Renewed Efficiency in Administrative Patent Revocation

I.     Introduction

More than six years since Congress empowered the Patent and Trademark Office (“Patent Office”) to reevaluate and revoke issued patents through administrative trials, the agency continues to seek an efficient structure for its system of review. Although administrative review of patent validity has existed in other forms since 1980, the 2011 America Invents Act (“AIA”) created the first truly adversarial agency trial proceedings that could act as credible substitutes for the federal courts in adjudicating patent validity.1 Court-agency substitution is a central feature of all three AIA proceedings: inter partes review, covered business method review, and post-grant review.2 What makes these substitutes attractive are several salient differences from judicial process.

Modern Patent Office review is intended to make it systematically easier to invalidate patents generally—so that patents of questionable quality will be more likely to fall. For example, patent validity review under the AIA has no standing requirement, unlike court proceedings under Article III.3 AIA reviews also allow patents to be revoked by a preponderance of evidence rather than by clear and convincing evidence, as courts demand.4 And before deciding whether a patent is so broad as to be invalid, AIA reviews assumed until recently that the patent has broader scope than it would in litigation.5 Indeed, AIA proceedings have been so successful at their intended purpose that the very idea of supplanting the primacy of Article III courts has provoked fundamental constitutional disputes about who can properly revoke vested patent rights.6 Last Term, the Supreme Court resolved the major constitutional dispute in favor of allowing Patent Trial and Appeal Board (“PTAB”) review to continue.7

Amid the grander debate, however, a quieter structural struggle has also been unfolding. Its focus is the efficient administration of Patent Office proceedings under the AIA and the agency’s control over its own docket as both a policy lever and a case management lever. The agency’s PTAB, which conducts all three types of reviews,8 has important institutional features that reflect how the agency has tried to fulfill its mandate from Congress and what normative choices it has made along the way. And in this more specific controversy, the Court in SAS Institute v. Iancu upended a cornerstone of PTAB administration—partial institution.9

Before SAS Institute, the PTAB did not always just institute or deny a petition for review. It routinely chose a third way, instituting petitions in part and denying them in part. The Patent Office view on this matter was straightforward. The authorizing statute was ambiguous about the criteria for instituting review, and the ability to pick and choose among patent claims arguments in a petition was a commonsense lever for efficiency. The power of partial institution let the PTAB focus on the most relevant and meritorious arguments, dispense up front with unavailing arguments, and proceed to trial as well as final judgment more quickly. This last point is especially important to case management, as the AIA imposes a deadline for PTAB reviews to conclude.10

As a result, the Court’s reversal in SAS Institute—forbidding partial institution—has already received much criticism for sacrificing efficiency on the altar of wooden statutory interpretation.11 The Court’s opinion drew two dissents, including a brief and important rejoinder by Justice Ginsburg and joined by Justices Breyer, Sotomayor, and Kagan.12 While detractors of the SAS Institute decision may have reasonable quarrels with the outcome, the charge of inefficiency is largely misplaced. In particular, Justice Ginsburg’s mock proposal to illustrate a “uselessly” duplicative route that the Patent Office could take to neutralize the impact of SAS Institute is actually a sound approach that would yield meaningful systemic benefits.13

Ending partial institutions was the reasonable and appropriate thing for the Court to do. The majority opinion marks a return to the efficiencies that Congress actually put in place when it made the Patent Office into a meaningful substitute for the federal courts. Part II elucidates an aberration in PTAB estoppel, which departs from traditional principles of res judicata and collateral estoppel. This aberration lies at the heart of the AIA’s system for patent revocation and animated the controversy over partial institution.

Part III connects this aberration to the SAS Institute dispute and the Court’s answer. Next is the aftermath of SAS Institute, including Patent Office guidance issued shortly after the decision and the incentives that are likely to drive both litigant behavior and agency behavior in currently pending cases where partial institution must now be corrected.

Part IV zooms out to evaluate the empirical scope of partial institution over its six-year lifespan, discussing operational data on how aggressively the PTAB managed its docket before SAS Institute. The data reveals that, depending on how the agency implements the Court’s decision, its workload going forward could rise substantially even if the incoming body of petitions were the same, though this is also unlikely given the effects on petitioner and patent owner incentives. Next is an earnest appraisal of Justice Ginsburg’s suggestion in her rejoinder on the SAS Institute opinion, which was offered as a straw man, but offers real benefits. The discussion concludes by addressing important open questions that are now likely to receive renewed attention from the Federal Circuit in the near future.

II.     The Preclusion Paradox in Patent Office Revocation

The animating principle of patent revocation under the AIA is a basic tradeoff. Petitioners who wish to challenge a patent’s validity can do so in an agency forum that is more accessible, affordable, expeditious, and expert than the Article III courts.14 Patent owners, meanwhile, are to be insulated from relitigation of their patents.15 For example, it would not be appropriate that a patent owner that successfully defends its patent in the PTAB should have a court invalidate it anyway. Similarly, a challenger that fails to invalidate a patent in the PTAB should not be permitted simply to bring another PTAB case against the same patent repeating the same arguments. In short, the familiar procedural virtues that judgments should have finality, resources should be conserved, and prevailing parties should have repose are all as relevant for agency proceedings as for the court proceedings they replace.

A.     PTAB Estoppel and Its Aberration

The AIA mediates this tradeoff primarily through a set of statutory estoppel provisions. Estoppel is a feature of all three PTAB validity review systems and applies to future proceedings both in the Patent Office and in other fora, including the U.S. district courts and the U.S. International Trade Commission. However, the peculiar doctrinal details of PTAB estoppel deviate fundamentally from traditional principles of res judicata and collateral estoppel. As a result, PTAB estoppel paradoxically undermines the very substitutionary purpose that the PTAB was intended to serve. The PTAB practice of partial institutions simply aggravated the problem and led to a variety of further systemic distortions.

Estoppel is strategically essential to administrative revocation under the AIA. A final written decision by the PTAB in inter partes review or in post-grant review bars the petitioner in a future proceeding from raising not only the same grounds but also other grounds that the petitioner reasonably could have raised (but did not) during the original agency proceeding.16 Moreover, not only is the petitioner itself estopped, but so is any real party in interest or privy of the petitioner.17 In covered business method review, the estoppel is only slightly softer: It applies to future PTAB proceedings but not to future civil actions and other proceedings.18 Real parties in interest and privies remain on the hook.19 This framework creates an aberrant hybrid estoppel that combines features of res judicata and collateral estoppel.

Both forms of preclusion require a final judgment on the merits in the first case.20 Res judicata further requires that the second case present the same claim or cause of action that the first case resolved.21 Although doctrinal specifics vary by jurisdiction, two claims are generally considered the same for res judicata purposes if they arise out of the same transaction (i.e., the same relevant facts): “whether the facts are related in time, space, origin, or motivation, whether they form a convenient trial unit, and whether their treatment as a unit conforms to the parties’ expectations or business understanding or usage.”22 The parties must also be the same from the first action to the second, though this is a pragmatic inquiry as well: Nonparties to the first action may be bound if their interests were adequately represented.23 As a result, because the claims and parties are the same, all relevant arguments from the first case—whether actually raised or not—are deemed foreclosed.

Collateral estoppel cuts more narrowly but more deeply than res judicata. Rather than foreclose relitigation of a whole case, it reaches only specific issues within the first case, such as individual elements of a prima facie claim or relevant questions of fact.24 These issues can be barred from reassertion not only in the same claim but also in different, unrelated claims.25 However, to be precluded, the issue in dispute must have been essential to the first judgment.26

PTAB estoppel combines these features in a peculiar way.27 Petitioners who lose in the PTAB, with respect to a claim in a patent,28 lose the ability to raise “any ground that the petitioner raised or reasonably could have raised”29 from that first case in any subsequent case, whether in the PTAB again or in a civil action or ITC proceeding.30 Privies and real parties in interest are estopped as well.31 This appears on its face to be a straightforward codification of res judicata. It deals throughout with requests for a declaration of patent invalidity. It binds only parties or those who share closely connected interests. And it forecloses all available arguments, both those actually raised and those that reasonably could have been raised.

However, doctrinal elements of collateral estoppel are present as well. Although the three PTAB proceedings under the AIA all differ in how they handle requests for patent invalidation, they are all equally concerned with the overall issue of whether the patent is valid. Whereas post-grant review applies only to patents issued subject to the first-inventor-to-file, provisions of the AIA went into effect32 and covered business method review is limited to certain data processing-related inventions,33 both proceedings allow a patent to be challenged as to all major legal grounds of patentability. The same is not true of inter partes review, which allows challenges based only on two legal grounds: that the invention lacks novelty or that the invention is obvious.34 Similarly, inter partes review allows only certain kinds of evidence—patents and printed publications—in support of a validity challenge, while post grant review and covered business method reviews have no such limitation.35 Thus, it is unclear at what level of generality to define the issue that may be precluded: patentability in general, or a particular ground. Outside the PTAB, civil actions and ITC proceedings allow an even wider range of legal claims and causes of action in which arguments about invalidity may arise.36

In short, the relevant transactional facts that define PTAB claims for res judicata purposes can vary considerably, even though they may all be concerned generally with patent invalidity. PTAB estoppel applies to all of these different contexts alike, suggesting a collateral estoppel-like depth of reach. However, unlike common-law collateral estoppel, the issue in dispute need not have been essential to the initial PTAB judgment on which estoppel will be based. Indeed, the issue need not even have been raised.

