Opinion for the court filed by Circuit Judge Taranto.
Opinion dissenting in part filed by Circuit Judge Dyk.
Taranto, Circuit Judge.Cellular Communications Equipment, LLC (CCE) owns U.S. Patent No. 7,218,-923, which describes and claims systems and methods for preventing third-party applications on smartphones from sending fraudulent and other problematic messages. Several groups of companies, including HTC Corp., HTC America, and ZTE (USA), Inc. (jointly, HTC), petitioned for an inter partes review (IPR) of certain claims of the ’923 patent under 35 U.S.C. §§ 311-312. The Patent Trial and Appeal Board of the.Patent and Trademark Office, acting as the delegate of the Office’s Director, 37 C.F.R. §§ 42.4(a), 42.108, instituted a review of claims 1, 2, 4, 5, 8, 24, 25, and 31. The Board subsequently rejected HTC’s challenges. HTC Corp. v. Cellular Commc’ns Equip., LLC, No. IPR2014-01133, 2016 WL 67220, at *7 (P.T.A.B. Jan. 4, 2016) (Final Written Decision). HTC appeals. We affirm.1
I
According to the ’923 patent, the existence of open development platforms for mobile communication terminals (such as smartphones) creates a risk that users of those communication terminals will download applications that may “behave contrary to the agreements made with the operator of the network, for example.” ’923 patent, col. 1, lines 38-47. The patent offers a solution. Messages that certain applications attempt to send over the network are diverted to a “controlling entity” that “resides in a tamper resistant area of the terminal.” Id., col. 1, line 59, though col. 2, line 10. The controlling entity then checks the message to determine “whether any changes are needed in the message or in the behavior of the application,” and will either allow the message through, modify it, or stop it. Id. Claims 1 and 24 are the *980only independent claims and are representative:
1. A method for controlling application programs in a communication terminal, the method comprising:
sending messages from an application program towards a communication network, the application program residing in a communication terminal;
diverting a message of the messages to a controlling entity residing in the communication terminal; and
based on the message, controlling in the controlling entity whether the application program behaves in a predetermined manner in the communication terminal, the - controlling being performed before the message is transmitted from the communication terminal to the communication network.
Id., col. 9, lines 10-22.
24. A terminal for a communication system, the terminal comprising:
an application program configured to send messages towards a communication network; and a diverting unit configured to divert a message of the messages sent from the application program and destined for the communication network to a controlling entity residing in the terminal,
wherein the controlling entity is configured to control, based on the message and before the message is transmitted to the communication network, whether the application program behaves in a predetermined manner in the communication terminal, and
wherein the terminal is a terminal of a communications system.
Id., col. 10, line 58, though col. 11, line 5.
The Board instituted a review of the patentability of claims 1, 2, 4, 5, 8, 24, 25, and 31 on two grounds: (1) anticipation by U.K. Patent Pub. No. 2,376,766 (D’Aviera); and (2) obviousness over U.S. Patent Pub. No. 2002/0065869 (Calder) and U.S. Patent No. 7,836,494 (Richardson). Final Written Decision, 2016 WL 67220, at *1.
In its petition, HTC argued that D’Avi-era disclosed every element-of the claims, relying on D’Aviera’s “isolator engine” to disclose the “diverting” and “controlling” limitations. For its obviousness challenge, HTC relied on the “interception module” of Calder as disclosing those limitations.
After the Board instituted the review, CCE, in its Patent Owner Response, argued: “The plain language of independent claims 1 and 24 clearly requires that the claimed ‘divert[ing]’ limitation is an intervening step that must be performed after a message has been sent from the claimed application program and before the message is' received and controlled by the claimed controlling entity.” J.A. 495 (alteration in original). CCE further argued that “to meet the claimed ‘diverging]’ step, there must be something that ‘diverts’ a message of messages after the message is sent from the application program and before the message is received by the controlling entity.” J.A. 501 (alteration in original). CCE contended that that “something” had to be “between the application program 210 and the isolation engine 225 (i.e., the alleged controlling entity)” of D’Aviera, J.A. 500, and “between the application program and the interception module 415 (i.e., the alleged controlling entity)” of Calder, J.A. 508.
