The Court invokes Chevron U.S.A. Inc. v. Natural Resources Defense Council, Inc., 467 U.S. 837, 104 S.Ct. 2778, 81 L.Ed.2d 694 (1984), and United States v. Mead Corp., 533 U.S. 218, 121 S.Ct. 2164, 150 L.Ed.2d 292 (2001), to resolve one of the questions presented in this case. See ante, at 2133 - 2137, 2142 - 2146. But today's decision does not rest on Chevron 's fiction that ambiguity in a statutory term is best construed as an implicit delegation of power to an administrative agency to determine the bounds of the law. In an appropriate case, this Court should reconsider that fiction of Chevron and its progeny. See Michigan v. EPA, 576 U.S. ----, ----, 135 S.Ct. 2699, 2712, 192 L.Ed.2d 674 (2015) (THOMAS, J., concurring) ("Chevron deference raises serious separation-of-powers questions"); see also Department of Transportation v. Association of American Railroads, 575 U.S. ----, ----, 135 S.Ct. 1225, 1242, 191 L.Ed.2d 153 (2015) (THOMAS, J., concurring in judgment) ("[T]he discretion inherent in executive power does not comprehend the discretion to formulate generally applicable rules of private conduct"); Perez v. Mortgage Bankers Assn., 575 U.S. ----, ---- - ----, 135 S.Ct. 1199, 1217, 191 L.Ed.2d 186 (2015) (THOMAS, J., concurring in judgment) ("Those who ratified the Constitution knew that legal texts would often contain ambiguities.... The judicial power was understood to include the power to resolve these ambiguities over time"); Cass, Is Chevron 's Game Worth the Candle? Burning Interpretation at Both Ends, in Liberty's Nemesis 57-69 (D. Reuter & J. Yoo eds. 2016).
The Court avoids those constitutional concerns today because the provision of the America Invents Act at issue contains an express and clear conferral of authority to the Patent Office to promulgate rules governing its own proceedings. See 35 U.S.C. § 316(a)(4) ; ante, at 2142 -2143. And by asking whether the Patent Office's preferred rule is reasonable, ante, at 2144 - 2146, the Court effectively asks whether the rulemaking was "arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law," in conformity with the Administrative Procedure Act, 5 U.S.C. § 706(2)(A). I therefore join the Court's opinion in full.
Justice ALITO, with whom Justice SOTOMAYOR joins, concurring in part and dissenting in part.
Congress has given the Patent and Trademark Office considerable authority *2149to review and cancel issued patent claims. At the same time, Congress has cabined that power by imposing significant conditions on the Patent Office's institution of patent review proceedings. Unlike the Court, I do not think that Congress intended to shield the Patent Office's compliance-or noncompliance-with these limits from all judicial scrutiny. Rather, consistent with the strong presumption favoring judicial review, Congress required only that judicial review, including of issues bearing on the institution of patent review proceedings, be channeled through an appeal from the agency's final decision. I respectfully dissent from the Court's contrary holding.1
I
In the Leahy-Smith America Invents Act (AIA), 35 U.S.C. § 100 et seq., Congress created three new mechanisms for Patent Office review of issued patent claims-inter partes review, post-grant review, and covered business method patent review (CBM review). This case involves the first of these proceedings, inter partes review.
Under inter partes review, anyone may file a petition challenging the patentability of an issued patent claim at almost any time. §§ 311(a), (c). The grounds for challenge are limited to the patentability of the claim under § 102 (which requires patent claims to be novel) and § 103 (which requires patent claims to be nonobvious). § 311(b).
The statute imposes other restrictions as well. A petition for inter partes review "may be considered only if" the petition satisfies certain requirements, including (as relevant here) that the petition "identif[y], in writing and with particularity, each claim challenged, the grounds on which the challenge to each claim is based, and the evidence that supports the grounds for the challenge to each claim." § 312(a)(3). Additionally, "inter partes review may not be instituted" if the party challenging the patent previously filed a civil action challenging the patent's validity or was sued for infringing the patent more than a year before seeking inter partes review. §§ 315(a)(1), (b). Finally, the Patent Office may not institute inter partes review "unless the Director [of the Patent Office] determines that the information presented in the [challenger's] petition ... and any response [by the patent owner] shows that there is a reasonable likelihood that the petitioner would prevail with respect to at least 1 of the claims challenged in the petition." § 314(a).2
The statute provides that "[t]he determination by the Director whether to institute an inter partes review under this section shall be final and nonappealable." § 314(d). If inter partes review is instituted, the Patent Office conducts a trial that culminates in a "final written decision" on the patentability of the challenged claims. § 318(a). Any patent owner or challenger that is "dissatisfied" with that decision may appeal to the Federal Circuit. § 319.
