT. TRADEMARK & COPYRIGHT J. (BNA) 1676, 1676 (2014) (quoting Federal Circuit Judge Raymond T. Chen). Given these views, it is perhaps not surprising that the approach that Judge Chen has taken to abstractness is very close to that of the one the PTAB uses in its institution decisions. of reversals could be reduced. And now that the Supreme Court has insisted that the Federal Circuit carefully distinguish between ultimate questions of law and subsidiary questions of fact, such reclassifications would also simplify trials and appeals. 202 The cases suggest that the PTAB can also be helpful in reducing the effects of isolation. As many observers have noted, the Federal Circuit sometimes backslides into old practices. 203 The PTAB could play a role in moderating that tendency. Once again, nonobviousness furnishes an example. As we saw, the Supreme Court has emphasized that people of ordinary skill are creative and not automatons. But in K/S HIMPP v. Hear-Wear Technologies LLC, 204 the Federal Circuit limited the reach of the decision to the specific context of KSR, where the question was whether a person with ordinary skill could combine references without a specific suggestion to do so. Thus, it refused to consider an ordinary artisan creative enough to make use of an element not mentioned in the cited prior art (a plurality of prongs that provide a detachable mechanical and electrical connection), but which was publicly available at the time of the invention. The court held that ordinary knowledge in the field can be used only on "peripheral issues." 205 In his dissent, Judge Dyk argued that the inability of examiners to rely "on their expert knowledge and common sense about what is well known in the art" is inconsistent with the Supreme Court's concern about rigid rules. 206 Had the majority considered what is happening in the PTAB, it might have found the dissent more persuasive. The cases there show that the use of background knowledge can come up in many contexts, making it rather unlikely that the KSR Court intended to spawn satellite litigation on whether a particular circumstance was peripheral or central, or to limit consideration of an artisan's creativity and other motivations to only certain issues. 207 202 See Teva Pharm. USA, Inc. v. Sandoz, Inc., 135 S. Ct A less intuitive example is claim construction, where a strong argument can be made that the Federal Circuit should allow itself to be influenced by the PTAB. That is, it should abandon its reliance on the meaning the claim "would have to a person of ordinary skill in the art in question at the time of the invention" 208 and instead adopt the practice of giving unexpired claims their broadest reasonable interpretation (BRI). To be sure, this use of the BRI is a PTO innovation. Unlike the standard for proving unpatentability, it is not statutorily imposed. It is by far the most controversial aspect of the administration of the three new procedures; 209 indeed, bills pending before Congress would reject its use. 210 As we saw, using a broad interpretation leads to the invalidation of claims that would have been upheld in court (even on a preponderance of the evidence standard). 211 This comes across as extremely unfair. Of course, if it were as easy to amend claims in CBMs, IPRs, and PGRs as it is to amend them during examination, the system would make some sense: the patent holder would retain exclusivity over advances that it had the right to protect. However in Idle Free Systems, Inc. v. Bergstrom, Inc., the PTAB took a stringent approach to motions to amend: it put the burden of establishing patentability on the patent holder (in examination, the PTO would bear the burden of showing unpatentability). 212 The patent holder must also set forth the knowledge of one skilled in the art and the closest prior art. Further, the amendment must respond directly to the reason the Board invalidated the original claim, and it must be narrower than the claim it replaces. 213 The Idle Free approach has led to very few successful amendments: as of the date of the Love and Ambwani study, the only amendment the PTAB granted in an IPR was one that was unopposed and filed by the United States. 214 limited to the consideration of patents and publications Surprisingly however, in In re Cuozzo Speed Technologies, LLC, the Federal Circuit's first encounter with these procedures, it approved use of the BRI. 215 However, the use of two standards, one for court and one for the PTAB, creates considerable inefficiency. If both tribunals used the same rule, the PTAB could dispense with claim construction in any case in which the district court had already interpreted the claims. Conversely, construction by the technologically sophisticated experts at the PTAB would provide a wealth of information to district courts later considering the same or related claims. Beyond efficiencies, adopting the BRI for judicial proceedings would have an even more important benefit, for the Federal Circuit's own "ordinary meaning" rule has presented many difficulties, including notoriously high reversal rates, contradictory interpretations, and demoralized trial court judges. 216 In Teva Pharmaceuticals USA, Inc. v. Sandoz, Inc., the Supreme Court tried to ameliorate the problem by distinguishing between underlying issues of fact, which are resolved by extrinsic evidence and are reviewable only for clear error, and legal issues, which rely on intrinsic evidence (the patent and its prosecution history) and receive de novo review. 217 However, that decision is likely to produce collateral litigation on whether a construction requires a factual inquiry and lead attorneys to shield district court wins by relying more on experts. 218 Although proposals before Congress would converge on the claim construction standard used in court, 219 serious consideration should be given to instead adopting the PTO's standard in litigation. In most cases, the approaches yield similar results in any event, 220 but the PTO uses the BRI for unexpired patents in all office proceedings for good reason: it believes the BRI to yield "more uniform and satisfactory results." 221 It is not surprising that this is so. The BRI provides a benchmark that the Federal Circuit's current standard lacks. In an earlier era, claims were construed to preserve their validity (and in light of the claimed infringement); once the Supreme Court's decision in Markman v. Westview Instruments, Inc. 222 made claim construction an issue of law and courts began to interpret claims in separate hearings held before evidence on invalidity (or infringement) was adduced, that canon became "a last resort, not a first principle." 223 Courts were then forced to make unguided determinations on what an ordinary artisan would understand, using a variety of tools that have changed throughout the course of the Federal Circuit's history. 224 In contrast, although the BRI has some wiggle room ("reasonable"), it includes an objective standard ("broadest") while still incorporating ordinary meanings and the traditional approach of allowing the patent holder to be its own lexicographer. As the PTAB articulates it, "[u]nder the broadest reasonable construction standard, claims terms are given their ordinary and customary meaning as would be understood by one of ordinary skill in the art in the context of the entire disclosure." 225 Of course, if valid claims are construed more broadly, the scope for infringement liability will increase. However, the BRI standard closely matches the construction that risk-averse potential users of the invention likely employ in deciding their freedom to operate. Accordingly, it may not have a significant impact on public access. Furthermore, using this standard would have a disciplining effect. If the BRI is more likely to lead to cancellation and invalidation, then drafters may begin to claim more narrowly; if it is used in enforcement actions, it may deter some infringement. Given that damages awards have been reduced and injunctive relief has become less certain, this approach could restore balance to the system. 226 As this Section demonstrates, there is much in the PTAB's decisions that could be of considerable value to the examining corps and to the courts.

The Board could, however, do better. Like the Federal Circuit, the PTAB tends to suppress almost all reference to underlying policy. Its contribution would be considerably greater if its reasoning were transparent. A dialogue between the tribunals that included candid analyses of policy options would be more fruitful, and joint consideration of these issues by the PTAB and the Federal Circuit would also enrich ultimate consideration by the Supreme Court.