B.     Why PTAB Estoppel is Problematic

Understandably, then, petitioners who mount patent challenges in the PTAB are quite resistant to treat the agency review process as an actual substitute, for the strategic benefit of investing in the strongest possible set of arguments for today’s PTAB review could well be outweighed by the strategic cost of squandering those arguments in a single PTAB review that might turn out to fail. For example, say a challenger considering inter partes review could conduct a basic search and discover prior art Reference A or, alternatively, could conduct a more extensive and costly search and discover prior art References A and B.

For estoppel purposes, the strategic choice is whether to invest more now in inter partes review or to conserve resources for a future proceeding. If discovered today, References A and B could both be asserted—indeed, would have to be asserted or else lost—in support of the current inter partes review. If the challenge succeeded, then the matter would be settled, at least as to the relevant patent claims. But if it failed, then estoppel would bar the assertion of both pieces of prior art in any subsequent case, even a subsequent case involving a different cause of action but the same issue.

By contrast, investing in merely a basic search today, and discovering and asserting only Reference A in the current inter partes review, would leave open an option that if the current inter partes review failed, then additional resources might still uncover Reference B for a future case. Under this approach, because Reference B was not known to the challenger at the time of the first PTAB challenge, there is at least room to argue that Reference B was not a ground that the petitioner raised or reasonably could have raised (i.e., that Reference B is not subject to estoppel in the future). Ultimately, the decision whether estoppel applies or not will turn on whether the adjudicator in the second case agrees that the challenger should have known about Reference B. The problem, therefore, is that there may be strategic incentive for the challenger to punt on this question and force a difficult adjudicatory decision later based on limited information.

In short, a balance must be struck. Investing less in today’s PTAB challenge leaves options for the future in case today’s challenge ends in defeat, but investing less also makes today’s challenge that much more likely to end in defeat. Thus, the effect of hybrid estoppel in the PTAB is that PTAB reviews that were intended as a substitute for judicial process, and whose estoppel provisions were crafted with the same intent, may nevertheless result in the opposite effect. The petitioner, rather than make its best case the first time around, may be better served in many situations by making only a second-best case, frustrating the AIA’s substitutionary goals of finality, repose, and minimizing duplication.

Partial institution made the problem considerably worse. Because this practice allowed some legal grounds, factual grounds, or both to be excluded from the proceeding that the petitioner had sought, it added another layer of uncertainty to estoppel. In addition to worrying whether it would be better to invest in making all possible arguments today or to preserve option value for the future, petitioners had to worry how grounds asserted in the petition but excluded from the merits’ adjudication would be treated by a later tribunal.

Would the later tribunal hold that these grounds had been raised because they were asserted in the petition, or would their non-institution be enough to exempt these grounds from estoppel? If a petitioner omitted such grounds in the first place based on expectations about what the PTAB would partially accept or deny anyway, would the later tribunal nevertheless hold that these grounds reasonably could have been raised, or would reliance on PTAB precedent be enough to exempt these grounds from estoppel?

III.     SAS Institute and Its Aftermath

These uncertainties were no idle theoretical speculation but came before the PTAB and Federal Circuit in multiple cases during the early years of AIA review. Even in cases where the Federal Circuit purported to resolve certain aspects of the problem,37 the cases often arose in odd procedural postures or came as split panel decisions that kept the underlying jurisprudential debate alive.38

A.     Partial Institution and Estoppel

The problems of partial institution and its estoppel effects were at the heart of SAS Institute v. Iancu, which began as an inter partes review by SAS against a software-related patent held by ComplementSoft.39 Although the SAS petition alleged that all sixteen claims of the patent were invalid, the PTAB instituted review as to patent “claims 1 and 3–10,” denying institution as to the rest.40 Based on the decision to institute, which the AIA makes “final and non-appealable,”41 the PTAB held a trial and issued a final written decision addressing the claims on which it had instituted review.42 SAS appealed, arguing that it was entitled to “a final written decision with respect to the patentability of any patent claim challenged by the petitioner”—and that this included every claim challenged in the petition.43 ComplementSoft declined to defend its victory beyond the Federal Circuit, and the Patent Office intervened.44

The Patent Office argued that it was required to issue final written decisions only as to those patent claims on which it had instituted review.45 In other words, the agency claimed that it could define its statutory obligations for itself by making partial institutions. Moreover, relying on the Supreme Court’s 2016 decision in Cuozzo v. Lee, the Patent Office also argued that its nonappealable discretion to institute or not placed the entire dispute over partial institution outside the jurisdiction of the federal courts to review at all.46 This extraordinarily broad view of agency autonomy reveals how important a building block partial institution was to the adjudicatory apparatus of the PTAB. 47

Indeed, it had been this way from the start. The power to make binary decisions was in the statutory text of the AIA itself, which referred to “[t]he determination by the Director whether to institute . . . review.”48 The supposedly lesser-included power to make partial decisions rested primarily on the syntax of the institution criterion. In the case of inter partes review, “[t]he Director may not authorize . . . unless . . . there is a reasonable likelihood” of success as to at least one challenged patent claim.49 For post-grant and covered business method review, “[t]he Director may not authorize . . . unless” there is a preponderance of success as to at least one challenged patent claim.50

The “may not/unless” framing meant that likelihood of eventual success was clearly a necessary condition to institute review, but it was unsettled whether this was also a sufficient condition. Moreover, the need to find a likelihood of success “with respect to at least [one] of the claims challenged” potentially meant that institution called for claim-by-claim evaluation and selection.51 In this, the Patent Office saw ambiguities and resolved them by concluding that likelihood of success is not a sufficient condition, and that textual support for claim-by-claim analysis permitted the agency to proceed piecemeal, on “all or some” of the claims challenged and grounds asserted.52

The Court in SAS Institute held, however, that the textual structure of the AIA is unambiguous as to the binary nature of the institution power.53 And even if likelihood of success is not a sufficient condition, said the Court, there is no warrant in the statute for claim-by-claim agency curation of the case. The Court explained that the statute’s text does not require each claim be examined separately but instead to decide whether at least one claim will succeed and that “a reasonable prospect of success on a single claim justifies review of all.”54 Central to this conclusion was the substitutionary purpose of PTAB review to stand in for federal courts. “Much as in the civil litigation system it mimics,” the Court explained, “in an inter partes review the petitioner is master of its complaint and normally entitled to judgment on all of the claims it raises, not just those the decisionmaker might wish to address.”55

The Court similarly dispatched arguments connecting partial institution to the ultimate question of final written decisions, concluding that partial institution is outside the agency’s powers under the statute and that a final written decision must, indeed, address every claim challenged in the petition.56 Importantly, the Court also held as a threshold matter that the nonappealability of PTAB institution decisions does not shield ultra vires practices such as partial institution, thus placing a further limit on the Court’s earlier decision in Cuozzo.57 This point is of particular salience to open questions that are now likely to come before the Federal Circuit, as discussed in Part IV.58

The opinion of the Court drew two dissents. The lengthier dissent came from Justice Breyer arguing point by point against Justice Gorsuch’s textual analysis for the majority.59 Of particular relevance to efficient administration in the PTAB, however, was a one-paragraph dissent from Justice Ginsburg rejecting what she termed the majority’s “wooden reading” of the statute.60 She suggested instead that the PTAB could easily restore its past practice of filtering unmeritorious arguments through an additional formalistic step, rendering the majority opinion little more than a requirement for “the Board to spend its time . . . uselessly.”61 As Part IV explains, Justice Ginsburg’s mock proposal is actually a sound plan and would offer significant benefits in both the short and long term.62

Now that the Court has issued its decision and partial institution is no longer available to the Patent Office, three important strategic questions have arisen about the incentives both of the litigants and of the agency.63 First is the question of how the PTAB will treat future petitions seeking review.64 Second is how the PTAB will treat the cases that were pending when SAS Institute was decided, where the agency made partial institutions and must now correct that error.65 Third and finally is how litigants are likely to respond to the changed agency landscape that now awaits them.66

B.     Future Petitions after SAS Institute

Two days after the decision in SAS Institute, the Patent Office issued its Guidance on the Impact of SAS on AIA Trial Proceedings laying out the agency’s policy for coming into compliance with the decision.67 Going forward, “the PTAB will institute as to all claims or none”—with no partial institution.68 Where the PTAB does institute review, it “will institute on all challenges raised in the petition.”69

It is remarkable that the guidance goes beyond what SAS Institute seems to require, which in its essential form is that that petitions be instituted fully as to all patent claims or not at all.70 The difference matters because a “challenge,”71 as the PTAB guidance uses that term, may consist of more than just a patent claim that is being attacked. A challenge may also include each statutory basis asserted for the invalidity of the patent claim, or the prior art cited as evidence of the claim’s invalidity, or both. In short, a challenge for PTAB purposes may be defined at different levels of generality, each with its own implications for agency workload as well as agency obligations of reasoned decision making.

Consider, for example, an inter partes review petition that seeks to invalidate only two patent claims, by a petitioner who challenges each claim on both available statutory grounds for invalidity (lack of novelty under 35 U.S.C. § 102 and lack of nonobviousness under § 103)72 and who cites two pieces of prior art, Reference A and Reference B.