II
In this case, the Patent Office instituted inter partes review of claims 10 and 14 of Cuozzo's patent based on prior art that the challenger's petition did not cite with respect to those claims. After trial, the Patent Office issued a final written decision holding those claims unpatentable, and Cuozzo appealed that decision to the Federal Circuit. In its appeal, Cuozzo argued (among other things) that the Patent Office had violated the requirement that a *2150petition for inter partes review "may be considered only if" the petition identifies "the grounds on which the challenge to each claim is based, and the evidence that supports the grounds for the challenge," "with particularity." § 312(a)(3).
The Federal Circuit held that it could not entertain this argument because § 314(d) provides that the Patent Office's decision to institute an inter partes review is "final and nonappealable." See In re Cuozzo Speed Technologies, LLC, 793 F.3d 1268, 1273 (2015). This Court now affirms.
I disagree. We have long recognized that "Congress rarely intends to prevent courts from enforcing its directives to federal agencies. For that reason, this Court applies a 'strong presumption' favoring judicial review of administrative action." Mach Mining, LLC v. EEOC, 575 U.S. ----, ----, 135 S.Ct. 1645, 1651, 191 L.Ed.2d 607 (2015) (quoting Bowen v. Michigan Academy of Family Physicians, 476 U.S. 667, 670, 106 S.Ct. 2133, 90 L.Ed.2d 623 (1986) ). While the "presumption is rebuttable," "the agency bears a 'heavy burden' in attempting to show that Congress 'prohibit[ed] all judicial review' of the agency's compliance with a legislative mandate." Mach Mining, supra, at ---- - ----, 135 S.Ct., at 1651 (quoting Dunlop v. Bachowski, 421 U.S. 560, 567, 95 S.Ct. 1851, 44 L.Ed.2d 377 (1975) ). If a provision can reasonably be read to permit judicial review, it should be.
Our decision in Lindahl v. Office of Personnel Management, 470 U.S. 768, 105 S.Ct. 1620, 84 L.Ed.2d 674 (1985), illustrates the power of this presumption. The statute at issue there provided that agency " 'decisions ... concerning [questions of disability and dependency] are final and conclusive and are not subject to review.' " Id., at 771, 105 S.Ct. 1620. The Federal Circuit concluded that the statute cut off all judicial review of such decisions, stating that " '[i]t is difficult to conceive of a more clear-cut statement of congressional intent to preclude review than one in which the concept of finality is thrice repeated in a single sentence.' " Id., at 779, 105 S.Ct. 1620. We reversed. We acknowledged that the statute "plausibly c[ould] be read as imposing an absolute bar to judicial review," but we concluded that "it also quite naturally c[ould] be read as precluding review only of ... factual determinations" underlying the agency's decision, while permitting review of legal questions. Ibid. In light of the presumption of reviewability, we adopted the latter reading. We observed that "when Congress intends to bar judicial review altogether, it typically employs language far more unambiguous and comprehensive," giving as an example a statute that made an agency decision " 'final and conclusive for all purposes and with respect to all questions of law or fact' " and " 'not subject to review by another official of the United States or by a court by mandamus or otherwise.' " Id., at 779-780, and n. 13, 105 S.Ct. 1620.3
This is a far easier case than Lindahl . There is no question that the statute now before us can naturally-perhaps most naturally-be read to permit judicial review *2151of issues bearing on the Patent Office's institution of inter partes review. Section 314(d) reads: "The determination by the Director whether to institute an inter partes review under this section shall be final and nonappealable." Unlike the statutes we addressed in Lindahl (including the one we found to permit review), § 314(d) does not say that an institution decision is "not subject to review." Instead, it makes the institution decision "nonappealable." This is fairly interpreted to bar only an appeal from the institution decision itself, while allowing review of institution-related issues in an appeal from the Patent Office's final written decision at the end of the proceeding. See § 319. Our cases have used the term "nonappealable" in just this way-to refer to matters that are not immediately or independently appealable, but which are subject to review at a later point.4 Thus, while the decision to institute inter partes review is "final and nonappealable" in the sense that a court cannot stop the proceeding from going forward,5 the question whether it was lawful to institute review will not escape judicial scrutiny. This approach is consistent with the normal rule that a party may challenge earlier agency rulings that are themselves "not directly reviewable" when seeking review of a final, appealable decision. 5 U.S.C. § 704. And it strikes a sensible balance: The Patent Office may proceed unimpeded with the inter partes review process (which must normally be completed within one year, see 35 U.S.C. § 316(a)(11) ), but it will be held to account for its compliance with the law at the end of the day.