In its Reply, HTC treated the central issue presented by CCE’s Response as whether, under the ’923 patent claims, the same structure could perform both the diverting and controlling functions. See J.A. 567-78. It pointed to authority explaining that different functions can be performed by the same structure, and it argued that the language of the claims did not preclude such a result here. HTC *981made no argument that claim 1, the independent method claim, differed from claim 24, the independent device claim, in whether it required different structures to perform the diverting and controlling functions.
In the Final Written Decision issued under 35 U.S.C. § 318(a), the Board concluded that the claims required “the diverting and controlling steps in claim 1” to be “performed by separate component's” and that the “diverting unit and controlling entity in claim 24” must be “separate components.” Final Written Decision, 2016 WL 67220, at *4.2 The Board observed that “[n]either party argues that claim 1 should be treated differently than claim 24 because claim 1 is a method claim.” Id. at *4 n.4. Applying that requirement of the claims to the prior art asserted by HTC, the Board determined that HTC had neither shown that “D’Aviera discloses a separate component for the diverting limitation in the challenged claims,” id. at *4, nor “identified] separate components in Calder and Richardson for performing the diverting and controlling steps in claim 1, or for the diverting unit and controlling entity in claim 24,” id. at *6. The Board therefore rejected HTC’s challenges to the patentability of the claims at issue. Id. at *5-6.
HTC appeals under 35 U.S.C. §§ 141(c) and 319. We have jurisdiction under 28 U.S.C. § 1295(a)(4)(A).
II
With one small exception, addressed in section IV, HTC’s entire argument to this court rests on its contention that the Board erred, as a matter of claim construction, in ruling that the “diverting” and controlling steps must be performed by separate components. Appellants’ Br. 20-49. We reject that contention.
As an initial matter, we agree with HTC that the Board’s ruling in this regard is a ruling on claim construction. Although the Board stated that “no claim terms require express construction,” Final Written Decision, 2016 WL 67220, at *3, the Board engaged in claim construction when it proceeded to determine whether D’Avi-era (and Calder) disclose the “diverting” and “controlling” limitations. It specifically concluded that the ’923 patent requires those steps to be performed by “separate components,” which is a conclusion about the scope of the ’923 patent, not the prior art, and which the Board drew using the standard tools of claim construction. Id. at *4-5 (citing claim construction cases, addressing the language of the claims, and examining the claims in light of the specification). Despite the heading under which the Board’s analysis took place, the Board’s ruling about the requirement of separate components was clearly a claim construction: it “establishfed] the scope and boundaries of the subject matter that is patented.” Netword, LLC v. Centraal Corp., 242 F.3d 1347, 1350 (Fed. Cir. 2001); see Regents of the Univ. of Minn. v. AGA Med. Corp., 717 F.3d 929, 939-40 (Fed. Cir. 2013).
By regulation, the Board, in construing claims in an IPR, must adopt the broadest reasonable interpretation in light of the specification.. 37 C.F.R. § 42.100(b); see Cuozzo Speed Techs., LLC v. Lee, — U.S. —, 136 S.Ct. 2131, 2142-46, 195 L.Ed.2d 423 (2016) (upholding regulation). The Board here relied entirely on intrinsic evidence in concluding that the patent claims require separate components for the di*982verting and controlling functions. Accordingly, we review the construction de novo. Microsoft Corp. v. Proxyconn, Inc., 789 F.3d 1292, 1297 (Fed. Cir. 2015).
We begin with device claim 24 because its language is on its face more informative than the language of claim 1. We agree with the Board that the broadest reasonable construction is one that requires separate components for diverting and for controlling. The strongest evidence for this separation is the claim language itself, which plainly recites two different structures: a “diverting unit” and a “controlling entity.” ’923 patent, col. 10, lines 62-65. The separate naming of two structures in the claim strongly implies that the named entities are not one and the same structure. See Becton, Dickinson & Co. v. Tyco Healthcare Grp., LP, 616 F.3d 1249, 1254 (Fed. Cir. 2010) (‘Where a claim lists elements separately, ‘the clear implication of the claim language’ is that those elements are ‘distinct component[s]’ of the patented invention.” (alteration in original) (quoting Gaus v. Conair Corp., 363 F.3d 1284, 1288 (Fed. Cir. 2004))).