Focusing, as SAS Institute does, on the patent claims challenged in the petition would give the PTAB only two arguments to consider:

  1. Claim 1 is invalid; and
  2. Claim 2 is invalid.

Focusing, in somewhat more detail, both on patent claims and on relevant statutory grounds for asserting invalidity would give the PTAB four arguments to consider:

  1. Claim 1 lacks novelty (§ 102);
  2. Claim 1 is obvious (§ 103);
  3. Claim 2 lacks novelty (§ 102); and
  4. Claim 2 is obvious (§ 103).

And finally, focusing in greatest detail on claims, statutory grounds, and prior art would give the PTAB—even in this trivial example—as many as ten arguments for purposes of an institution decision:

  1. Claim 1 lacks novelty (§ 102) based on Reference A;
  2. Claim 1 lacks novelty (§ 102) based on Reference B;
  3. Claim 1 is obvious (§ 103) based on Reference A;
  4. Claim 1 is obvious (§ 103) based on Reference B; and
  5. Claim 1 is obvious (§ 103) based on combining References A and B.
  6. Claim 2 lacks novelty (§ 102) based on Reference A;
  7. Claim 2 lacks novelty (§ 102) based on Reference B;
  8. Claim 2 is obvious (§ 103) based on Reference A;
  9. Claim 2 is obvious (§ 103) based on Reference B; and
  10. Claim 2 is obvious (§ 103) based on combining References A and B.

As this example reveals, conceptualizing “challenges” in broad, general terms reduces the number of distinct arguments the PTAB must consider, and vice-versa. The more distinct arguments the PTAB must consider, the higher its trial workload will be and the more fully it will have to explain its final decision. Coming on the heels of a defeat in SAS Institute, where the Patent Office had staked out a firmly minimalist view of those very obligations, why would it now issue guidance that goes even further beyond what the Court required?

There are at least two plausible explanations, and both reflect institutional caution. One is that the agency, freshly chastened by the Court, is now interested in leaving no doubts about its compliance. Indeed, this is consistent with the full language of the Patent Office regulation that authorized partial institution:

(a) When instituting inter partes review, the Board may authorize the review to proceed on all or some of the challenged claims and on all or some of the grounds of unpatentability asserted for each claim.

(b) At any time prior to institution of inter partes review, the Board may deny some or all grounds for unpatentability for some or all of the challenged claims. Denial of a ground is a Board decision not to institute inter partes review on that ground.73

Having earned the Court’s disapproval in cherry-picking among patent claims, the Patent Office may have lost its political appetite for cherry-picking among grounds for unpatentability as well.74

The other explanation, which is not mutually exclusive to the first, is that looking to claims alone does not adequately describe the workload associated with the case or the necessary fullness of the eventual agency decision. The key premise of the example above, after all, is that the true depth of a given case must account for two things. One is the set of patent claims that the petitioner challenges. The other is the set of statutory grounds that the petitioner asserts and the prior art evidence that the petitioner cites in support. Regardless of how general or specific the conceptualization of the issues involved, the analytical work actually to be done remains the same. And wherever on this spectrum the Patent Office chooses to specify how it will approach PTAB cases, the agency should state its choice and explain it.

The reason is that the Patent Office, like all administrative agencies, is obliged to reach its conclusions by reasoned decision making.75 In general, this means there must be rationality in the process of deciding, not that the decision itself must be correct.76 Accordingly, the Patent Office must strike a balance between the benefits of reducing its workload by framing PTAB issues in general terms and the costs of potential reversal by the Federal Circuit for failing to frame the issues specifically enough to support its conclusions fully.

In fact, the Patent Office has good reason to fear such reversals. In a recent string of appeals from the PTAB, the Federal Circuit has been sending cases back to the agency for further explanation.77 The basis for these remands is the familiar Chenery doctrine, which holds that courts can uphold agency action only on the grounds that the agency itself has articulated, not on other adequate grounds that the court itself might find satisfactory.78 Appropriately, Chenery itself is also a bulwark of reasoned decision making by agencies, ensuring that courts respect even agency judgments with which they disagree—so long as those judgments are rooted in transparent and defensible explanations.79

Thus, it seems likely that the Patent Office’s commitment to institute review on all “challenges” refers not only to patent claims but also to arguments of law (i.e., the available statutory grounds for invalidity). It might additionally refer to arguments of fact such as prior art, though this would push reason-giving to its maximum, with no cost savings from conceptualizing “challenges” for institution purposes even somewhat narrowly.

C.     Currently Pending Cases

More problematic than future petitions are currently pending cases in which the PTAB has already made partial institutions. The guidance that the Patent Office issued just after SAS Institute states a policy that seems surprising at first. Where a partial institution is already in place, says the guidance, “the panel may issue an order supplementing the institution decision to institute on all challenges raised in the petition.”80 Now that PTAB institution is a binary, all-or-nothing matter, the PTAB may retroactively institute the rest of the petition. What makes this approach surprising is its discretion.

After SAS Institute, it might seem that the PTAB must institute the additional challenged claims, but this is not so for two reasons. First, though the guidance does not spell it out, the PTAB may also deny the entire petition, including what it had previously instituted, and terminate the proceeding. Retroactively granting-all and retroactively denying-all are equally permissible because of the “may not/unless” structure of the institution standard. As discussed above, this framing makes a reasonable likelihood of success a necessary condition.81SAS Institute now makes clear that it is not a sufficient one. Once the standard is met, whether or not to institute is discretionary: “The text says only that the Director can decide ‘whether’ to institute the requested review—not ‘whether and to what extent’ review should proceed.”82 The Court also notes that “the language anticipates a regime where a reasonable prospect of success on a single claim justifies review of all.”83 The language is telling. Under these circumstances, review is justified but not compelled.

Moreover, the parties themselves may render further PTAB action moot. A joint request to terminate the case has multiple benefits. The settling petitioner in such a case can escape without estoppel attaching.84 In exchange, the patent owner can likely end the inter partes review altogether. Such a bargain is even more likely in cases where the PTAB has not yet reached the merits or sunk significant decision costs into the case.

To be sure, the PTAB can proceed to a final written decision even if all petitioners have settled and dropped out.85 However, in this type of situation above all, it makes little sense for the PTAB to disregard party settlement and force itself forward to a final written decision. If it did so, the panel would have to adjudicate claims on which it had previously denied institution, expecting at the time that it would not have to issue any decision at all. Indeed, that was precisely the agency’s litigation position in SAS Institute itself.86 Quiet resolution of these cases is likely to be attractive not only to the parties but also to the PTAB.

D.     Litigant Incentives

Finally, while currently pending cases are resolved, potential petitioners face an ongoing choice about whether to mount new challenges and, if so, how to do it. For them, SAS Institute raises the stakes considerably. To be clear, these higher stakes are separate from cost constraints, including the PTAB fee increase that took effect in late 2017, before the SAS Institute decision was even handed down.87 They are also separate from more longstanding constraints such as page limits on petitions, which are difficult to circumvent only as far as they impose the additional cost of filing more petitions in order to make all desired arguments.88

Before SAS Institute, petitioners faced little reason not to be overinclusive in their challenges. Full denials were possible, but the routine usage of partial institution meant that the risk-reward calculus could be split. Meritorious arguments could go forward on their own, and failed arguments were no real loss because the estoppel effects were murky at best. Going forward, however, even petitions that contain meritorious arguments might fail in their entirety.

After SAS Institute, the PTAB faces the responsibility of writing a final written decision as to every claim challenged and has additionally assumed responsibility for every challenge in the petition.89 The Federal Circuit’s increased interrogation of agency reasoning means that these responsibilities also require more thorough explanations, under Chenery, in the PTAB’s final written decisions.90 As a result, petitioners would do well to focus their petitions. Whether in the context of choosing which patent claims to challenge, which statutory grounds to assert, and which prior art to cite, PTAB petitioners must now seek a difficult balance that the Patent Office never required them to seek before. Petitioners must now try to ensure that the likelihood of full institution is greater than the likelihood of full denial.

IV.     Toward More Efficient PTAB Administration

In this changed landscape, the likely incentive of the Patent Office is to deny institution more often, at least for the time being. The Court’s opinion has no effect on the PTAB’s ability to grant full institutions. PTAB panels could already do so and still can. What they now confront is the prospect of granting full institution even where arguments in the petition lack merit. It is no longer possible to filter out these potentially unavailing arguments at the outset, where estoppel would at least arguably not attach. Instead, the only way to entertain any arguments is to adjudicate all arguments—even the plainly deficient ones—with all the Chenery obligations of reason-giving and the full scope of estoppel that it imposes on the petitioner. This requirement represents a potentially significant increase in the PTAB’s workload.

A.     Caseload Effects of Partial Institution

The magnitude of this increased workload merits closer scrutiny. Until now, the Patent Office has largely based its estimates on the number of pending cases (i.e., the number of petitions filed).91 Looking with more granularity and detail at the content of those petitions, however, reveals just how much unappreciated filtering work partial institution has been doing until now. This Section presents the results of an empirical study of each of the nearly 6,000 inter partes review petitions in which the PTAB had rendered institution decisions starting from the beginning of AIA review through April 2018—when the SAS Institute decision ended partial institution.92

As an initial matter, the agency’s reliance on case-level measures has led it to understate the scope of partial institution. At the level of case petitions filed, the total shares of petitions that were fully instituted, partially instituted, and fully denied through April 2018, were 39.2%, 25.8%, and 35.0%, respectively. By this finding, partial institution seems to have accounted for only a quarter of the PTAB’s workload. As demonstrated by Figures 1 and 2, looking at this measure over time tells an even more modest story, that petitions in which the PTAB made partial institutions accounted for a steadily declining share, falling to about 19% by April 2018. In other words, even before SAS Institute came along and ended the practice, partial institution was apparently already on the decline.