While the procedures instituted by the AIA could contribute significantly to the sound administration of patent law, enjoying the benefits will depend heavily on whether the Federal Circuit is willing to grant a healthy degree of deference to PTAB decisions (that is, share its authority over patent jurisprudence) and on what it makes of the PTO's rules regulating the proceedings in which these decisions are issued. These are not easy questions. Congress interposed dispute resolution in a specialized agency under review in a specialized court. However, it failed to make clear how the two tribunals ought to relate to one another. The statute gave the PTO authority to "prescribe regulations . . . establishing and governing . . . [the three new methods of] review,"foot_8 but it did not mandate notice and comment,foot_9 thus arguably withholding formal rulemaking authority. Similarly, Congress directed the PTO to entertain inter partes PGR, IPR, and CBM review and specifically mentioned discovery, responsive filings, oral hearings, and final written determinations; the AIA also renamed the BPAI the Patent Trial and Appeal Board. 229 However the statute nowhere incorporates the phrase "on the record after opportunity for an agency hearing," 230 which is the best evidence that Congress intended the procedures to constitute formal adjudica-tions and thus intended to delegate to the agency interpretive authority entitled to judicial deference. 231 Not only is the statute unclear about how much authority Congress meant to vest in the PTO, the Federal Circuit's past approach to agency review has been decidedly mixed. As Arti Rai and Stuart Benjamin documented in 2007, despite Dickinson v. Zurko, 232 which admonished the Federal Circuit to apply mainstream administrative law to its review of agency action, the Federal Circuit has not consistently done so. Tafas v. Doll, 233 the Federal Circuit's previous foray into PTO rulemaking, produced a split decision. The court rejected a PTO rule on continuations because, in one judge's view, no deference was owed the PTO, and in the other two judges' view, because the rule had a substantive effect inconsistent with the Patent Act. 234 Substantively, while the Federal Circuit has reviewed factual findings on a substantial evidence standard (or for clear error), it has reviewed PTO legal determinations de novo. 235 The AIA's delineation of the PTO's new authorities and obligations ought to have prompted the Federal Circuit to reconsider its approach to administrative deference. But the early cases reviewing the PTO's authority and PTAB decisions, In re Cuozzo Speed Technologies, Microsoft Corp. v. Proxyconn, Inc., and Versata Development Grp., Inc. v. SAP America, Inc., suggest that the court is not ready to take systematic heed of congressional willingness to rely on the PTO. 236 Quite the reverse.

To be sure, in Cuozzo, the court referenced the Chevron framework in reaching the decision to approve the PTAB's use of the BRI. It assumedarguendo-that since Congress was silent on claim construction, the PTO could adopt a reasonable interpretation. It found the BRI to be reasonable because it was supported by the policy rationales applicable to other forms of examination. 237 The Chevron analysis was, however, extremely short (two paragraphs) and mechanical. Indeed, the court withdrew its initial opinion in the case 238 in order to add a paragraph emphasizing that it did not, in fact, agree to the use of the BRI on Chevron grounds. In the revised opinion, it denied that "Congress has newly granted the PTO power to interpret substantive statutory 'patentability' standards." 239 According to the court, "[s]uch a power would represent a radical change in the authority historically conferred on the PTO by Congress, and we could not find such a transformation effected by the regulation-authorizing language of § 316." 240 Instead, the court devoted most of its discussion of the BRI (eight paragraphs) to its own interpretation of the AIA. Noting that the BRI has been applied "for more than 100 years in various types of PTO proceedings," 241 it reasoned that Congress must have been aware of it and thus can be said to have implicitly approved its use. 242 As to PTO-imposed limitations on amendment (which, as we saw, is a major reason for the BRI controversy), the court did not apply Chevron at all. Rather, it found the availability of amendment sufficient, 243 once again reasoning that congressional awareness can be taken as adoption.

Despite the court's careful clarification that its approval of the BRI was not because of deference to the PTO, the decision to deny a rehearing en banc still drew two strong dissents. 244 The dissenters mainly claimed that since post-grant opposition was meant to be a cheap substitute for litigation, it was more appropriate to use what they termed the "actual" construction of challenged claims. 245 But echoing her position in Tafas, Chief Judge Prost also questioned whether the standard of construction fell within the PTO's procedural authority. Because in her view it did not, the choice of the BRI merited no Chevron deference at all. 246 Given the Cuozzo court's heavy reliance on past practice regarding claim construction, it is not surprising the court has been even more skeptical of PTO authority on issues unique to post-grant practices. In Proxyconn, the question was whether the Board's Idle Free decision appropriately required the patent holder to prove an amendment was patentably distinct over all the prior art in the record (and not merely the art the petitioner raised). Although this time the court expressly referenced the standard articulated in the Administrative Procedure Act (APA), 247 it mainly provided its own assessment of the rule and the procedure used to create it. 248 Versata, a CBM, was much the same. The Federal Circuit acknowledged that Congress had left it to the PTO to define the term "covered business method" and carefully noted that the PTO had considered both the legislative history and public responses to the notice it published in the Federal Register. Nonetheless, the court conducted its own assessment. 249 In the end, it decided that it agreed with the PTO that covered business methods are not limited to products and services of the financial sector, noting only at the very end of that section of the opinion that the PTO's expertise entitles it to "substantial deference in how it defines its mission." 250 It handled another issue that arose in the case similarly, deciding for itself that in CBMs, the PTAB has authority to decide whether the claims are drawn to patentable subject matter. 251 The fate of matters like the narrow constraints the PTO placed on how the parties present their cases (the time and page limits; the minimal discovery), which have no history and are not tied to the PTAB's mission, are even more difficult to determine. It is, however, worth nothing that although page, time, discovery, and argument limitations have an impact on the parties' ability to present their cases (especially in light of the broad view of the grounds that can be asserted, at least in CBMs), these limits do not appear to significantly impair argumentation, and they are well tailored to Congress's goal of creating a quick and less expensive way to weed out invalid claims. 252 Furthermore, PTAB panels enjoy the discretion to alter many of these limits in the interest of justice or for good cause. 253 Thus, if a Chevron standard were applied, these regulations appear to lie squarely within the regulatory authority Congress accorded to the PTO and to be consistent with the Patent Act and permissible interpretations of the AIA. 254 249 See Versata Dev. Grp., Inc. v. SAP Am., Inc., 793 F.3d 1306, 1323-26 (2015). For the legislative history considered, see supra text accompanying notes 94-95. 250 Versata, 793 F.3d at 1325. The court also considered the PTO's regulation defining "technological invention" but apparently found it so devoid in meaning that it did not bother to state a standard for reviewing it. 254 The PTAB procedures have also been challenged for depriving the parties of a trial by jury and for violating separation of power principles. In Cooper v. Lee, the issues were raised in a declaratory judgment action but dismissed for failure to exhaust administrative procedures. No. 1:14-cv-00672-GBL-JFA, 2015 WL 686041, at *9 (E.D. Va. Feb. 18, 2015). The district court did, however, note that the Federal Circuit had upheld the constitutionality of ex parte reexamination in a decision the Supreme Court declined to review. Id. at *2 (citing Joy Techs., Inc. v. Manbeck, 959 F.2d 226 (Fed. Cir. 1992)). B&B Hardware, Inc. v. Hargis Industries, Inc., 135 S. Ct. 1293 (2015), which held that an issue decided by the Trademark Trial and Appeal Board (TTAB) is preclusive of later litigation in an Art. III court, suggests that the Supreme Court is unlikely to find that the PTAB deprives the patent holder of a right to a jury trial.