The specification reinforces the inference that the claimed “diverting unit” and “controlling entity” are separate components. The specification nowhere gives an example of a single structure performing both functions. Instead, the specification consistently describes the “diverting” step as performed' by either the “SIP protocol stack” or a “middleware modification module.” See ’923 patent, col. 4, lines 46-61; id., col. 6, lines 20-26; id., col. 6, lines 49-55 (“The functionality needed in the SIP protocol stack may be introduced as changes made within the protocol stack, as is assumed in the above examples. Alternatively, the functionality may be introduced as a separate middleware modification module that resides between the applications(s) and the protocol stack and which thus also provides an Application Program Interface (API) for the applications,”). In contrast, the specification describes the “controlling” step as being performed by a “controlling entity,” identified as the “trusted agent,” id., col. 2, lines 10-11, which “may be a dedicated software agent or a Digital Rights Management (DRM) agent whose normal functionality has been modified for the method of the invention,” id., col. 3, lines 63-66. The specification thus describes two possible structures for performing the “diverting” step, both of them distinct from the two possible structures for performing the “controlling” step. The implication of distinctness is further bolstered by the specification’s later statement that “[i]n the above examples, it is assumed that the SIP stack (or the middleware modification module) can be changed but the trusted agent offers its services only to a valid SIP stack (or mid-dleware modification module).” Id., col. 7, lines 9-12. That statement distinguishes the trusted agent from the SIP stack and middleware module.
Furthermore, the specification repeatedly describes the “trusted entity” as residing in a “tamper resistant area of the terminal.” Id., col. 2, lines 3-4; id., col. 3, lines 58-63. Every time the “trusted entity” is shown in a figure, it is shown inside the “tamper resistant area” and separate from the “SIP-protocol stack,” which is doing the diverting and resides in the “open platform area.” Id., figs.2, 6-8. The specification contemplates that the SIP stack might sometimes be in the “tamper resistant area,” but speaks of it as a separate entity being wholly moved, not incorporated into the “controlling entity.” Id., col. 7, lines 5-9 (“If the use [of] unauthorized applications is also to be eliminated, the SIP stack must be moved to the tamper-proof area or the possibility to change the SIP stack is to be eliminated other*983wise.”)- For all of these reasons, the only reasonable construction of the claim language, in light of the specification, is that claim 24 assumes separate structures for the “diverting unit” and the “controlling entity.”
HTC counters that, in some patents, “a single prior art structure that performs two functions can satisfy two functional Claim elements.” Appellants’ Br. 14; id. at 32-33.3 But, as HTC’s main authority makes explicit, “can satisfy” does not mean “always satisfies”; rather, the claims of any given patent must be evaluated in light of the specification to determine whether their scope requires more than one structure. In re Kelley, 305 F.2d 909, 916 (C.C.P.A. 1962). In this case, for the reasons already given, claim 24 requires two structures for the “diverting unit” and the “controlling entity.”
Turning to claim 1, the independent method claim, we conclude that HTC has forfeited any argument for a conclusion different from the one we (and the Board) have drawn for claim 24. In this court, HTC separately addressed the language of claims 1 and 24, see Appellants’ Br. 22-38, but it made no argument in its briefs that the two claims could support different conclusions regarding whether two structures are required .for the diverting and controlling functions. And at oral argument, HTC stated that, while the claims use different language, there is no “meaningful distinction” between claims 1 and 24 for this purpose. Oral Arg. 1:04— 1:25, http://oralarguments.cafc.uscourts. gov/default.aspx?fl=2016-1858,mp3. That position, moreover, reflects how HTC presented its case to the Board. Even though HTC recognized in its Reply that the central issue was whether a single structure could perform the two functions, and discussed claims 1 and 24 under separate headings, J.A. 567-78, HTC never argued to the Board that it could prevail on claim 1 even if it lost on claim 24, see J.A. 685-66 (HTC arguing to the Board that claims 1 and 24 are “very similar”). The Board so recognized. Final Written Decision, 2016 WL 67220, at *4 n.4, A fortiori, HTC never presented a developed explanation for why the Board could distinguish between those claims with respect to the separate-components requirement. Nor has HTC presented such an argument in this court; indeed, it has asserted that the textual differences are ultimately not meaningful.