Figure 1. Shares of Inter Partes Review Petitions Fully Instituted, Partially Instituted, and Fully Denied (through April 2018)




Figure 2. Shares of Cases Fully Instituted, Partially Instituted, and Fully Denied (retrospective 6-month average)


Given the claim-by-claim approach that the PTAB itself advocated in defense of partial institution for so long, it would seem appropriate to parse workload more finely, at least at the level of claims challenged within a case. For example, a case involving thirty challenged patent claims on which institution was fully granted presents quite a different workload than a case involving only ten challenged claims on which institution was only partly granted.

The grounds on which patent claims are challenged also matter, especially in light of the Federal Circuit’s renewed push to elicit fuller and more reasoned decision making in PTAB cases.93 An argument about the novelty requirement for patentability generally rests on a single prior art reference,94 whereas an argument about the nonobviousness requirement95 generally rests on a combination of multiple prior art references.96 These, too, present quite different workloads for a panel of PTAB judges.

Thus, in empirically exploring the workload associated with PTAB cases and with the work of partial institution as a filter, it is appropriate under the circumstances to take as the unit of observation each combination of patent claim and asserted legal ground for unpatentability. For example, an inter partes review petition challenging ten patent claims each on novelty and nonobviousness grounds represent twenty claim-ground pairs.

This more detailed definition of workload results reveals, surprisingly, that the use of partial institution prior to SAS Institute had been cutting the agency’s workload substantially. Figures 3 and 4 demonstrate that out of all the claim-ground pairs that came before the PTAB in inter partes review petitions on which the PTAB made a partial institution decision until April 2018, the PTAB accepted 60.8% and rejected the remaining 39.2% as unmeritorious. This finding is also consistent across technologies.


Figure 3. Share of Claim-Ground Pairs Instituted versus Denied in Partially Instituted Cases (Overall, Dec. 2012–Apr. 2018)



Figure 4. Share of Claim-Ground Pairs Instituted versus Denied in Partially Instituted Cases, by Technology (Overall, Dec. 2012–Apr. 2018)


The reality that partial institution was cutting the PTAB’s workload is also strikingly stable over time. Indeed, the PTAB reached this relative equilibrium almost immediately. Institution decisions began issuing in December 2012, some three months after AIA reviews came into effect.97 The six-month moving average rate of institution for claim-ground pairs, starting in June 2013, was about 62.8%, corresponding to about 37.2% of the PTAB’s workload being filtered out at the institution stage. Since then, the six-month average share of the workload that was being filtered has occasionally risen and fallen, as low as 31.1% in April 2014 and as high as 45.7% in July 2016, but has mostly stayed close to 40% ever since, as demonstrated by Figure 5.

Figure 5. Share of Claim-Ground Pairs Instituted versus Denied in Partially Instituted Cases, over Time (retrospective 6-month average)


These findings show that, even as the share of petitions that were partially instituted has been declining, the share of claim-ground pairs that are being filtered out has remained persistent. As a result, measuring rates of institution by evaluating PTAB cases as a whole rather than claim-ground pairs within each case distorts the observed workload of the PTAB in AIA reviews. When looking at cases as a whole, two common measures for institution are (1) the share of cases in which at least some institution was made and (2) the share of cases in which at least some institution was denied. It might seem intuitive that the “true” rate of institution lies somewhere between the bounds of these two case-based measures. As demonstrated by Figure 6, however, measuring by claim-ground pairs reveals that the rate at which the PTAB grants institution is systematically lower than both case-based measures.

 Figure 6. Estimating Partial Institution by Case vs. Claim-Ground Pair


In fact, this tendency toward lower institution rates and correspondingly greater reductions in workload appears to be a natural trend in the PTAB. When the volume of challenges in a given time period is relatively low—say, less than 1,000 claim-ground pairs asserted in petitions within a six-month window—the rate of institution varies widely both across and within technologies. As demonstrated in Figure 7, for petitions on “Chemical”-related patents, the institution rate may be higher than 95% whereas for petitions on “Mechanical”-related patents, the rate may be as low as 20%. However, as the volume of challenges increases—between 1,000 and 2,000 claim-ground pairs in a six-month window—the range of variation decreases to institution rates between 30% and 80%. In this range are most petitions on “Drugs & Medical”- and “Electrical”-related patents. And finally, as the volume increases to 3,000 or more claim-ground pairs within a six-month window, the range of variation decreases to institution rates between 40% and 60%. In this range, all petitions are on “Computers & Communications”-related patents.


Figure 7. Share of Claim-Ground Pairs Instituted versus Denied within Successive 6-Month Periods, by Technology



The key lessons of this empirical study are that partial institution’s actual effect on the PTAB’s workload has been to cut it substantially, and that as petition volume rises within a given time frame, the corresponding tendency of the PTAB has been to filter its workload more aggressively. Caution is appropriate, therefore, in responding to calls that the PTAB should take cases that might have been partial institutions and presumptively favor making full institutions, whether in currently pending cases or in future petitions. As the argument goes, the PTAB in such cases has already found an adequate likelihood of success as to at least one claim and so should grant full institution. To do otherwise would recognize a deficient patent and consciously do nothing about it.

However, a policy of full institution could immediately and dramatically raise the PTAB’s workload, bringing into review all of the claim-ground pairs that had previously been denied institution in cases where partial institution was made. A strong preference for institution (rather than a per se policy) would raise similar concerns; only the magnitude of the problem would differ. Calls for full, or near-full, institution ignore the flip side of the Court’s actual holding in SAS Institute—the PTAB must institute fully or it must deny fully.98

B.     Incentives of the Agency

To deny fully can be a valuable action, and not just because it properly rejects petitions that consist largely of unmeritorious arguments. The use of denials also invites better filing in the future. In fact, full denial of petitions was the very workaround proposed in Justice Ginsburg’s dissent. That opinion expressly contemplates full denial of a petition, except that the PTAB in its decision to deny institution would also identify which claims were worthy of review and which claims were not.99 Petitioners could then refile in light of that advisory information. Justice Ginsburg described this exercise as the PTAB “spend[ing] its time so uselessly”100—in contrast simply to allowing partial institutions and reaching the same point without the added step of refiling.

This is not only a reasonable idea but also one that comports with agency obligations of reasoned decision making.101 Petitioners are now obliged to pare down their challenges in order to balance the hope of full institution against the risk of full denial, but they do not make this choice in a vacuum. The PTAB also has a valuable—indeed, necessary—role to play.

Decisions to grant review already explain what arguments the panel has found meritorious in order that the parties, especially the patent owner, have clear and ample notice of the issues to be litigated.102 Now after SAS Institute, the PTAB can do the same to explain in its denials of institution precisely what it finds worthy or unworthy of review, and why.103 By channeling petitioners “to file new or amended petitions shorn of challenges the Board finds unworthy,”104 the PTAB can generate both static and dynamic benefits. In cases already filed, these reasoned denials would be a roadmap for petitioners to craft petitions without undue guesswork.

In cases to come, other petitioners would also have had the benefit of earlier decisions, enabling them to file challenges that are more likely in the first place to be fully granted than fully denied. In particular, the realigned incentive of the PTAB to reject overinclusive and repetitive petitioning would do much to discourage those behaviors among petitioners. These were especially pernicious habits that partial institution cultivated in petitioners for the simple reason that the PTAB had a way to manage its workload without having to discipline extravagant petitioners. That is no longer the case. Now when the PTAB allows a petitioner to bring a patent owner into AIA review, the defensive burden that the patent owner faces is more closely connected to the workload than the PTAB itself faces.

C.     Upcoming Questions for the Federal Circuit

The Court’s mandate in SAS Institute raises important procedural questions about how to reconcile administrative adjudication of patent validity with the need for binary institution decisions. Having rejected the expansive Patent Office claim to nonappealable discretion,105 however, the Court has also endorsed a jurisdictional rule of vigorous appellate oversight of the PTAB.106 This is entirely appropriate. The stable and predictable substitution of expert agency adjudication for slower and costlier district court litigation is more likely to arise if the procedural answers to post-SAS Institute questions bear the imprimatur of the Federal Circuit and are not merely the transient policy preferences of the Patent Office.

Two questions are especially likely to command prompt attention. One is the scope of the one-year time bar, which compels district-court defendants in a patent infringement case to seek inter partes review within a year or forgo it altogether.107 Another is the decision of a district court to stay its hand pending resolution of an inter partes review or other PTAB proceeding on the same patent.108

1.     Inter Partes Review’s One-Year Time Bar

Litigants are likely to dispute the scope of the one-year time bar particularly soon in the wake of SAS Institute because a denial of institution may not leave the petitioner enough time to refile within the one-year deadline.109 The question then is what happens when a party timely files its petition in the first place, the PTAB plans to deny review, and the one-year deadline has passed in the meantime. Can, and should, the PTAB in such cases allow the petitioner to refile notwithstanding the time bar? The question is unsettled, but the answer is probably no.