Whereas procedurally, Cuozzo, Proxyconn, and Versata send a somewhat mixed message, substantively, they do no such thing: nothing in these decisions even gestures at the notion that the Federal Circuit is prepared to defer to PTAB decisions. 255 In reviewing the merits, Cuozzo and Proxyconn cited Teva. 256 In other words, the court adopted the standard the Supreme Court imposed on review of district court claim constructions to consider the PTAB's claim construction. Indeed, in Proxyconn, the court reversed the PTAB's construction of the claim, 257 even though the court is less versed in the application of the BRI than is the PTAB. In Versata, it went much further. First, despite having decided in Cuozzo that it had no jurisdiction to review PTAB decisions on whether to institute IPRs, 258 the Versata court redefined the issue: it decided it could review whether the PTAB had exceeded its CBM authority by hearing challenges to patents that it misidentified as drawn to "covered business methods." 259 And even though the court accepted the Board's definition of what constituted a "covered business method," and despite its own complete lack of case law on the meaning of a "covered business method" or a "technological invention," it nevertheless reviewed the PTAB decision de novo. 260 Similarly, it reviewed the question whether the invention at issue was patentable subject matter in the same way as it would have reviewed a decision of a trial court. 261 This approach may be a function of the court's skepticism as to whether the APA applies with full force to the PTO (which was created a century before the rise of the administrative state and is mentioned only marginally in the legislative history of the APA). 262 Or it may stem from a fear that the PTAB is, indeed, acting like a "death squad." Alternatively, it may derive from what Judge Hughes, in dissenting from Versata's holding that institution decisions are reviewable, claims is an "appetite for arrogating to the court the Board's statutory authority." 263 But to the extent that the court's attitude derives from the perception that the PTO lacks formal adjudicatory authority, it should be reconsidered. Plenary review is not, as Judge Hughes put it, "how Congress designed the AIA to work." 264 As Melissa Wasserman has cogently argued, the imposition of procedural safeguards, as well as the mention of "court-like proceeding[s]" in the House Report on the AIA, suggest that these procedures were to "be accompanied with a policy-making or lawmaking ability." 265 She also notes that it is hard to understand why Congress would have provided for post-grant review of "novel or unsettled legal question[s] that [are] important to other patents or patent applications" if it had not intended the PTAB to speak with the force of law and for its decisions to be entitled to deference. 266 As the system has gone into operation, other reasons to believe that Congress expected the court to give substantial deference to the PTAB have emerged. First, as the next Section describes in more detail, the same claims, patents, and families of patents, all using similar terms and concepts, can be the target of multiple proceedings; often, the challenges are brought by the same petitioners. Occasionally, what are essentially protective petitions are filed-that is, petitions challenging claims the PTAB has already cancelled, apparently filed out of concern that the cancellation decision will be reversed. 267 Unless the patent holder can rely on the near-binding effect of PTAB determinations on the parties (and a stare decisis effect on nonparties), the tax that these multiple procedures impose will be formidable. Since, presumably, Congress did not (in Judge Hughes's words) adopt these procedures to create "unnecessary and counterproductive litigation costs" or to chill innovation, it can be inferred that it intended the PTAB's decisions to be entitled to deference. 268 Notably, Congress made the estoppel effect of PTAB decisions dependent on the issuance of a final written decision of the Board. 269 In contrast, the Federal Circuit has suggested that the estoppel effect of other PTO determinations-to which deference is not accorded-is triggered only after all 263 Versata, 793 F.3d at 1337 (Hughes, J., concurring in part and dissenting in part). 264 appeals are exhausted. 270 Although the estoppel provision is tempered by the timing of the issuance of the certificate of cancellation and amendment (which is after appeal),foot_12 the difference in the trigger for estoppel suggests an intent that PTAB determinations bind not only the patent holder and the challenger, but also allow the public free use of the claimed invention as soon as the PTAB determines the claim should be cancelled. It is difficult to square that intent with de novo review of institution decisions or legal determinations. While review has been sought on almost every final written decision, 272 Congress did not provide for expedited appeals. Given the time it takes to hear appeals, and especially the historical frequency with which the Federal Circuit has reversed trial courts on the issue of claim construction, which is of central importance in these procedures, 273 it would be impossible to bring about finality in the rapid time frame Congress had in mind if such a high level of scrutiny were applied. 274 I leave the question of the precise effect of the AIA on the standard of review to the administrative law experts. For these purposes, it is enough to say that it behooves the Federal Circuit to do more than reflexively reverse everything that does not accord with its own initial assessment. Not only must it consider the more robust safeguards mandated for these procedures, it should also take the relative capacities of the two institutions into account. After all, members of the Federal Circuit have often expressed frustration with the Supreme Court's lack of expertise and intermittent exposure to patent cases. 275 But as between the Federal Circuit and the PTAB, it is the latter that is closer to the technology and sees the issues more regularly. 276 This is not to say that no review should be accorded (particularly on new issues), but given the many differences between these procedures and examination, review of facts on the "substantial evidence" standard the APA mandates for formal proceedings, 277 and review of law on the deferential end of the Skidmore-Chevron spectrum required for substantive legal and policy decisions, is not inappropriate from both a normative and an institutional perspective. 278 While the AIA can be understood as requiring deference to the PTO on both rulemaking and adjudication, experience nonetheless demonstrates the need for the PTO (or Congress) to make some changes. If the Federal Circuit does not adopt the BRI for litigation, then the dual approach will be very difficult to maintain. 279 Thus, even though the Federal Circuit approved use of the BRI, the PTO may wish to rethink it. If the PTO retains use of the BRI, then it ought, at least, to reconsider the availability of amendments. It could, for example, alter the regulations and Idle Free. For example, one reason the PTAB may be reluctant to allow amendments is that the amended claims go into force without further examination. 280 A better approach might be to have the patent holder announce proposed amendments early in the process and put the burden on the petitioner to prove these amendments are invalid. 281 It would not be a complete solution to the absence of examinationafter all, some amendments may not be of enough interest to a petitioner to trigger opposition. But if it were clear to potential petitioners that there is a viable amendment procedure and that they will be required to protect their own interests in it, perhaps fewer petitions for review would be filed. Alternatively, the PTO could clarify when a patent holder whose claims were cancelled can use ex parte reexamination or, in the case where error can validly be claimed, reissue, in order to obtain further (albeit narrower) protection. 282 276 See Wasserman, supra note 265, at 2007-17 (comparing the institutional features of the PTAB with the Federal Circuit); see also Benjamin There are other matters also worth revisiting. For example, the petitioner currently has no right to reply directly to the patent owner's preliminary response to a petition to institute. As a result, the petitioner must anticipate the arguments the patent owner will make and reply to these conjectural arguments in the initial petition itself. This uses precious pages and also requires the PTAB to wade through arguments and hypothetical counterarguments that may never become a part of the merits decisions. Allowing a reply to the preliminary response may be preferable. Additionally, the page limits and other rules appear to be leading parties to file more than one petition challenging claims in the same patent, or similar claims in a family of patents. 283 While the Board can refuse to institute when successive petitions raise "substantially the same . . . argument[ ]," that practice is within the PTAB's discretion; 284 the Board may not be willing to exercise that power if the petitioner did not have the space to fully flesh out its arguments. 285 To be sure, the Board can sometimes obtain efficiencies through formal consolidation or by assigning the same panel to hear all the cases, but multiple institution decisions and final written decisions must often be handed down and the efficiencies the PTAB achieves internally are not necessarily shared with the parties themselves. 287 The multiplicity of proceed- 287 See, e.g., Groupon, Inc. v. Blue Calypso, LLC, No. CBM2013-00033, -00034, -00035, 00044, -00046, at *1 n.1 (P.T.A.B. May 7, 2014) (decision on admission pro hac vice) (considering five petitions together, with the admonition that the parties "are not authorized to use this style heading in any subsequent papers"). Another example is furnished by the cases between Liberty Mutual Insurance Co and Progressive Casualty Insurance Co., including CBM2012-00002 (regarding a method for determining an automobile insurance premium based on data collected from monitored motor vehicle operational characteristics and operator's driving characteristics); and -00003 (regarding a vehicle monitoring system); -00010 (regarding a system that allows policyholders to access policy information on the internet); CBM2013-00002 (accessing policy information on the internet) (oral argument merged with CBM 2012-00010); -00004 (regarding a system for monitoring the operation of a motor vehicle); and -00009 (regarding a vehicle monitoring system). All these were decided by the same panel of administrative patent judges but (with one exception) were entertained as separate cases. The Federal Circuit similarly obtained efficiencies by reviewing the cases together, see Progressive Cas. Ins. Co. v. Liberty Mut. Ins. Co., ings is also problematic for trial courts, for they can never be certain when practice before the PTAB is over and any stays they issued can be dissolved. It may, in the end, be more efficient to liberalize the page and time limits and give potential challengers a longer period in which to join ongoing proceedings. Moreover, because the PTAB reconsiders issues decided at the time of institution (like claim construction) at the merits stage as well, 288 it might also be useful to give parties more scope for argument earlier in the proceedings.