In the particular circumstances presented, we will not consider giving claim 1, in the respect at issue, a different scope from that of claim 24, which, we have concluded, requires two separate components for the claimed “diverting” and “controlling.” We rely on what HTC has and has not argued in both this court and in the PTO. In this court, HTC stated that there is no meaningful difference between the claims, and presented no explanation for differentiating the claims’ scope in the respect that is centrally at issue in this IPR. See SmithKline Beecham Corp. v. Apotex Corp., 439 F.3d 1312, 1319-20 (Fed. Cir. 2006) (treating an argument not meaningfully developed on appeal as waived). In the PTO, HTC did not “fairly put [the Board] on notice as to the substance of’ any request to distinguish claims 1 and 24 regarding this critical issue of scope. See Nelson v. Adams USA Inc., 529 U.S. 460, 469, 120 S.Ct. 1579, 146 L.Ed.2d 530 (2000) (explaining that the critical issue regarding preservation for appeal is fair notice in initial tribunal). *984In Interactive Gift Express, Inc. v. Compuserve Inc., 256 F.3d 1323 (Fed. Cir. 2001), this court discussed the preservation of contentions regarding claim construction for appeal. It explained that presenting a “new claim construction position” that “change[s] the scope of the claim construction” on appeal is generally impermissible, whereas “clarifying or defending the original scope of [a] claim construction” or “supporting [the] existing claim construction position with new citations to the specification” is generally permissible. Id. at 1346-47. But the issue of differentiating claim l’s scope from claim 24’s scope does not fit neatly into either category. And, as noted, even in this court, HTC has not argued for such differentiation.
Judgment and discretion are involved in assessing whether an argument has been both adequately preserved and adequately developed and whether to overlook deficiencies in either respect. See, e.g., Icon Health & Fitness, Inc. v. Strava, Inc., 849 F.3d 1034, 1040 (Fed. Cir. 2017); SmithKline, 439 F.3d at 1320; Harris Corp. v. Ericsson Inc., 417 F.3d 1241, 1251 (Fed. Cir. 2005); see also 16AA Charles Alan Wright et al., Federal Practice & Procedure § 3974.1 n.36 (4th ed. 2017). Here, we think, the interests in fair notice to, the initial tribunal and to the opposing party (both before the Board and in this court) would be ill-served were we now to distinguish claim 1 from claim 24 where HTC never directly argued for such different scopes or explained why it justifiably should win on claim 1 even if it lost on claim 24, and where HTC stated in this court there was no meaningful difference between the claims. Those interests have particular weight in the context before us. It is not plausible that HTC inadvertently overlooked the potential for a half win, which it might well have deliberately chosen to discourage by taking an all-or-nothing approach. Moreover, the IPR regime at issue places a special premium on efficient Board resolution of patentability challenges. See 35 U.S.C. §§ 316(a)(11), 318(d); Cuozzo, 136 S.Ct. at 2144; H.R. Rep. No. 112-98, pt. 1, at 39-40, 45 (2011), as reprinted in 2011 U.S.C.C.A.N. 67, 69, 75. In these circumstances, we will not decide whether claim 1 should be accorded a different scope from that of claim 24 in the respect at issue.