The reason why this issue is likely to arise lies in the power-allocation function of the one-year time bar, which is an important structural border provision in the AIA’s substitution of agency adjudication for judicial process.110 The one-year clock begins when a party (or its privy or real party in interest) “is served with a complaint alleging infringement of the patent.”111 Prior research confirms that, when measuring the lag from prior district-court litigation to an eventual inter partes review petition on the same patent, the large majority do fall within a one-year window.112 As Figures 8 and 9 demonstrate, for the lag between first district-court case to first petition, 76.6% of patents come within the one-year window.113 For the lag between last pre-PTAB district-court case to first petition, the share is 88.6%.114 Most importantly, a large share of parties seeking inter partes review wait as long as possible, filing right at the one-year mark.115


Figure 8. Lag from First District-Court Case to First Inter Partes Review Petition, in Years116


Figure 9. Lag from Last Pre-Inter Partes Review District Court Case to First Inter Partes Review Petition117



As for how the issue might be resolved, there are at least two colorable theories on which petitioners might seek leave to refile notwithstanding the time bar, and why the PTAB might think itself empowered to entertain such requests. However, neither is ultimately persuasive. The first, more intuitive theory is that the time bar is like a statute of limitations and might be subject to tolling. This theory fails because of the general rule that a limitation period can be tolled within a court’s equitable discretion only if the time limit is non-jurisdictional in nature.118 Failure to comply with a jurisdictional time bar, on the other hand, “deprives a court of all authority to hear a case.”119 The Federal Circuit’s en banc decision in Wi-Fi One v. Broadcom suggests that the one-year time bar for inter partes review is, indeed, jurisdictional in nature.120 As the en banc majority explained, satisfying the time bar “is a condition precedent to the Director’s authority to act”121 and “sets limits on the Director’s statutory authority to institute, balancing various public interests.”122 Judge O’Malley, writing separately, put the point even more plainly, framing the time bar as a matter of the PTAB’s “statutory jurisdiction.”123

It is also instructive to consider how the law treats the initial district-court complaint that starts the one-year clock itself. In Click-to-Call v. Ingenio, the Federal Circuit recently held that the voluntary dismissal of a prior district-court complaint does not reset the one-year clock for filing a petition seeking inter partes review.124 Prior case law had hinted that such dismissals might reset the clock, recognizing “the effect of dismissals without prejudice as leaving the parties as though the action had never been brought.”125 An en banc majority of the court, however, agreed that leaving the parties as if the intervening action had never been brought is a reason to apply the time bar as is—not to toll it indefinitely.126 By a similar logic, if a denial of institution is viewed as equivalent to a dismissal that leaves the parties as though the case had never been brought, then presumably the time bar should apply as is.

The second theory fares no better. Rather than denying institution through an outright dismissal, the PTAB might grant the petitioner leave to amend its petition, as courts often do when facing motions to dismiss.127 Allowing amendments to a petition would spare the PTAB the need to dismiss an entire case, including meritorious arguments, where the one-year time bar would forbid refiling the case “shorn of challenges the Board finds unworthy.”128 This would certainly show solicitude for petitioners who have invested in testing the validity of patents, a policy position that the Patent Office might plausibly adopt. It would also mirror the modern judicial policy to “freely give leave when justice so requires,”129 a policy whose justifications include solicitude for certain favored categories of litigants, such as pro se plaintiffs.130

Meanwhile, the potential burden to patent owners of having to defend their patent rights against later-raised arguments is not necessarily an insurmountable obstacle. Just as amendments in civil litigation may come after a statute of limitations but can nevertheless relate back to the date of the original pleading in certain circumstances,131 amendments to an inter partes review petition could be conditioned on finding that the original petition provided enough notice to the patent owner so as not to inflict undue prejudice.132

Ultimately, the risk in allowing petitions to be amended as a way around the one-year time bar is that recent case law is unclear whether the PTAB has the authority to do it. Unlike the liberal joinder rules of federal civil litigation that generally allow “as many claims as [a party] has” to be introduced into a pleading,133 the adjudicatory powers of the Patent Office are enumerated by statute and were consciously designed as a zero-sum substitute for the power of district courts to adjudicate patent validity.134 The agency cannot enlarge its domain to balance fairness and prejudice between petitioners and patent owners without intruding upon the domain of the courts—indeed, intruding upon the allocation of power that Congress already put in place through structural border provisions such as the one-year time bar itself.135 The Federal Circuit’s en banc decision in Wi-Fi One v. Broadcom recognized the gravity of this balance and forcefully rejected Patent Office attempts to reallocate power to itself at the expense of the courts.136 That said, the issues at stake in Wi-Fi One were whether the Federal Circuit could review the PTAB’s application of the one-year time bar and whether the PTAB’s application was correct with regard to real parties in interest. For the distinct question of how the PTAB applies the time bar to amended petitions, the Federal Circuit would almost surely find judicial reviewability again but might view the correctness of the PTAB’s position differently.

Still, in all, the Court’s decision in SAS Institute counsels against an amended-petition workaround to the one-year time bar. Fundamentally, the encouragement of liberal amendment137 and liberal joinder138 in civil litigation are rooted in the modern preference for a unitary form of action without procedural distinctions or different modes of seeking relief.139 In other words, if two civil actions that might otherwise be filed separately are related enough to be litigated together, they should be.140 Importantly, this does not give courts jurisdiction that they would not otherwise have, but merely encourages the exercise of existing powers in ways that “secure the just, speedy, and inexpensive determination of every action and proceeding.”141

Thus, even if this analytical framework were to be applied to the PTAB to permit amending petitions and joining arguments merely as a way of consolidating multiple cases that would have been brought instead, the analogy would still fail. The Court in SAS Institute already expressly ejected the Patent Office’s attempt to curate the contents of a petition in the PTAB.142 Moreover, the text of the AIA, which specifically provides for joinder of multiple inter partes reviews by the same or different parties, limits such joinders only to petitions that are each independently determined to warrant institution.143 The agency simply does not have the broad adjudicatory discretion that courts have in civil litigation.

The upshot is that petitioners who file close to the one-year deadline and are denied institution after the deadline has passed are likely to be time-barred from refiling, as are their privies and real parties in interest. Nevertheless, others may learn from the PTAB’s opinion explaining the denial, and this remains a valuable reason for the agency to engage in reasoned decision-making.

2.     Judicial Stays Pending PTAB Review

Beyond the one-year time bar, the end of partial institution is also likely to push a renewed interest in seeking stays of district-court litigation pending PTAB review, especially inter partes review. Most of this interest will naturally be directed to district courts themselves, which “have broad discretion to manage their dockets, including the power to grant a stay of proceedings.”144 The end of partial institution after SAS Institute, however, will especially affect two of the factors that district courts weigh in deciding whether to stay a patent case pending PTAB review of the same patent. This impact, in turn, is likely to invite Federal Circuit oversight on whether district courts are taking the change in circumstances after SAS Institute into account enough not to constitute an abuse of discretion.

In weighing a stay pending inter partes review, district courts generally seek to simplify the issues involved, to avoid letting costs already sunk into the litigation before them go to waste, and to minimize prejudice to the non-moving party.145 Indeed, together with a fourth factor—“reduc[ing] the burden of litigation on the parties and on the court”—this standard is actually required to be considered where the co-pending PTAB proceeding is a covered business method review.146 As a result, this fourth factor has begun to inform district court stays for inter partes review as well.147 The Federal Circuit, for its part, approves this cross-fertilization as being well within the sound discretion of district courts.148

Partial institution, by its very nature, made this flexible, discretionary standard hard to apply. Where some patent claims and arguments were admitted for review and others were excluded, the degree to which a PTAB judgment would simplify litigation would necessarily be incremental. This is over and above the structural limits on how much inter partes review can overlap with Article III litigation, as inter partes review can consider legal arguments only about anticipation and obviousness and prior art evidence consisting only of patents and printed publications.149

It is little surprise, then, that district courts around the country grant stays with high variability, more than 70% in the Northern District of California and less than 40% in the Eastern District of Texas.150 As with estoppel, partial institution took an existing problem with PTAB review and made it even worse.151 In fact, growing awareness of this variability—and apparently growing uncertainty about how useful PTAB review would prove to be to courts—has also coincided with an overall decrease in the granting of stays, from about 80% over the 2011–2014 period to about 70% over the 2011–2016 period.152 Now that institution decisions must fully grant or fully deny a petition, both inter partes review and covered business method review suddenly represent a far clearer signal from the PTAB about what issues stand to be simplified if the agency grants review, and vice-versa.