The point is not that current practice is perfect. Rather, it is that if Congress fails to step in, then changes should be within the sound discretion of the PTO. Federal Circuit judges, many of whom lack case management experience, are not institutionally best situated to decide how to balance the congressional goal of efficiency against the needs of the parties to present their cases effectively.

With the addition of a procedure permitting third parties to submit information to the PTO prior to patent issuance, 289 the AIA envisioned a "cradle to grave" opportunity for the public to challenge rights to exclusivity. Preissuance submission hands off to PGR, and then to IPR, with CBMs (temporarily) hovering over claims that are especially questionable. Through the use of separate procedures, Congress attempted to strike a careful balance between promoting public access to unpatentable advances and recognizing reliance interests that mature as patent holders and licensees pour resources into exploiting their inventions. However, three different review procedures create a great deal of complexity. As we have seen, patent families can get caught in a series of successive challenges. These give rise to complicated questions on how the results of one review affect the positions that patent holders, petitioners, their privies or real parties in interest, and strangers can take in later actions.

Some of these procedures allow any interested party to challenge claims. While this, too, is commendable from an access-to-knowledge perspective, it raises procedural questions and can create opportunities for harassment. Furthermore, as Part I described, the review procedures permit the parties to rely on (and challenge) experts, make motions to exclude, amend, file objections to amendments, and present oral arguments. In practice, these oppor-No. 2014-1466, 2015 WL5004949 (Fed. Cir. Aug 24, 2015). Similarly, a combined oral hearing was held on two cases involving Apple and Sightsound Technologies. This produced two separate written opinions, issued on the same day. tunities are used in many cases. The expense raises questions as to whether public domain interests are as well protected as the proponents of these procedures hoped. The answers depend on whether problems raised by complexity can be resolved in a fair and evenhanded manner. This Part discusses these issues.

As described earlier, the AIA and PTO introduced rules on estoppel that are designed to prevent the parties from obtaining several bites at the apple. Under the regulations, a patent holder is "precluded from taking action inconsistent with [an] adverse judgment, including obtaining in any patent: (i) a claim that is not patentably distinct from a . . . canceled claim; or (ii) [a]n amendment of a specification . . . that was denied during the trial." For IPR and PGR challengers, the statute provides that once a final written decision is handed down, the petitioner, the real party in interest, and the petitioner's privies are precluded from maintaining an action in the PTO with respect to that claim "on any ground that the petitioner raised or reasonably could have raised" in the proceeding. 291 Nor can the petitioner argue in court or in the ITC that the claim is invalid on any ground the petitioner "raised or reasonably could have raised." 292 CBMs are somewhat different: once a final written decision is issued, the petitioners or the real party in interest cannot assert, in courts or in the ITC, an argument that the claim is invalid on any ground that the petitioner "raised" during the CBM. 293 But despite the considerable detail written into these provisions, it is already clear that they fail to cover all the ways in which those involved in these proceedings can nibble at the apple. Thus, for example, Colleen Chien and Christian Helmers observe that 82% of IPRs run parallel to litigation. 294 And yet the rules do not say whether the findings of fact and law made in the course of these procedures are binding on other tribunals as a matter of collateral estoppel (issue preclusion) or stare decisis. 295 The poster child for these omissions is the dispute between Versata and SAP America, where a party that lost in the trial court on the argument that certain claims were invalid, 296 then brought a CBM petition to challenge the validity of the same claims. 297 Such an action is not barred by the AIA or the regulations adopted by the PTO. And because potential infringement is continuing, there is no question of claim preclusion. Nor does issue preclusion apply: since the PTAB uses a lower burden of proof on validity and a different rule for claim construction, the issues in the review are not the same as the issues decided by the court. 298 The PTAB instituted an action and cancelled the claims. 299 But by the time the PTAB handed down a final written decision, much had happened: the Federal Circuit had affirmed the trial court's damage award, vacated part of the order for injunctive relief, and remanded to the district court; 300 that court had then awarded damages in the amount of $391 million (the injunction was abandoned). Arguably, the subsequent invalidation of the claims in the PTAB should have provided grounds for the infringement defendant to petition for relief from the $391 million judgment under Federal Rule Of Civil Procedure 60. 301 However, in a short per curiam order that offered no explanation, the Federal Circuit denied a motion to stay or for relief from judgment. 302 In a later proceeding, the Federal Circuit essentially gave the patent holder a right to contest the institution decision as part of its review of the merits of the PTAB's invalidation determination. 303 However, the Federal Circuit affirmed the PTAB on both issues, with the result that a party is now forced to pay a rather large award on account of a patent that is not valid. 304 Troubling as Versata looks, the result can be justified on the ground that it motivates those who believe a patent is invalid to challenge it quickly and put the invention it covers into the public domain. In contrast, the opposite rule-one that absolves challengers of past liability-encourages sharp practice. A party with private information about invalidity could infringe, safe in the knowledge that if it were sued, it would invalidate the patent and free itself of the obligation to pay damages, all the while retaining any extra profits it captured as a result of, essentially, sharing exclusivity with the patent holder. The Versata outcome is not even entirely unprecedented. For example, a licensee who successfully challenges a patent must pay royalties up to the point where it took the affirmative step of prompting adjudication of validity. 305 Still, if such cases proliferate, U.S. procedure will take on an aspect of German practice, where infringement suits and validity determinations are bifurcated and decided at different times by different tribunals. That practice, which regularly leads to damages based on invalid patents, has been heavily criticized for imposing unjustified cost on the technology industry. 306 Better, perhaps, would be to require the successful challenger to pay damages, but only up to the time that the petition for review was filed. Particularly for patents subject to PGR review, a rule that dates the termination of liability this way would encourage very early weeding out of invalid claims. The incidence of these problems would also be minimized if the courts and the PTAB used the same approach to claim construction, for they would then more often reach the same conclusion on validity. In addition, if the Federal Circuit holds to the position it announced on stays in VirtualAgility, and if the PTAB develops efficient procedures for consolidating multiple challenges to the same patents, it will less often happen that a patent is upheld in court before it is invalidated by the PTAB. 307 But even if the system largely avoids the Versata problem, many more subtle questions on the relationships among these procedures remain to be worked out. Not only does the AIA allow a party who lost in litigation to petition for review, it also appears to permit a party that lost a CBM to bring an IPR on any issue not actually raised in the CBM, and for their privies to maintain any action in the PTO or in court. 308 Furthermore, because IPRs can raise only questions about novelty and nonobviousness, IPR losers who are then threatened with suit could bring CBMs on other matters. More generally, because the estoppel applies only to specific claims, the same parties can file multiple challenges to other claims, even if they are within the same patent or family of patents. 309 Also, the AIA did not abolish ex parte reexamination-yet it says nothing about its relationship to the new procedures. Furthermore, both PGRs and IPRs can be brought by anyone who is not the patent holder: there are no standing requirements, as there are in court. Many (according to Dolin, nearly one third) of the patents in IPR proceedings are subject to multiple IPR requests; 310 the same may eventually happen for PGRs.