In rejecting this commonsense interpretation, the Court gives short shrift to the presumption in favor of judicial review. Its primary reason for disregarding the presumption reduces to an assertion-devoid of any textual analysis-that surely § 314(d) must bar review of legal questions related to institution decisions. Ante, at 2139 - 2140. As I have explained, the statute's text does not require that conclusion.
Moving (further) away from the statutory text, the Court next objects that allowing judicial review "would undercut one important congressional objective, namely, giving the Patent Office significant power to revisit and revise earlier patent grants." Ante, at 2139. I am not sure that the Court appreciates how remarkable this assertion is. It would give us cause to do away with judicial review whenever we think that review makes it harder for an agency to carry out important work. In any event, the majority's logic is flawed. Judicial review enforces the limits that Congress has imposed on the agency's power. It thus serves to buttress, not "undercut," Congress's objectives. By asserting otherwise, the majority loses sight of the principle that "no legislation pursues its purposes at all costs." Rodriguez v. United States, 480 U.S. 522, 525-526, 107 S.Ct. 1391, 94 L.Ed.2d 533 (1987) (per curiam ). "Every statute purposes, not only to achieve certain ends, but also to achieve them by particular means-and *2152there is often a considerable legislative battle over what those means ought to be. The withholding of agency authority is as significant as the granting of it, and we have no right to play favorites between the two." Director, Office of Workers' Compensation Programs v. Newport News Shipbuilding & Dry Dock Co., 514 U.S. 122, 136, 115 S.Ct. 1278, 131 L.Ed.2d 160 (1995). The inter partes review statute is no exception. It empowers the Patent Office to clean up bad patents, but it expressly forbids the Patent Office to institute inter partes review-or even consider petitions for inter partes review-unless certain conditions are satisfied. Nothing in the statute suggests that Congress wanted to improve patent quality at the cost of fidelity to the law.
The Court also observes that the inter partes review appeal provision, § 319, "limit[s] appellate review to the 'final written decision.' " Ante, at 2140. The majority reads too much into this provision. Section 319 provides simply that "[a] party dissatisfied with the final written decision ... may appeal the decision." The statute does not restrict the issues that may be raised in such an appeal. As the Patent Office once explained (before having a change of heart), the "plain language of the statutory text" recognizes a "right of judicial review ... for any party 'dissatisfied' by the [Patent Office's] ultimate 'written [decision],' " and "[n]othing in the statutory scheme limits the reasons that a party might be so 'dissatisfied.' " Memorandum of Law in Support of Defendant's Motion to Dismiss in Versata Development Group, Inc. v. Rea, Civ. Action No. 1:13cv328 (ED Va., May 16, 2013), p. 16. A party may be dissatisfied with a final written decision in an inter partes review because the Patent Office lacked authority to institute the proceeding in the first place, or because the Office committed some other error in the lead-up to its final decision. Neither § 314(d) nor § 319 prevents a party from pressing such issues on an appeal from the final decision. This is familiar practice under 28 U.S.C. § 1291, which similarly limits appeals to "final decisions of the district courts" but allows appellants to challenge earlier rulings as part of those appeals. See Quackenbush v. Allstate Ins. Co., 517 U.S. 706, 712, 116 S.Ct. 1712, 135 L.Ed.2d 1 (1996) ("The general rule is that a party is entitled to a single appeal, to be deferred until final judgment has been entered, in which claims of district court error at any stage in the litigation may be ventilated" (internal quotation marks omitted)); 15A C. Wright, A. Miller, & E. Cooper, Federal Practice and Procedure § 3905.1, pp. 250, 252 (2d ed. 1992) (noting "the general rule that an appeal from final judgment ... permits review of all rulings that led up to the judgment" and observing that "[t]he variety of orders open to review on subsequent appeal from a final judgment is enormous"). And, as noted above, judicial review of "final agency action" likewise encompasses earlier rulings that are "not directly reviewable." 5 U.S.C. § 704 ; see supra, at 2151.