, Other cases involving different circumstances could justify a different judgment. This is not a case in which “ ‘the proper resolution is beyond any doubt, or where injustice might otherwise result.’ ” Interactive Gift, 256 F.3d at 1344 (quoting Singleton v. Wulff, 428 U.S. 106, 121, 96 S.Ct. 2868, 49 L.Ed.2d 826 (1976)); see also Icon Health, 849 F.3d at 1040. As to the proper resolution, the seemingly relevant considerations are mixed. On one hand, the claims differ in their language: claim 1 makes no mention of a “diverting unit,” and we see no reason that a “controlling entity” cannot “divert” a message to itself as a matter of English language alone, putting aside what the context of this particular patent may require. On the other hand, such context cannot be put aside, and claim 1, through its ■ claiming of a “controlling entity,” is linked to claim 24, which claims the same “controlling entity,” raising a question about whether the common entity could be properly understood as doing more (ie., diverting) in claim 1 than it does in claim 24. See In re Varma, 816 F.3d 1352, 1363 (Fed. Cir. 2016). Moreover, as we have explained, the specification gives no example of a single-structure embodiment, whether of a device or a method. At least for those reasons, the dissent’s view — that this patent is reasonably read to claim a method invention broader (in the respect at issue) than the device version of the invention — is not “be*985yond any doubt.”4
Nor does this case qualify for the narrow “injustice” exception. HTC bypassed a clear opportunity to present an argument for distinguishing the results as to claim 1 and claim 24, here and before the Board, in circumstances of a type that often involve tactical choices to omit arguments. See Icon Health, 849 F.3d at 1040 (explaining that whether there was an “opportunity to raise the objection” is a factor in overlooking waiver (internal quotation marks omitted)). And we have no basis for thinking that, by not deciding claim l’s scope on the merits, we are resolving “sig-. nificant questions of general impact or of great public concern.” Id. (internal quotation marks omitted). We have a patent-specific claim-construction issue, not a matter of statutory interpretation. See Teva Pharm. USA, Inc. v. Sandoz, Inc., — U.S. —, 135 S.Ct. 831, 840, — L.Ed.2d — (2015) (distinguishing claim construction from statutory interpretation regarding breadth of impact). And our conclusion leaves other potential challengers (if not otherwise barred by, e.g., estoppel) free to argue the point on its merits. We therefore adopt the same construction for claim 1 as for claim 24.
III
HTC makes one additional claim-construction argument. It attacks the Board’s requirement of separate components as ambiguous and impossible to apply. HTC’s point here boils down to an argument that, even if the claims require “separate elements,”- that requirement is met by any piece of software because, as long as the claims are not specific as to the kind of structure that might be required (such as an application, executable file, library, object code file, etc.), the “structure” may be a single instruction, and the “diverting” instruction necessarily will be a “separate” line of code from the “controlling” instruction. Appellants’ Br. 43-48.
HTC waived this argument. As its Reply filed in the IPR makes clear, HTC recognized that the central issue was whether the claims required two separate components for the diverting and controlling functions, yet it never argued to the Board that such a requirement was improper because it would be ambiguous, impossible to apply, hollow, or the like, given the nature of software. It argued only that, as a matter of law, the same structure could be relied on to meet two functional claim elements. See J.A. 567-78. And that is all that is said in the one sentence from the Reply that HTC quotes to this court in arguing against waiver: “That such structure [for diverting] might be included in the same ‘box’ as the controlling entity is irrelevant, so long as the elements of the claim are met.” J.A. 571.5
IV
HTC makes a final argument to the effect that the Board should have looked at *986the entirety of certain prior-art references for assessing obviousness even if the separate-component requirement was not satisfied by those references. But it does not point to any argument in its petition that certain claim requirements, including the separate-component one, would be obvious in light of prior art even if no prior art disclosed that element. Therefore, this argument is waived.
V
For the foregoing reasons, we affirm the Board’s decision that HTC did not carry its burden to show the unpatentability of claims 1, 2, 4, 5, 8, 24, 25, and 31 of the ’923 patent.
AFFIRMED
. HTC was joined in the IPR by co-petitioners NEC Corporation of America; NEC CAISO Mobile Communications, Ltd.; LG Electronics, Inc.; LG Electronics U.S.A., Inc.; Pan-tech Co., Ltd.; and Pantech Wireless, Inc. By the time HTC filed its opening brief in this court, only HTC remained as a challenger.
. As the case comes to us, no separate issue is presented as to the dependent claims; only claims 1 and 24 need be discussed.
. The Summary of the Invention describes "one embodiment” as the "provision of a method,” without specifically calling for two structures, and "another embodiment” as a "terminal,” which includes two (functionally defined) structures. ’923 patent, col. 2, lines 11-31. HTC has not pointed to that passage and relied on it as distinguishing the method and device claims in the respect at issue.
. The Board observed that HTC made some suggestion that certain aspects of D’Aviera and Calder met the two-structure requirement, but the Board found HTC’s showing insufficient. Final Written Decision, 2016 WL 67220, at *5. That is a conclusion about application of the claim construction to the prior art. HTC’s argument to this court is entirely about claim construction, not application (with the minor exception noted in section IV).