Indeed, the availability of these procedures appear to be creating new defensive strategies and business models and extending old ones. Patent risk management services like RPX Corporation, and public interest groups such as the Electronic Frontier Foundation (EFF), have begun to file petitions for review. 311 There are even firms-UnifiedPatents is an example-formed specifically to use IPRs to clear the technology space of their members. 312 Further, generic drug companies are filing IPRs instead of, or in addition to, engaging in Hatch-Waxman litigation, 313 and there are a few cases brought by well-known patent aggregators. 314 Another purported technique, alleged to be practiced by the principals behind the Coalition for Affordable Drugs, a PGR (which must be filed within nine months of issuance) could follow a written decision in a CBM (which requires that the petitioner have been accused of infringement). However, because standing to bring a CBM requires only that the challenger be "charged" with-and thus, not necessarily "sued" for-infringement, AIA, Pub. L. No. 112-29, § 18(a)(1)(B), 125 Stat. 284, 330 (2011), an IPR, which cannot be brought more than a year after an enforcement action is filed, 35 U.S.C. § 315(b) (2012), could in some cases be instituted even after a CBM proves to be unsuccessful. is to simultaneously file a petition and short the patent holder's stock in the hope that the PTAB's actions will depress the stock's price. 315 The procedures may also be encouraging "reverse trolling": soliciting payments from patent holders for the favor of not filing petitions for review or for settling cases that have been filed. 316 As Greg Dolin describes it, the result is that "the Damocles sword of post-issuance review perpetually hang[s] over the patentee's head." 317 Furthermore, while the PTO has created procedures that are swift and efficient, they are far from cheap-what with use of experts and discovery, projected costs are in the range of $150,000 to $300,000 per party. 318 This may be an order of magnitude less than a full-blown trial, but if it is multiplied several times, the cost can put a patent holder in a precarious position. Some of these problems may be solved once the concepts mentioned in the estoppel provisions ("reasonably could have raised," "privy," "real party in interest") are better defined. But the effect of that effort is likely to be limited.

As to the first phrase ("reasonably could have raised"), the patent holder would clearly have been much better protected had Congress prevented parties from asserting grounds they "could have raised"-that is, had the legisla-315 See Michael Loney, The Fine Line Between Abuse and Fair Use at the PTAB, MANAGING INTELLECTUAL PROPERTY (May 27, 2015) http://www.managingip.com/Article/3456658/ The-fine-line-between-abuse-and-fair-use-at-the-PTAB.html (describing how Bass filed fifteen IPR petitions against seven pharmaceutical companies and allegedly shorted the stock of at least some of these firms); Gene Quinn, Patent Abuse or Genius? Is Kyle Bass Abusing the Patent System?, IPWATCHDOG (Apr. 8, 2015), http://www.ipwatchdog.com/2015/04/08/is-kyle-bass-abusing-the-patent-system/id=56613/. So far, the PTAB has declined to institute these petitions. See, e.g., Coal. for Affordable Drugs v. Acorda Therapeutics, No. IPR2015-00720 (P.T.A.B. Aug. 24, 2015) (institution decision). However, the PTAB has not been willing to sanction Bass. See Kelly Knaub, PTAB Won't Sanction Bass for AIA Reviews of Drug Patents, LAW360 (Sept. 25, 2015), http://www.law360.com/articles/707663. 316 An example may be New Bay Capital, LLC, an entity that appears to make nothing but which filed several IPRs against VirnetX. See New Bay Capital, LLC v. VirnetX Inc., Nos. IPR2013-00375, -00376, -00377, and -00378 (P.T.A.B. June 23, 2013). According to VirnetX, it then received a request from New Bay for ten percent of a verdict that VirnetX had won in a patent infringement action it had brought against Apple. New Bay later filed motions to terminate the IPRs, possibly in response to threatened subpoenas. ture adopted the formulation in the old inter partes reexamination statute. 319 However, out of caution that courts would interpret the phrase to require a "scorched-earth search around the world" for prior art, the term was softened. 320 The softening leaves open the question whether "reasonably" is to be interpreted in light of what the challenger could have easily uncovered prior to filing, what it actually knew when the petition was filed, or what it thought could be handled in a single petition, given the strict page and time limits imposed by the PTO. 321 The more reasonably the challenger is treated, the higher the cost to the patent holder and to the PTAB. But even if "reasonably" is interpreted to provide the strongest possible protection to the patent holder, there would still be considerable vulnerability. The estoppel applies only to the claims that were challenged. And in CBMs, the estoppel extends only to grounds actually raised. While the PGR estoppel provision now reads the same as the IPR provision, there is reason to believe that the legislators thought that the short time frame for seeking review militated in favor of a narrow estoppel and that the error will be fixed in future legislation. 322 Even more important are the limitations inherent in the second and third phrases: the estoppel provisions apply only to the parties, their "privies," and "real parties in interest." While the legislative history indicates that these are to be equitable and practical determinations based on the activities of the parties, 323 a set of IPRs concerning patents held by VirnetX suggests these concepts may not go very far in protecting patent holders' interests. In that situation, Apple had been successfully sued by VirnetX for infringement. It then brought seven IPRs to challenge VirnetX's patent claims, but the petitions were dismissed as untimely, given the earlier litigation. 324 RPX Corporation, to which Apple subscribes, then brought seven IPRs of its own to challenge the same claims. 325 As might be expected, these petitions were all dismissed upon a finding that Apple was the real party in interest, with the Board reasoning that if Apple had been out of time, then so was RPX. 326 Notably however, Apple's mere membership in RPX was not considered enough to bind RPX to the Apple dismissal. Rather, what mattered was the compensation Apple paid to RPX, its suggestion that RPX file against patents of "questionable quality" (which the record showed included the VirnetX patents), the sharing of attorneys and experts, the similarity between Apple's and RPX's petitions, and the fact that Apple was the only RPX subscriber that had a significant interest in the case. 327 Absent these factors-and one must assume that parties will make sure that these factors are absent or, at least not ascertainable, in later cases-industry groups, trade associations, and other patent risk management organizations will likely be able to organize successive attacks on patent portfolios, aimed at debilitating the ability of the patent holder to maintain an effective defense. 328 Now that the Supreme Court has made it clear that decisions of administrative agencies are entitled to preclusive effect on the same standard as court decisions, 329 doctrines of issue preclusion and stare decisis could play an important role in protecting patent holders caught in such multiple challenges. However, because of the relative time frame of these procedures and appeals, these doctrines will be helpful only if cases like Versata, where the decision to institute was considered reviewable after the PTAB issued a final written decision, and Proxyconn, where the Federal Circuit refused to defer to the PTAB on the merits, are rare. Otherwise reversals will wreak havoc on the system. 330 Rehearings are possible, but the regulation requires they be made within thirty days of a final written decision, well before the Federal Circuit is likely to decide an appeal. 331 And even if it is assumed that patent holders can rely on PTAB decisions on law and fact, preclusion doctrines raise their own set of problems.

Once again, the thorniest issue is likely to be claim construction. As we have seen, the PTAB is frequently called upon to interpret the same claims in multiple actions brought by the same or different parties. 332 7, 2013). In those cases, the Board considered two challenges to the same set of claims, construed them, and declined to institute in No. CBM2013-000019, but did institute in No. CBM2013-00020; ultimately the Board held the claims unpatentable on the construction used upon institution. lem of whether the same construction should be used within the PTAB is solved by joining the petitions. More informally, sometimes the same panel decides all the related challenges. 333 Nevertheless, there have been situations where the question of a claim's construction arose after a final written decision on the same claim. In those cases, the PTAB appears to use the same construction, albeit without specifically stating that it is relying on preclusion principles. 334 When the claim construction is used against the patent holder, this outcome is unremarkable: the construction of the patent holder's claim is a core issue in every case and as long as the claim has not expired, the PTAB always uses the BRI. Accordingly, the same issue arising in the second action was actually litigated, decided, and essential to the judgment in a prior adjudication involving that party that reached a final conclusion. 335 Of course, if a new challenger comes along, that party has the right to seek a different interpretation of the claims. However, unless the challenger has a radically different argument, stare decisis should lead to the same result (in this way, stare decisis promotes uniform interpretation of patent claims).