The Court next contends that my interpretation renders 35 U.S.C. § 314(d)"superfluous." Ante, at 2140. Reading the statute to defer review of institution decisions is "unnecessary," the Court says, because the "Administrative Procedure Act already limits review to final agency decisions" and a "decision to initiate inter partes review is 'preliminary,' not 'final.' " Ibid. But Congress reasonably may have thought that the matter needed clarifying, given that § 314(d) itself calls such a decision "final" (albeit in a different sense, see supra, at 2150 - 2151). Language is not superfluous when it "remove[s] any doubt" about a point that might otherwise be unclear.
*2153Ali v. Federal Bureau of Prisons, 552 U.S. 214, 226, 128 S.Ct. 831, 169 L.Ed.2d 680 (2008). More important, my reading prevents an appeal from a decision not to institute inter partes review, which is plainly final agency action and so-absent § 314(d) -might otherwise trigger immediate review. The Court asserts that this too is unnecessary because, in its view, a decision to deny inter partes review is "committed to agency discretion by law" and so unreviewable under normal principles of administrative law. 5 U.S.C. § 701(a)(2) ; see ante, at 2153. I agree that one can infer from the statutory scheme that the Patent Office has discretion to deny inter partes review even if a challenger satisfies the threshold requirements for review. But the law does not say so directly and Congress may not have thought the point self-evident. Again, 35 U.S.C. § 314(d) plays a clarifying role. This gives the provision plenty of work to do. There is no need to read it more broadly.6
III
A
None of this is to say that courts must-or should-throw out an inter partes review decision whenever there is some technical deficiency in the challenger's petition or in the Patent Office's institution decision. Although § 314(d) does not preclude review of issues bearing on institution, normal limits on judicial review still apply. For example, errors that do not cause a patent owner prejudice may not warrant relief. See 5 U.S.C. § 706 ("[D]ue account shall be taken of the rule of prejudicial error"). Some errors may also be superseded by later developments. Most notably, once the Patent Office issues its final written decision, the probabilistic question whether a challenger is "reasonabl[y] likel[y]" to prevail on the merits, 35 U.S.C. § 314(a), will be subsumed by the ultimate question whether the challenger should in fact prevail.7 And while I have no occasion here to decide the matter, it may be that courts owe some degree of deference to the Patent Office's application of the statutory prerequisites to inter partes review.
I would leave these considerations for the Court of Appeals to address in the first instance. But I must confess doubts that Cuozzo could ultimately prevail. As noted above, Cuozzo argues that the Patent Office improperly granted inter partes review of claims 10 and 14 on grounds not asserted in the petition for inter partes review, in violation of the statutory requirement *2154that a petition must state the grounds for challenge "with particularity." § 312(a)(3). The problem for Cuozzo is that claim 17-which the petition properly challenged-incorporates all of the elements of claims 10 and 14. Accordingly, an assertion that claim 17 is unpatentable in light of certain prior art is necessarily an assertion that claims 10 and 14 are unpatentable as well. Assuming that Cuozzo must show prejudice from the error it alleges, it is hard to see how Cuozzo could do so here.
B
But any perceived weakness in the merits of Cuozzo's appeal does not mean that such issues are unworthy of judicial review. Section 312(a)(3)'s particularity requirement is designed, at least in part, to ensure that a patent owner has sufficient notice of the challenge against which it must defend. Once inter partes review is instituted, the patent owner's response-its opening brief, essentially-is filed as an opposition to the challenger's petition. See § 316(a)(8) ; 37 CFR § 42.120. Thus, if a petition fails to state its challenge with particularity-or if the Patent Office institutes review on claims or grounds not raised in the petition-the patent owner is forced to shoot into the dark. The potential for unfairness is obvious.