Moreover, beyond the simplification of issues, a court’s evaluation of costs already sunk into litigation, through discovery and the setting of a trial date, is also likely to be affected by SAS Institute. Where the PTAB institutes a petition in full, even a case that is somewhat far down the litigation timeline may still benefit from a stay because the savings from a likely invalidation in the PTAB could be greater than the additional judicial resources that would be spent if the court declined a stay.153 The PTAB can do much to strengthen the quality of this informational signal from the agency to the courts by explaining its reasoning in decisions to grant or deny institution even more fully than it already does.154

The propriety of how district courts use this clearer information from the agency about simplification and sunk litigation cost could come before the Federal Circuit in two ways. One is ordinary appeal following a final judgment, which is naturally likely to follow a decision to deny, rather than grant, a stay. The other is interlocutory appeal immediately upon the grant or denial of a stay. Interlocutory review is a meaningful possibility in the face of a co-pending PTAB case. For covered business method reviews, Congress expressly provided such immediate appeal, with de novo oversight from the Federal Circuit “to ensure consistent application of established precedent.”155

For inter partes review, meanwhile, interlocutory jurisdiction is available where the district court’s decision about a stay materially affects some other issue over which interlocutory appeal would otherwise be available, such as a motion for preliminary injunction.156 In Procter & Gamble v. Kraft Foods, for example, the district court granted alleged infringer Kraft Foods’s motion to stay the infringement litigation pending inter partes reexamination of the patent that Procter & Gamble had asserted.157 Procter & Gamble had sought a preliminary injunction, which the district court denied as moot in light of its stay, and the Federal Circuit concluded that it had interlocutory jurisdiction.158 Thus, for a party who loses in district court over whether a stay should have been issued pending inter partes review, it is paramount to be able to show that the decision would have “serious, perhaps irreparable, consequence” or “cannot effectually be reviewed after the trial.”159

In both contexts, post-judgment appeal as well as interlocutory appeal, the Federal Circuit will face essentially the same question: whether the district court has taken adequate account of the dramatic change in Patent Office practice following SAS Institute. Agency decisions to grant review should, at the margin, correspond to a greater likelihood of a judicial decision to grant a stay, and vice-versa. It would be appropriate, then, for Federal Circuit review to take this expectation as a baseline for evaluating the district court’s exercise of discretion, though particular case facts might always justify a departure.

Similarly, as to the expenditure and timeline of litigation costs for weighing a stay, an institution decision generally comes six months after filing, with three months for the patent owner’s preliminary response and another three months for the PTAB to determine,160 based on the petition and the response, whether review is appropriate.161 In light of this relatively stable agency schedule, the Federal Circuit could promote the substitutionary efficiency of the AIA by rewarding, with affirmance, the stay decisions of district courts that not only account for where the litigation calendar was at the time but also set the litigation calendar with an eye toward concurrent PTAB review in the first place.162 Indeed, bellwether courts such as the Eastern District of Texas, the Northern District of California, and the District of Delaware might even cultivate a norm of calendaring patent cases to conserve initial judicial resources where a PTAB petition is likely to be filed.

V.     Conclusion

Although the Court’s decision in SAS Institute has undoubtedly upended a cornerstone of PTAB administration, the overall effect on administrative patent revocation is likely to be greater efficiency, in accordance with the original intent of the AIA to substitute agency review for judicial review. The initial Patent Office response has been sound, going further than the Court’s opinion might strictly have required, but the agency should exercise caution as it proceeds. The Patent Office seems to have underappreciated the sheer scale of the adjudicatory workload that partial institution was allowing it to filter out. As a result, the stakes now of accepting petitions too liberally are correspondingly high.

Ultimately, the virtue of ending partial institution is that it imposes greater discipline both upon patent challengers and upon the agency that evaluates their challenges. The PTAB’s own workload is now directly connected to the burdens that petitioners are allowed to impose upon patent owners. Moreover, if the Patent Office takes Justice Ginsburg’s mock proposal seriously, as it should, the result will be a desirable increase in reasoned decision-making by the PTAB from which current and future petitioners alike can learn much about how to focus their arguments for success.

The opinion of the Court quite consciously concerned itself with respecting the structural choices that Congress made in the AIA, for better or worse, without interposing the Court’s own policy judgments about efficiency. Nevertheless, and perhaps ironically, the SAS Institute decision will promote efficiency after all.


  1. [1]. See Saurabh Vishnubhakat et al., Strategic Decision Making in Dual PTAB and District Court Proceedings, 31 Berkeley Tech. L.J. 45, 50–64 (2016) (summarizing the history of ex post administrative review prior to, and under, the AIA).

  2. [2]. Id. at 64–81 (discussing court-agency substitution through AIA proceedings).

  3. [3]. See 35 U.S.C. § 311(a) (2012); Consumer Watchdog v. Wis. Alumni Research Found., 753 F.3d 1258, 1262 (Fed. Cir. 2014).

  4. [4]. Compare 35 U.S.C. § 316(e) (allowing patents to be revoked after an inter partes review proceeding upon a showing of preponderance of the evidence), with Microsoft Corp. v. i4i Ltd. P’ship, 564 U.S. 91, 95 (2011) (holding that asserting patent invalidity as a defense to infringement requires clear and convincing evidence).

  5. [5]. See 37 C.F.R. § 42.100(b) (2012); Cuozzo Speed Techs., LLC v. Lee, 136 S. Ct. 2131, 2136 (2016). The Patent Office in late 2018 changed this broader standard for claim construction and aligned it with the interpretive approach that governs in infringement litigation. See 37 C.F.R. § 42.100(b) (2018).

  6. [6]. See generally Oil States Energy Servs., LLC v. Greene’s Energy Grp., LLC, 138 S. Ct. 1365 (2018) (questioning whether the PTAB has the authority to revoke a patent through inter partes review); see also generally MCM Portfolio LLC v. Hewlett–Packard Co., 812 F.3d 1284 (Fed. Cir. 2015) (rejecting the argument that inter partes review violates the Constitution), cert. denied, 137 S. Ct. 292 (2016).

  7. [7]. Oil States, 138 S. Ct. at 1379.

  8. [8]. 35 U.S.C. § 6 (including reexaminations, derivation proceedings, and inter partes reviews).

  9. [9]. SAS Inst. Inc. v. Iancu, 138 S. Ct. 1348, 1359–60 (2018).

  10. [10]. 35 U.S.C. §§ 316(a)(11), 326(a)(11).

  11. [11]. SAS Inst., 138 S. Ct. at 1360 (Ginsburg, J., dissenting).

  12. [12]. Id. (Breyer, J., dissenting).

  13. [13]. Id. (Ginsburg, J., dissenting).

  14. [14]. Vishnubhakat et al., supra note 1, at 47–50.

  15. [15]. Id. at 58–59.

  16. [16]. 35 U.S.C. §§ 315(e), 325(e) (2012).

  17. [17]. Id.

  18. [18]. Id. § 325(e); Leahy–Smith America Invents Act, Pub. L. No. 112-29, § 18(a)(1)(A), 125 Stat. 283, 329 (2011).

  19. [19]. 35 U.S.C. § 325(e); Leahy–Smith America Invents Act § 18(a)(1)(A).

  20. [20]. Restatement (Second) of Judgments § 17 (Am. Law Inst. 1982).

  21. [21]. Id. § 24(1).

  22. [22]. Id. § 24(2).

  23. [23]. Id. §§ 34, 41.

  24. [24]. Id. § 17(3); see also id. § 17 cmt. c.

  25. [25]. Id. § 27.

  26. [26]. Id.

  27. [27]. The potential for confusion over whether PTAB estoppel is a species of res judicata or of collateral estoppel was the subject of a recent panel presentation by Megan La Belle. See Project on the Foundations of Private Law, The Administrative-Private Law Interface in IP Conference: Panel 3: The Limits of Preclusion and Deference, YouTube (Apr. 5, 2018),
    watch?v=J-Nb9zKDqr0 (beginning at 13:50, remarks of Megan La Belle).

  28. [28]. This use of “claim” is specific to patents, which contain a series of individual statements about particular embodiments of the invention that the patent covers. See 35 U.S.C. § 112(b) (2012). It is not synonymous with “claim” in the procedural sense of litigation.

  29. [29]. Id. §§ 315(e)(1), 325(e)(1).

  30. [30]. Id. §§ 315(e)(2), 325(e)(2).

  31. [31]. Id. §§ 315(e), 325(e).

  32. [32]. Id. § 321(c); Leahy–Smith America Invents Act, Pub. L. No. 112-29, §§ 3(n)(1), 6(f)(2)(A), 125 Stat. 284, 293, 311 (2011).

  33. [33]. Leahy–Smith America Invents Act § 18(d)(1).

  34. [34]. 35 U.S.C. § 311(b).

  35. [35]. Id. §§ 311(b), 321.

  36. [36]. Id. § 282(b)–(c).

  37. [37]. See, e.g., Shaw Indus. Grp., Inc. v. Automated Creel Sys., Inc., 817 F.3d 1293, 1305 (Fed. Cir. 2016) (“[T]he lack of a reasoned basis [in a PTO decision] deprive[d] [any] future tribunals of the necessary basis to determine whether estoppel should apply.”).

  38. [38]. See Saurabh Vishnubhakat, The Non-Doctrine of Redundancy, 33 Berkeley Tech. L.J. 777, 792–93 (2019) (discussing Judge Reyna’s departure from the panel majority in Shaw Industries and the competing analysis that his concurring opinion set forth).

  39. [39]. SAS Inst. Inc. v. Iancu, 138 S. Ct. 1348, 1354 (2018).

  40. [40]. Id.

  41. [41]. 35 U.S.C. § 314(d); Cuozzo Speed Techs., LLC v. Lee, 136 S. Ct. 2131, 2136 (2016).

  42. [42]. SAS Inst., 138 S. Ct. at 1354.

  43. [43]. Id. (quoting 35 U.S.C. § 318(a) (emphasis added)).

  44. [44]. Id.

  45. [45]. Id. at 1355.

  46. [46]. Brief for the Federal Respondent at 12–13, SAS Inst., 138 S. Ct. 1348 (No. 16-969), 2017 WL 3948437, at *12–13.