Much harder is the question whether the construction announced by the PTAB binds the challenger in court. This problem will, of course, not arise if the claim is cancelled. However, if the claim is upheld under the BRI, the patent holder may well seek to enforce the patent against the unsuccessful challenger. In that case, can the accused infringer-who previously argued for a broad claim construction in the PTAB-now argue for a different, narrower, construction in court? 336 The estoppel provisions carefully left it open to the unsuccessful challenger to argue it did not infringe, but infringement is more likely if the claim is interpreted broadly (as in the PTAB) than narrowly (as in court).

Under current practice, there should be no preclusion because the issue of what the claim means to the ordinary artisan is not the same as the issue of what constitutes the broadest reasonable interpretation. 337 In a way, use of a narrower standard protects the public interest because at the margin, it allows for freer use of advanced technologies. At the same time, however, there is something unseemly about allowing a party to argue for a broad construction 333 See supra note 332. 334 See SAP Am., Inc. v. Lakshmi Arunachalam, No. CBM2013-00013, at 9 (P.T.A.B. Sept. 18, 2014) ("We addressed the construction of [a certain term] in . . . IPR2013-00194. See SAP Am., Inc. v. Arunachalam . . . . We apply the same construction in this proceeding."); see also Apple, Inc., No. CBM2013-00020, at 3; Apple Inc. v. Sightsount Techs. LLC, No. CBM2013-00023, at 26 (P.T.A.B. Oct. 7, 2014) (involving different patents, which were heard together and decided separately, but appear to have been resolved through various common issues, including some claim construction, in tandem). 335 RESTATEMENT (SECOND) OF JUDGMENTS § 27 (AM. LAW INST. 2015). 336 Chien and Helmers refer to this as the "Angora cat" play: the cat is fluffed up in the validity proceeding so it looks quite large and wetted down in the infringement action so it appears to be very small. Chien & Helmers, supra note 35, at 16. 337 See B & B Hardware, Inc. v. Hargis Indus., Inc., 135 S. Ct. 1293, 1309-10 (2015) (noting that the tribunals must be using the same legal standard).

when it suits its interest in invalidating the patent, and then a narrower construction when the issue is its own infringement. 338 While that would argue for barring the challenger (and only the challenger) from rearguing claim construction, allowing strangers to obtain a different construction-one that allows them to evade a patent enforceable against PTAB challengers-would distort competition among rival users of the same technology. Such a result might also chill use of the review mechanisms, for no one would want to be placed at a competitive disadvantage.

The discrepancy could be avoided by limiting use of the BRI to cases where there was no prior construction of the claim by a district court. 339 But that would require the PTAB to use two different approaches, depending on the order in which parallel proceedings are adjudicated. As argued earlier, it would be much better to eliminate the discrepancy by requiring district courts to adopt the PTAB's approach and apply the BRI or, as Congress is considering,foot_19 requiring the PTAB to use the district court standard. Either way, strangers would of course still not be estopped by the prior construction, but stare decisis would apply. 341 Allowing patent holders to rely on prior PTAB claim constructions in enforcement actions would streamline litigation and provide winning patent holders with a degree of repose.

There are many other issues that arise in the course of PTAB review that could also have relevance in later proceedings. 342 These include factual matters, such as the priority date accorded the application, the effective date of references, the level of skill of an ordinary artisan, what the prior art teaches and its accessibility, whether the written description is adequate, and legal questions such as whether the invention is anticipated or obvious, whether the claim is drawn to nonpatentable subject matter, or whether it is fully enabled and distinctly claimed. In some instances, these decisions are made at the institution stage-for example, when the PTAB rejects a challenge based on a finding that the written description is adequate or that certain prior art does not render the invention obvious. 343 Such decisions (decisions not to institute) are, in a sense, final and using them would provide strong protection to the repose interests of patent holders. However, they cannot be accorded preclusive effect. Panels do not regard themselves as bound by decisions such as claim construction made at that stage, even though they give the issues considerable thought and often write long opinions. Furthermore, the decisions lack the safeguards that give rise to administrative estoppel: most glaringly, decisions concerning institution are appealable, if at all, only in connection with a final written decision. 344 Nor can they be set aside by suing the PTAB in a district court. 345 Notably, the estoppel provisions of the statute turn on whether the review "results in a final written decision," not a decision to institute. 346 In cases where the issues are resolved in a final written decision, different considerations come into play. Because the PTAB and the courts use the same approach to issues other than claim construction, 347 there is good reason to give decisions preclusive or stare decisis effect. But there are complications. In some cases, the PTAB grapples with more than one distinct ground of invalidity and there may be a question as to which ground led to the cancellation, for only issues actually decided and essential to the judgment are entitled to preclusive effect. 348 In some cases, neither ground may be precluded; in others (novelty and nonobviounsess), the relationship between the issues may be such that the decision is preclusive as to both grounds. In addition, courts distinguish between ultimate and penultimate issues: mediate facts may not be entitled to the same preclusive effect. 349 For instance, in a novelty determination, there may be several ways to reach the conclusion that the prior art predated the invention; no particular date should be given preclusive effect because the significance of that date in a later case may be different. Thus, neither the parties nor the decisionmaker may have fully appreciated the need to give it the optimal degree of scrutiny. 350 The bottom line is that the approaches that the PTAB takes to these issues may be extremely valuable later, for they may form the starting points of (inquiring as to whether the stakes in the two proceedings were equivalent). further analysis of particular issues. More important, if potential challengers realize that they are unlikely to procure different views from different PTAB panels, some of the current filing frenzy may peter out. But like the definition of "reasonably" and "privity," concepts of preclusion will not make Damocles' sword disappear.

As the PTO is beginning to recognize, 351 there are many ways in which it could improve the situation itself. The page, discovery, and other limitations may be forcing challengers to split their arguments into multiple petitions (some of which may later be joined by the PTAB). 352 Petitioners may also be dividing challenges among industry groups, lest they run afoul of the estoppel provisions. As suggested earlier, it may make sense to alter some of the limits in order to allow petitioners to challenge a single patent or a family of patents more efficiently. There may also be ways to make better use of the power the AIA provides to stay, transfer, consolidate, or terminate petitions that substantially repeat the arguments in other petitions. 353 In particular, the PTO could institutionalize the process of consolidating cases in ways that benefit the patent holder as well as the PTAB.

The PTO (or an ambitious researcher) might also take a hard look at settlements to get a fuller picture of what is going on. The AIA requires that all settlement agreements and collateral agreements referring to settlements be in writing and filed with the PTO. 354 Upon request of the parties, these can be kept confidential, but they are available to government agencies and to any person on a showing of good cause. 355 As the number of settlements grows, it becomes increasingly important to learn what is going on. Are multiple proceedings forcing patent holders into improvident settlements? Are reverse trolls extracting nuisance payments? Are patent holders paying off petitioners to drop challenges that would have led to the cancellation of claims? As noted earlier, the PTO does not necessarily terminate review upon settlement (Cuozzo was such a case), 356 but the decision to terminate is discretionary. 357 With more information about the terms of past settlements, the PTO might, in the future, choose to ask for more information before agreeing to terminations.

The PTO cannot, however, alter the statutory terms of estoppel. Furthermore, it is unlikely to interpret them in ways that significantly cut down on the number of potential challenges or challengers. A broad rule on access has an important benefit: it provides more opportunity to ensure the quality of issued patents. It opens the system to parties with a variety of perspectives on the impact of patents on the public interest. Moreover, it forces the PTO to engage with the consumers of the patent system and to consider arguments that its ordinary "customers"-the patent community-would be unlikely to raise.