Other problems arise if the Patent Office fails to enforce the prohibitions against instituting inter partes review at the behest of challengers that have already sued to invalidate the patent or that were sued for infringement more than a year before seeking inter partes review. 35 U.S.C. §§ 315(a)(1), (b). Allowing such a challenge exposes the patent owner to the burden of multiplicative proceedings-including discovery in both forums, see § 316(a)(5) -while permitting the challenger to exploit inter partes review's lower standard of proof and more favorable claim construction standard. Congress understandably thought that the Patent Office's power should not be wielded in this way. Yet, according to the Court, Congress made courts powerless to correct such abuses.
Even more striking are the consequences that today's decision portends for the AIA's other patent review mechanisms, post-grant review and CBM review, see supra, at 2148 - 2149 which are subject to a "no appeal" provision virtually identical to § 314(d). See § 324(e) ("The determination by the Director whether to institute a post-grant review under this section shall be final and nonappealable"); see AIA § 18(a)(1), 125 Stat. 329, note following 35 U.S.C. § 321, p. 1442 (CBM review generally "shall be regarded as, and shall employ the standards and procedures of, a post-grant review"). Post-grant review and CBM review allow for much broader review than inter partes review. While inter partes review is limited to assessing patentability under § 102 and § 103, in post-grant review and CBM review, patent claims can also be scrutinized (and canceled) on any invalidity ground that may be raised as a defense to infringement, including such grounds as ineligible subject matter under § 101, indefiniteness under § 112, and improper enlargement of reissued claims under § 251. See § 321(b) ; §§ 282(b)(2), (3). But this broader review comes with its own strict limits. A petition for post-grant review must be filed within nine months after a patent is granted. § 321(c). And while CBM review is not subject to this time limit, Congress imposed a subject-matter restriction: The Patent Office "may institute a [CBM review] proceeding only for a patent that is a covered business method patent," which Congress defined to cover certain patents with claims relating to "a financial product or service." AIA
*2155§§ 18(a)(1)(E), (d)(1), at 1442; see § 18(a)(1)(A), ibid.8
Congress thus crafted a three-tiered framework for Patent Office review of issued patents: broad post-grant review in a patent's infancy, followed by narrower inter partes review thereafter, with a limited exception for broad review of older covered business method patents. Today's decision threatens to undermine that carefully designed scheme. Suppose that the Patent Office instituted post-grant review on a petition filed 12 months (or even 12 years) after a patent was issued, and then invalidated a patent claim as indefinite under § 112 -a ground available in post-grant review but not in inter partes review. This would grossly exceed the Patent Office's authority and would be manifestly prejudicial to the patent owner. Can Congress really have intended to shield such shenanigans from judicial scrutiny? The Court answers with a non sequitur: Of course the Patent Office cannot cancel a patent under § 112"in inter partes review." Ante, at 2154. The Court seems to think that we could overturn the Patent Office's decision to institute "post-grant review" based on an untimely petition and declare that the agency has really instituted only "inter partes review." But how is that possible under today's opinion? After all, the petition's timeliness, no less than the particularity of its allegations, is "closely tied to the application and interpretation of statutes related to the Patent Office's decision to initiate ... review," and the Court says that such questions are unreviewable. Ibid. ; see § 321(c) ; § 312(a)(3).
To take things a step further, suppose that the Patent Office purported to forgive the post-grant review petition's tardiness by declaring the challenged patent a "covered business method patent," even though the patent has nothing to do with financial products or services (it claims, say, a new kind of tempered glass). Again, this involves the application of statutes related to the Patent Office's institution decision. See AIA § 18(a)(1)(E), at 1442 (Patent Office "may institute a [CBM review] proceeding only for a patent that is a covered business method patent"). So is this specious determination immune from judicial scrutiny under the Court's reasoning?
If judicial review of these issues is unavailable, then nothing would prevent the Patent Office from effectively collapsing Congress's three-tiered review structure and subjecting all patents to broad post-grant review at all times. Congress cannot have intended that.