  47. [47]. See generally Saurabh Vishnubhakat, The Porous Court-Agency Border in Patent Law, 51 Akron L. Rev. 1069 (2017) (providing detailed criticism of progressively broader Patent Office interpretations of the PTAB nonappealability provisions, including and especially the agency’s position in SAS Institute).

  48. [48]. 35 U.S.C. §§ 314(d), 324(e) (emphasis added).

  49. [49]. Id. § 314(a).

  50. [50]. Id. § 324(a).

  51. [51]. Id. §§ 314(a), 324(a).

  52. [52]. 37 C.F.R. §§ 42.108, 42.208 (2018).

  53. [53]. SAS Inst. Inc. v. Iancu, 138 S. Ct. 1348, 1355–56 (2018).

  54. [54]. Id. at 1356.

  55. [55]. Id. at 1355.

  56. [56]. Id. at 1359–60.

  57. [57]. Id. at 1359.

  58. [58]. See infra Section IV.C.

  59. [59]. SAS Inst., 138 S. Ct. at 1360–65 (Breyer, J., dissenting).

  60. [60]. Id. at 1360 (Ginsburg, J., dissenting).

  61. [61]. Id.

  62. [62]. See infra Section IV.B.

  63. [63]. The discussion throughout the rest of this Part draws heavily from a blog commentary just after SAS Institute was decided. See Saurabh Vishnubhakat, First Steps After SAS Institute, Patently-O (Apr. 27, 2018),

  64. [64]. Id.

  65. [65]. Id.

  66. [66]. Id.

  67. [67]. Guidance on the Impact of SAS on AIA Trial Proceedings, USPTO (Apr. 26, 2018) [hereinafter PTAB Guidance],
    of_sas_on_aia_trial_proceedings_%20(april_26,_2018).pdf [].

  68. [68]. Id.

  69. [69]. Id.

  70. [70]. SAS Inst. Inc. v. Iancu, 138 S. Ct. 1348, 1352–53 (2018).

  71. [71]. PTAB Guidance, supra note 67.

  72. [72]. 35 U.S.C. § 311(b) (2012).

  73. [73]. 37 C.F.R. § 42.108 (2018) (emphasis added). The regulation governing institution of post-grant review and covered business method review corresponds fully to that of inter partes review. See id. § 42.208.

  74. [74]. The initial Federal Circuit response to this agency caution seems positive. In PGS Geophysical AS v. Iancu, a nonprecedential panel decision held that SAS Institute requires the patent claims being challenged and the grounds for the challenge to be treated equally for institution purposes. PGS Geophysical AS v. Iancu, 891 F.3d 1354, 1360 (Fed. Cir. 2018). A month later in Adidas AG v. Nike, Inc., the Federal Circuit reiterated that conclusion from PGS Geophysical AS, this time in a precedential opinion. Adidas AG v. Nike, Inc., 894 F.3d 1256, 1258 (Fed. Cir. 2018).

  75. [75]. Motor Vehicle Mfrs. Ass’n v. State Farm Mut. Auto. Ins. Co., 463 U.S. 29, 57 (1983).

  76. [76]. Richard E. Levy & Robert L. Glicksman, Agency-Specific Precedents, 89 Tex. L. Rev. 499, 529–30 (2011).

  77. [77]. See Arista Networks, Inc. v. Cisco Sys., Inc., 726 F. App’x 787, 788 (Fed. Cir. 2018); Rovalma, S.A. v. Bohler–Edelstahl GMBH & Co. KG, 856 F.3d 1019, 1021 (Fed. Cir. 2017); Personal Web Techs., LLC v. Apple, Inc., 848 F.3d 987, 989 (Fed. Cir. 2017); In re Nuvasive, Inc., 842 F.3d 1376, 1379 (Fed. Cir. 2016).

  78. [78]. SEC v. Chenery Corp., 318 U.S. 80, 88 (1943).

  79. [79]. Levy & Glicksman, supra note 76, at 530.

  80. [80]. PTAB Guidance, supra note 67 (emphasis added).

  81. [81]. See supra text accompanying notes 50–52.

  82. [82]. SAS Inst. Inc. v. Iancu, 138 S. Ct. 1348, 1356 (2018) (emphasis in original).

  83. [83]. Id.

  84. [84]. 35 U.S.C. § 317(a) (2012).

  85. [85]. Id.

  86. [86]. Brief for the Federal Respondent, SAS Inst., 138 S. Ct. 1348 (No. 16-969), 2017 WL 3948437, at *30.

  87. [87]. Press Release, USPTO, USPTO Finalizes Revised Patent Fee Schedule (Nov. 17, 2017),

  88. [88]. See 37 C.F.R. § 42.24 (2016).

  89. [89]. PTAB Guidance, supra note 67.

  90. [90]. SEC v. Chenery Corp., 318 U.S. 80, 94–95 (1943).

  91. [91]. Chat with the Chief: SAS Guidance for AIA Trial Proceedings After Supreme Court Decision, USPTO (Apr. 30, 2018, 12:00 PM), (discussed during the question-and-answer session).

  92. [92]. PTAB data was analyzed from Docket Navigator. See Docket Navigator, (last visited Apr. 29, 2019). Although Docket Navigator does not permit republication of its data, researchers can obtain access and replicate this study fairly readily. See generally Vishnubhakat et al., supra note 1, app. A at 84 (discussing data and replication from Docket Navigator in Appendix A).

  93. [93]. See supra text accompanying notes 75–76.

  94. [94]. 35 U.S.C. § 102 (2012).

  95. [95]. Id. § 103.

  96. [96]. Compare Enplas Display Device Corp. v. Seoul Semiconductor Co., 909 F.3d 398, 405 (Fed. Cir. 2018) (discussing single-reference anticipation), with Realtime Data, LLC v. Iancu, 912 F.3d 1368, 1373–74 (Fed. Cir. 2019) (discussing multi-reference obviousness).

  97. [97]. Microsoft Corp. v. Proxyconn, Inc., No. IPR2012-00026 (P.T.A.B. Dec. 12, 2012), https://

  98. [98]. See SAS Inst. Inc. v. Iancu, 138 S. Ct. 1348, 1355–56 (2018).

  99. [99]. Id. at 1360 (Ginsburg, J., dissenting).

  100. [100]. Id.

  101. [101]. See supra text accompanying notes 74–81.

  102. [102]. See Office Patent Trial Practice Guide, 77 Fed. Reg. 48,756, 48,757 (Aug. 14, 2012) (to be codified at 37 C.F.R. pt. 42). From the beginning, the PTAB has made a practice of setting forth with specificity in its institution decisions both the scope of review and the reasons for instituting review—and has allowed dissatisfied parties to seek rehearing accordingly. See id. at 48,757, 48,765.

  103. [103]. Under current practice, a decision not to institute review is expected to “contain [only] a short statement as to why the standards [for instituting review] were not met,” and even “this may not be necessary in all cases.” Id. at 48,765.

  104. [104]. SAS Inst., 138 S. Ct. at 1360 (Ginsburg, J., dissenting).

  105. [105]. See supra text accompanying note 57.

  106. [106]. SAS Inst., 138 S. Ct. at 1358–60.

  107. [107]. 35 U.S.C. § 315(b) (2012).

  108. [108]. The power to grant stays is an inherent judicial power in the discretion of the court. See Procter & Gamble Co. v. Kraft Foods Glob., Inc., 549 F.3d 842, 848 (Fed. Cir. 2008).

  109. [109]. Certainly, there would be no immediate estoppel effect of the denial. Although institution decisions are made “final” by statute, 35 U.S.C. § 314(d), estoppel upon both future agency proceedings and future court proceedings arises only when a case “results in a final written decision under section 318(a).” Id. § 315(e). Such final written decisions, in turn, can come only in cases where “inter partes review is instituted and not dismissed.” Id. § 318(a).

  110. [110]. Vishnubhakat, supra note 47, at 1082, 1086–87.

  111. [111]. 35 U.S.C. § 315(b).

  112. [112]. Vishnubhakat et al., supra note 1, at 75–76.

  113. [113]. Id. at 76, 105 fig.17; see infra Figure 8.

  114. [114]. Vishnubhakat et al., supra note 1, at 76, 106 fig.18; see infra Figure 9.

  115. [115]. Vishnubhakat et al., supra note 1, at 76 (describing a modal spike at one year according to both measures). One might ask why any portion of these distributions might exceed the one-year deadline. The answer is that these measures look only at court-agency lags as to the same patent and not necessarily the same parties. Non-standard petitioners who have not been sued in district court at all would not appear in the distribution. Id. at 75.

  116. [116]. Id. at 105 fig.17.

  117. [117]. Id. at 106 fig.18.

  118. [118]. United States v. Kwai Fun Wong, 135 S. Ct. 1625, 1630–31 (2015) (en banc).

  119. [119]. Id. at 1631.

  120. [120]. Wi-Fi One, LLC v. Broadcom Corp., 878 F.3d 1364, 1371–72 (Fed. Cir. 2018).

  121. [121]. Id. at 1380 (Hughes, J., dissenting) (emphasis omitted).

  122. [122]. Id. at 1374 (majority opinion).

  123. [123]. Id. at 1375–76 (O’Malley, J., concurring).

  124. [124]. See Click-To-Call Techs., LP v. Ingenio, Inc., 899 F.3d 1321, 1335, 1337 (Fed. Cir. 2018).