The last observation is significant: it suggests that Damocles' sword has two edges. Congress enacted these procedures because it was concerned that patents of dubious validity were inhibiting innovation and taxing lawful enterprises. Complaints about NPEs asserting poor-quality patents, bringing bogus infringement actions, and extorting licenses had captured legislative attention. Further, there was concern that the cost of challenging these patents in court was too high, that small firms could not afford it and even large ones were waiting to see, hoping someone else would incur the costs so they could free ride on the result. 358 The Myriad case 359 exposed another flaw: because standing in court largely limits the class of potential challengers to entities within the same industrial sector as the patent holder, no one raises questions that call the entire industry's holdings into question. By the time someone manages, so many patents may have issued, the judges become reluctant to invalidate them. 360 Thus, while Congress could protect patent holders from multiple attacks by revising the AIA to limit standing and the PTO could probably, by regulation, enlarge the scope of privity or the concept of real party in interest, these changes would undermine core rationales for establishing the procedures. Affordable Drugs, and UnifiedPatents are, in a sense, the "good guys" in that they solve the collective action problem by pooling the resources of their members. Unified, for example, uses its subscription fees to monitor troll activity, investigate prior art, challenge troll-owned patents in the U.S. Patent and Trademark Office through ex-parte reexamination and inter partes review, and purchase patents before trolls can (but never to purchase patents from trolls). When Unified purchases a patent, all . . . members receive an immediate, perpetual license to that patent. 361 The Electronic Frontier Foundation claims to be saving podcasting; 362 the Coalition says it is attacking "pharmaceutical companies sitting on ridiculous patents that are stealing from the American public." 363 Admittedly, it is important to distinguish between these defensive patent aggregators and reverse trolls, who may be using the threat of an IPR for extortionate purposes. One reason to look carefully at what is happening in settlement is to identify criteria that would enable the PTO to sort the cases accurately and in ways that protect the interests of both the public and the patent holder.

If the public value of these procedures is to be maintained, three other matters require attention. First is the question of cost. In most of these cases, the parties rely on several experts, and they often use as much discovery and engage in as much motion practice as the rules and the PTAB permit. As noted earlier, the cost of multiple actions is a heavy imposition on patent holders. But the high cost of pursuing these actions (which also involve filing fees 364 ) may also discourage challengers, especially public interest groups, such as the ones that brought the Myriad case (the American Civil Liberties Union and coalitions of scientists interested in conducting fundamental research with the isolated genes Myriad had patented).

In very strong cases, these would-be challengers could bypass these procedures and sue in court, with the intent of asking for judgment on the pleadings. Where the grounds for invalidation are based entirely on recent case law (such as Alice), an evidentiary record may be unnecessary. Indeed, courts have shown themselves receptive to that approach. 365 Litigation has a significant advantage: it avoids having to make anticipatory arguments and lining up evidence and witnesses so the process can go forward on a tight time frame. The litigation route will not, however, be effective in cases where the lower burden of proof available in the PTAB is critical. Accordingly, the PTO might consider instituting an analogue to judgment on the pleadings. For example, it could develop a fast track for challenges that do not require the kinds of preparation currently devoted to so many of the cases.

Second is the question of attorney choice. Appearance before the PTAB is limited to practitioners registered to practice before the PTO, even though petitions rarely raise issues that only a patent prosecutor can answer. While the Board may recognize backup (but not lead) counsel pro hac vice 366 and has done so in many proceedings, the requirement limits the choices both parties can make and increases their costs. In many of these reviews, both sides are already involved in litigation, so the requirement can force them to add a new member to the team and invest in bringing him or her up to speed. Also troubling is that only attorneys with "established familiarity with the subject matter at issue in the proceeding" can be recognized pro hac vice, 367 for this limitation raises the question whether attorneys for groups like the American Civil Liberties Union, the Association for Molecular Pathology, or university counsel will be allowed to appear on behalf of researchers or the public. If not, then the rule will tend to limit the type of issues raised to ones that the patent system already considers on a regular basis.

Even more important is the question of appellate jurisdiction over PTAB decisions. The AIA contemplates broad authority to appeal. It provides that "[a] party dissatisfied with the final written decision" may appeal and "[a]ny party to the . . . review shall have the right to be a party to the appeal." 368 The Federal Circuit has, however, been extremely skeptical of patent challengers, 369 and its decision in Consumer Watchdog v. Wisconsin Alumni Research Foundation 370 (WARF) could make these procedures profoundly unattractive as a means of protecting the public interest.

In that case, a "not-for-profit public charity dedicated to providing a voice for taxpayers and consumers in special interest-dominated public discourse, government and politics," 371 challenged WARF's stem cell patents, which have historically created a substantial obstacle to important life sciences research. The charity lost before the PTAB; when it sought to appeal, 366 it was denied the right to do so on the ground it lacked standing. 372 The patent was upheld in a reexamination, and not in one of the new procedures. However, the reexamination statute likewise gives participants a right to appeal. 373 That provision made no difference to the Federal Circuit. Citing Lujan v. Defenders of Wildlife 374 for the proposition that Congress cannot evade the constitutional limits set out in Article III and Hollingsworth v. Perry 375 for the principle that the same "constitutional requirements for standing apply on appeal, just as they do before district courts," 376 Judge Rader reasoned that because Consumer Watchdog is not itself engaged in research or commercial activities involving stem cells, it had not sustained an injury in fact, traceable to the challenged action, that could be redressed by judicial action. Thus, he concluded it did not have standing to appeal.

The Supreme Court denied certiorari in the case. 377 However, the outcome certainly bears reconsideration in the context of the new procedures, for the inability to appeal will surely chill their use by parties Congress may have been particularly interested in attracting. Significantly, neither Lujan nor Hollingsworth is directly on point. Lujan involved standing to bring an action in an Article III trial court; Hollingsworth concerned the right of a party that had not participated in the trial to appeal the district court's decision. Thus, neither case addressed the right of a party who had suffered judgment to have the adverse decision reviewed. In contrast, the facts in ASARCO Inc. v. Kadish 378 (which Judge Rader failed to cite) are much closer to those of Consumer Watchdog. In ASARCO, the Court found jurisdiction to review a state court judgment that invalidated a mining claim, even though under Article III, the petitioners could not have raised their federal question in a federal court in the first instance. The Court applied the Lujan analysis (in that sense, the Federal Circuit was correct). However, the Court analyzed the case from an appellate perspective, taking into account the parties' position after the state court judgment was rendered. According to the Court,

[I]t is undisputed that the decision to be reviewed poses a serious and immediate threat to the continuing validity of [petitioners' interests] . . . . The state proceedings ended in a declaratory judgment adverse to petitioners, an adjudication of legal rights which constitutes the kind of injury cognizable in this Court on review from the state courts. 379 The Court therefore held the petitioners had standing to appeal.

A challenger who loses in the PTAB is in a very similar position to the parties that sought review in ASARCO. As we saw, the estoppel effect of PTAB decisions is triggered by the final written decision. Thus, under the statute, the challenger can no longer seek to invalidate the patent claim on any ground it raised-or for PGRs and IPRs, reasonably could have raised-during the PTAB proceeding. Furthermore, in both PRGs and IPRs, the parties who are in privity with the challenger or are considered its real parties in interest are also estopped. In an RPX-type case, for example, that would include RPX members such as Apple. If the concept of privity were expanded as described in the previous Part to protect patent holders from abuse, it would also include other RPX subscribers, all of whom would have to seek licenses to practice patents that may, in fact, be invalid.