I take the Court at its word that today's opinion will not permit the Patent Office "to act outside its statutory limits" in these ways. Ante, at 2141. But how to get there from the Court's reasoning-and how to determine which "statutory limits" we should enforce and which we should not-remains a mystery. I would avoid the suspense and hold that 35 U.S.C. § 314(d) does not bar judicial review of the Patent Office's compliance with any of the limits Congress imposed on the institution of patent review proceedings. That includes the statutory limit, § 312(a)(3), that Cuozzo alleges was violated here.
* * *
In enacting the AIA, Congress entrusted the Patent Office with a leading role in combating the detrimental effect that bad patents can have on innovation. But Congress did not give the agency unbridled authority. The principles I have set forth afford the Patent Office plenty of latitude to carry out its charge, while ensuring that *2156the Office's actions-no less than the patents it reviews-stay within the bounds of the law.
I would vacate the Federal Circuit's judgment and remand for that court to consider whether the Patent Office exceeded its authority to institute inter partes review with respect to claims 10 and 14 of Cuozzo's patent. With respect to claim 17, I agree with the Court that the judgment below must be affirmed. See n. 1, supra ; Part III, ante .
I agree with the Court that the Patent Office permissibly applies a "broadest reasonable construction" standard to construe patent claims in inter partes review, and I therefore join Parts I and III of its opinion.
The Director of the Patent Office has delegated his authority to institute inter partes review to the Patent Trial and Appeal Board (Board), which also conducts and decides the inter partes review. See 37 CFR §§ 42.4(a), 42.108 (2015) ; 35 U.S.C. §§ 316(c), 318(a). I therefore use the term "Patent Office" to refer to the Director, the Board, and the Patent Office generally, as the case may be.
The Court tries to recast Lindahl as a decision about "agenc [y] primacy" by focusing on its recognition that factual questions were unreviewable under the relevant statute (no one disputed that) and treating the case's holding that legal questions were reviewable as an afterthought. Ante, at 2140 - 2141. The review that Lindahl permitted-to correct "a substantial departure from important procedural rights, a misconstruction of the governing legislation, or some like error going to the heart of the administrative determination," 470 U.S., at 791, 105 S.Ct. 1620 (internal quotation marks omitted)-is quite similar to the review I envision of Patent Office decisions to institute inter partes review, as the discussion that follows makes clear.
See Mohawk Industries, Inc. v. Carpenter, 558 U.S. 100, 105, n. 1, 109, 130 S.Ct. 599, 175 L.Ed.2d 458 (2009) (agreeing with decisions holding that attorney-client privilege rulings are "nonappealable" because "postjudgment appeals generally suffice to protect the rights of litigants"); Coopers & Lybrand v. Livesay, 437 U.S. 463, 469, 472, n. 17, 98 S.Ct. 2454, 57 L.Ed.2d 351 (1978) (describing an order denying class certification as "nonappealable" but noting that it "is subject to effective review after final judgment").
Like the Court, I do not have occasion to address whether in extraordinary cases a patent owner might seek mandamus to stop an inter partes review before the proceeding concludes.
It is true that my interpretation leaves no apparent avenue (short of mandamus, at least) for judicial review of decisions not to institute inter partes review. This demonstrates that the presumption of reviewability has its limits. Nor is it surprising that Congress would design such a scheme. A patent challenger does not have nearly as much to lose from an erroneous denial of inter partes review as a patent owner stands to lose from an erroneous grant of inter partes review. Although such a challenger loses some of the advantages of inter partes review (such as a more favorable claim construction standard and a lower burden of proof), it remains free to challenge the patent's validity in litigation. A patent owner, on the other hand, risks the destruction of a valuable property right.
The Court recognizes that such issues are unreviewable even absent a statute like § 314(d), comparing the Patent Office's "reasonable likelihood" determination to an indicting grand jury's finding of probable cause. See ante, at 2153. But it draws the wrong analogy for this case. Cuozzo's complaint is that the petition for inter partes review did not articulate its challenge to certain patent claims with adequate particularity. This is more akin to an argument that an indictment did not sufficiently allege an offense and provide notice of the charges against the defendant, which is reviewable after trial and judgment. See, e.g., United States v. Carll, 105 U.S. 611, 612-613, 26 L.Ed. 1135 (1882) (overturning a conviction based on the insufficiency of the indictment).
Additionally, a challenger may file a petition for CBM review only if it has been sued for or charged with infringement of the patent. AIA § 18(a)(1)(B), at 1442.