  125. [125]. E.g., Shaw Indus. Grp., Inc. v. Automated Creel Sys., Inc., 817 F.3d 1293, 1301 (Fed. Cir. 2016); see also Bonneville Assocs., Ltd. P’ship v. Barram, 165 F.3d 1360, 1364 (Fed. Cir. 1999). However, because then-controlling precedent held that PTAB determinations about the one-year time bar were nonappealable, the issue remained unresolved. Shaw Indus. Grp., 899 F.3d at 1301 (citing Achates Reference Publ’g, Inc. v. Apple Inc., 803 F.3d 652, 658 (Fed. Cir. 2015)). The en banc opinion in Wi-Fi One reversed Achates and concluded that the time bar is within the Federal Circuit’s appellate jurisdiction. Wi-Fi One, 878 F.3d at 1367. Indeed, this was what precipitated rehearing in Click-To-Call. Click-To-Call Techs., 899 F.3d at 1324–25.

  126. [126]. See Click-To-Call Techs., 899 F.3d at 1335.

  127. [127]. See Fed. R. Civ. P. 12(b)(6).

  128. [128]. SAS Inst. Inc. v. Iancu, 138 S. Ct. 1348, 1360 (2018) (Ginsburg, J., dissenting).

  129. [129]. Fed. R. Civ. P. 15(a)(2).

  130. [130]. See, e.g., Dolan v. Connolly, 794 F.3d 290, 295 (2d Cir. 2015) (remanding a prisoner’s legal challenge because the lower court applied an unnecessarily-stringent standard of pleading to his complaint); El v. Wehling, 548 F. App’x 750, 752–53 (3d Cir. 2013) (remanding a pro se plaintiff’s case to allow him to amend his complaint for a fourth time); Eldridge v. Block, 832 F.2d 1132, 1135–36 (9th Cir. 1987) (observing that Rule 15 “is applied even more liberally to pro se litigants”); cf. Thomas A. Reuland, Note, Rule 15: A Limited Safety Net for 12(B)(6) Dismissal After Iqbal, 96 Iowa L. Rev. 1403, 1409 n.33 (2011) (reading Erickson v. Pardus, 551 U.S. 89, 94 (2007) (per curiam) to “suggest[] that courts must not forget that the limits of notice pleading depend on whether the litigant is pro se”).

  131. [131]. Fed. R. Civ. P. 15(c).

  132. [132]. See generally 61B Am. Jur. 2d Pleading § 804, Westlaw (database updated February 2019) (discussing prejudice to the non-amending party as a factor counseling against amendment and the desirability of allowing amendment when such prejudice is not asserted and proven).

  133. [133]. Fed. R. Civ. P. 18(a).

  134. [134]. Vishnubhakat et al., supra note 1, at 49–50.

  135. [135]. See Vishnubhakat, supra note 47, at 1075–78, 1081–83 (discussing power allocation between the courts and the Patent office through structural border provisions of the AIA); Amicus Curiae Brief of the N.Y. Intellectual Prop. Law Ass’n in Support of Neither Party at 14, Wi-Fi One, LLC v. Broadcom Corp., 878 F.3d 1364 (2018) (Nos. 2015-1944, -1945, -1946), 2017 WL 694463, at *14.

  136. [136]. Wi-Fi One, 878 F.3d at 1374 (noting that, “like § 315 as a whole, [the one-year time bar] governs the relation of IPRs to other proceedings or actions, including actions taken in district court”).

  137. [137]. Fed. R. Civ. P. 15(a).

  138. [138]. Id. R. 18(a).

  139. [139]. See id. R. 2; Advisory Comm. on Rules for Civil Procedure, Proposed Rules of Civil Procedure for the District Courts of the United States 2­–3 (1937),

  140. [140]. This is one of the basic conclusions in the broader procedural debate over what should, as a theoretical matter, constitute the “basic litigation unit.” See, e.g., Michael D. Conway, Comment, Narrowing the Scope of Rule 13(a), 60 U. Chi. L. Rev. 141, 145 n.21 (1993); Douglas D. McFarland, Seeing the Forest for the Trees: The Transaction or Occurrence and the Claim Interlock Civil Procedure, 12 Fla. Coastal L. Rev. 247, 276 (2011).

  141. [141]. Fed. R. Civ. P. 1.

  142. [142]. SAS Inst. Inc. v. Iancu, 138 S. Ct. 1348, 1353 (2018).

  143. [143]. 35 U.S.C. § 315(c) (2012).

  144. [144]. Murata Mach. USA v. Daifuku Co., Ltd., 830 F.3d 1357, 1362 (Fed. Cir. 2016) (quoting Procter & Gamble v. Kraft Foods Glob., Inc., 549 F.3d 842, 848–49 (Fed. Cir. 2008)); Procter & Gamble, 549 F.3d at 848–49 (citing Landis v. N. Am. Co., 299 U.S. 248, 254–55 (1936)); Gould v. Control Laser Corp., 705 F.2d 1340, 1341 (Fed. Cir. 1983).

  145. [145]. See, e.g., Murata, 830 F.3d at 1361.

  146. [146]. Leahy–Smith America Invents Act of 2011, Pub. L. No. 112-29, § 18(b)(1)(D), 125 Stat. 284, 331 (2011) (codified as amended at 35 U.S.C. § 321 (2011)).

  147. [147]. Anne S. Layne-Farrar, The Cost of Doubling Up: An Economic Assessment of Duplication in PTAB Proceedings and Patent Infringement Litigation, A.B.A.: Landslide (May 1, 2018), https://
    june/cost-doubling-up. Recent commentary seems to support this use of the fourth CBM stay factor in inter partes review as a way of advancing the overall substitutionary purpose of the PTAB. See, e.g., Joel Sayres & Julie Wahlstrand, To Stay or Not to Stay Pending IPR? That Should Be a Simpler Question, 17 Chi.-Kent J. Intell. Prop. 52, 63–65 (2018).

  148. [148]. Murata, 830 F.3d at 1362.

  149. [149]. 35 U.S.C. § 311(b) (2012).

  150. [150]. Sayres & Wahlstrand, supra note 147, at 55. There is also additional variation among individual judges. Id. (citing DocketNavigator, Success Rates on Requests to Stay Pending IPR, CBM, or PGR through 2017, DocketReport (Apr. 3, 2018),

  151. [151]. See supra Section II.B.

  152. [152]. Morgan Lewis, 2017 PTAB Digest: The Latest Trends and Developments in Post-Grant Proceedings 26 (2017),
    report/ptab-post-grant-proceedings_fin_screen.ashx; Layne-Farrar, supra note 147 (citing Brian J. Love & Shawn Ambwani, Inter Partes Review: An Early Look at the Numbers, 81 U. Chi. L. Rev. Dialogue 93, 94 (2014)).

  153. [153]. The correlation of a decision to grant PTAB review with a likelihood of invalidating the patent is not only empirically verified but also, of course, built into the statutory criteria themselves, which require “a reasonable likelihood” of petitioner success for inter partes review and a preponderant likelihood of petitioner success for covered business method review. 35 U.S.C. §§ 314(a), 324(a).

  154. [154]. See supra notes 102–04 and accompanying text.

  155. [155]. 35 U.S.C. § 321 note (Transitional Program for Covered Business Method Patents).

  156. [156]. Carson v. Am. Brands, Inc., 450 U.S. 79, 84–85 (1981).

  157. [157]. Procter & Gamble Co. v. Kraft Foods Glob., Inc., 549 F.3d 842, 844 (Fed. Cir. 2008). Inter partes reexamination was an earlier generation of patent validity reevaluation that the Patent Office administered until the AIA replaced it with inter partes review. See Vishnubhakat et al., supra note 1, at 57–58.

  158. [158]. Procter & Gamble, 549 F.3d at 846 (“[T]he stay order, as implemented in this case, can be deemed to have denied P & G’s motion for a preliminary injunction, meets the Carson requirements, and is therefore reviewable under [28 U.S.C.] § 1292(a)(1) and (c)(1).”).

  159. [159]. Id. (quoting Woodard v. Sage Prods., Inc., 818 F.2d 841, 851 (Fed. Cir. 1987)).

  160. [160]. 35 U.S.C. § 313; 37 C.F.R. § 42.107(b) (2018).

  161. [161]. 35 U.S.C. § 314(b).

  162. [162]. The effect of this approach will generally be limited to inter partes review, which is the only AIA validity review that is constrained by a one-year time bar. Compare id. § 315(b) (requiring an inter partes review be filed within one year of the complaint), with id. § 325(a)(1) (showing that post grant reviews are not subject to that same one-year time limit).


Associate Professor, Texas A&M University School of Law; Associate Professor, Texas A&M University Dwight Look College of Engineering; Fellow, Duke Law Center for Innovation Policy; former Expert Advisor, United States Patent and Trademark Office.

The arguments in this writing are the author’s and should not be imputed to the USPTO or to any other organization. Sincere thanks to Bryan Choi, John Duffy, Brian Holland, Dmitry Karshtedt, Jay Kesan, Megan La Belle, Mark Lemley, Adam Kelly, Adam Mossoff, David O’Brien, Lisa Ouellette, Tanya Pierce, Arti Rai, Josh Sarnoff, Stefani Shanberg, Jonathan Stroud, Chris Walker, and participants at the 2018 Iowa Law Review Symposium, the Ohio State Law Faculty Summer Workshop Series, and the 2018 Intellectual Property Scholars Conference for helpful comments.