Arguably, a party that does not have the right to appeal ought not be estopped. However, it is far from clear the Federal Circuit would see things that way. There is, after all, no federal right to an appeal. Moreover, ASARCO dealt with a very similar argument and rejected it on the ground that removing the preclusive effect of the state court judgment would undercut it and, in effect, impose federal standing requirements on a tribunal that is not subject to Article III. 380 Furthermore, even if the challengers were not technically precluded, the PTAB judgment could have effects equally adverse to the interests of the challenger, its privies, and its real parties in interestindeed, adverse to the interests of all members of the relevant industry. Given the lower burden of proof and broader claim construction used by the PTAB (as well as the win/loss statistics presented above), any patent that survives PTAB review becomes essentially "bullet proof." No jury, for example, is likely to find a patent invalid after being told that the experts on the PTAB (indeed, a "death squad") had upheld it. 381 To be sure, there are important differences between ASARCO and Consumer Watchdog. While the PTAB, like a state court, is a non-Article III tribunal, the respect that the Supreme Court accords to the decisions of the courts of other sovereigns is unlikely to be conferred on the Board, which is merely an arm of a federal administrative agency. Furthermore, in ASARCO, the lower court had invalidated the petitioners' mining claims. Thus, they were directly injured by the decision. Because the lower tribunal had upheld the patent claim, the harm to Consumer Watchdog is different. Nonetheless, a (permitting the United States to appeal a decision on the constitutionality of the Defense of Marriage Act even though the Executive agreed with the trial court); Camreta v. Greene, 131 S. Ct. 2020, 2030 (2011) (permitting a winning party to appeal a decision on qualified immunity in order to dispositively establish the boundaries of a doctrine that will "have a significant future effect"). 380 ASARCO, 490 U.S. at 622. 381 Cf. Gunn v. Minton, No. 133 S. Ct. 1059 (2013). In Gunn, the Supreme Court was not concerned about the effect of a state court decision on a patent law matter. Unlike the PTAB, which deserves considerable deference on its patentability determinations, state court determinations of patent issues have little influence on how federal courts analyze them. See id. validated patent causes harm that is immediate and concrete. For example, the patents RPX was challenging had already been enforced against Apple successfully.

Significantly, the Supreme Court has long endorsed special rules on justiciability in particular classes of cases, 382 and these appear to include patent validity cases. As Justice Brennan pointed out in Franchise Tax Board v. Construction Laborers Vacation Trust, federal courts have always adjudicated claims of invalidity brought by alleged (not merely proven) patent infringers. 383 In Cardinal Chemical Co. v. Morton International Inc., the Supreme Court, having "identified a strong public interest in the finality of judgments in patent litigation," prevented the Federal Circuit from vacating a finding of patent invalidity, even though the decision in the case rested on the adequate and independent ground that the patent was not infringed. 384 The Court was concerned that if the patent was not invalidated, it would "impose[ ] ongoing burdens." 385 Similarly, in Electrical Fittings Corp. v. Thomas & Betts Co., the Court permitted the party that had won below on a finding of noninfringement to pursue an appeal in order to have a judgment of validity vacated. 386 It is also worth noting that the Federal Circuit's several attempts to restrict the availability of declaratory judgments of invalidity have been rebuffed by the Supreme Court on the theory that "the public . . . has a paramount interest in seeing that patent monopolies . . . are kept within their legitimate scope." 387 Finally, narrowing standing in these cases would not serve the goals of Article III and could do serious mischief. 388 While standing is often thought to have a gatekeeping dimension, fees, coupled with the criteria for institution, effectively filter out frivolous challenges. Article III is also said to play a role in screening out cases raising federalism concerns, but patents are federal rights that preempt state law. 389 Nor does a narrow standing rule protect the integrity of the adjudicatory system the way that barring ideological litigants sometimes protects the judiciary from manipulation. Quite the contrary: if the participation of these litigants is chilled, the PTO will once again hear only from those it regulates. 390 Instead of being exposed to broader, public-interest-based considerations, the PTO will become as subject to capture as it was before these procedures were instituted. Additionally, these proceedings are important in part because they ameliorate the collectiveaction problem that the Supreme Court created with its decision in Blonder-Tongue Laboratories, Inc. v. University of Illinois Foundation to give nonparties the benefit of a judgment invalidating a patent. 391 It would be ironic if the Court's Article III standing jurisprudence stood in the way of mitigating Blonder-Tongue's damaging side effect.

While Congress cannot ignore the requirements of Article III, the question of who has standing to appeal raises very close questions; the court that considers them would do well to take account of the reasons these procedures were crafted as they were. 392 At the very least, the Federal Circuit should look behind the interests of the named petitioner. Unlike other cases involving standing to appeal agency actions, where evidence on the party's relationship to the decision is not readily ascertainable, 393 these procedures require the petition to identify the parties and related matters that would be affected by the proceeding. 394

As a member of the National Academies Committee that brought attention to the need for (what was then called) post-grant opposition, I read these decisions with deep satisfaction. The PTAB is, indeed, the "technologically sophisticated environment" in which questionable claims can be efficiently challenged. 395 The judges display considerable familiarity with the technologies at issue and bring to the adjudication a deep understanding of patent jurisprudence. As the statistics suggest, the petitions target patents of genuinely dubious value. The cases also restore a degree of public confidence in the patent system. For example, while the PTAB has instituted a large number of challenges to patents owned by NPEs, the invalidation rate for these patents has not been particularly high, suggesting that the so-called 390 Brilmayer, supra note 388, at 316. Brilmayer observes that standing is more likely to be granted when the issue will not otherwise be litigated-precisely the fear sparked by the Federal Circuit's reluctance to find standing in Myriad. Id. 391 402 U.S. 313 (1971). 392 Cf. Lujan v. Defs. of Wildlife, 504 U.S. 555, 580 (1992) (Kennedy, J., concurring in part and concurring in the judgment) ("In my view, Congress has the power to define injuries and articulate chains of causation that will give rise to a case or controversy where none existed before, and I do not read the Court's opinion to suggest a contrary view."). trolls may, in fact, be doing more good (by monetizing patent holdings and lowering transaction costs) than harm.

Still, there are many issues that demand further thought. So far, the Federal Circuit has displayed significant reluctance about sharing its authority. Nevertheless, there are both institutional and normative reasons to give the Patent Office a larger role in the development of patent jurisprudence and to allow the PTAB to serve as a true partner in improving patent law and the administration of justice. PTAB decisions are technically superb and offer opportunities for dialogue among courts on open questions, including implementation of Supreme Court cases and the substantive provisions of the AIA. However, the PTAB hews closely to the Federal Circuit tradition of suppressing all discussion of policy. As a result, percolation is not likely to be as effective as it might be. Difficult institutional questions are raised by injecting inter partes adjudication in a specialized agency into a system that also benefits from an expert appellate body.

The multiplicity of procedures creates a great deal of complexity. The rules on estoppel, although fairly detailed, nonetheless expose patent holders to multiple challenges. While there are several straightforward ways to achieve greater repose, the solutions tend to come at the expense of the public interest in weeding out bad patents and the benefits of exposing the PTO to a broader array of concerns. Finding the right balance will therefore be difficult. For both patent holders and challengers, cost is also a significant factor. While the procedures are less costly than litigation, the expense may discourage some challengers and impair the innovation industry's ability to defend-or indeed, to rely on patents to protect business interests. For district courts, multiple proceedings make it difficult to decide when to impose a stay or dissolve it. It may be possible to revise the procedures in ways that promote consolidation, but there are tradeoffs between efficiency and effective process.

While this Article attempted to look beyond the numbers, there is much further work to be done looking at the numbers. Empirical attention to how these challenges proceed could tease out information that might help bring down costs. More analysis of who is using the system would help answer questions about whether the opportunities for harassment outweigh the benefits, especially given the federal courts' current receptivity to judgment on the pleadings. The settlements in particular cry out for investigation. As the numbers rise, the question whether the settlements are abusive, collusive, or improvident becomes increasingly pressing. These procedures also create mixed incentives. Their availability may encourage more careful drafting and due diligence-or they may encourage patent holders to create deeper portfolios of rights in an effort to raise the cost of effectively challenging their exclusivity. Empirical work is necessary to determine which effect dominates.

The PTAB is, indeed, well positioned to give the Federal Circuit a run for its money. It will be interesting to see how the Board tackles the many