In two inter partes review proceedings requested by Samsung, the Patent Trial and Appeal Board ("the Board") upheld all challenged claims of U.S. Patent 8,917,772 ("the '772 patent"), which is owned by Infobridge. The Board upheld the claims because it found that Samsung failed to show that a certain prior art reference was publicly accessible before the '772 patent's critical date and thus could not be considered prior art. Samsung appeals the Board's decisions, insisting that it has standing to do so and arguing, among other things, that the Board applied the wrong legal standard in assessing public accessibility. We agree with Samsung on both issues. We therefore vacate the Board's decisions and remand for further consideration.
I. BACKGROUND
A. The Technology
The '772 Patent, titled "Method of Constructing Merge List," generally relates to encoding and decoding video data. Both parties agree for purposes of this appeal that the patented methods are essential to the High Efficiency Video Coding standard ("the H.265 standard").
B. The Prior Art
The sole prior art reference at issue on appeal is Working Draft 4 of the H.265 standard ("the WD4 reference"), which was developed by the Joint Collaborative Team on Video Coding ("JCT-VC"). Samsung Elecs. Co., Ltd v. Infobridge Pte. Ltd , IPR2017-00099, 2018 WL 1940480 at *2 (P.T.A.B. April 23, 2018) (" Final Written Decision ").1 The key question here is whether the WD4 reference was publicly accessible prior to the '772 patent's critical date. Samsung points to three examples of disclosures that it claims, independently or *1366together, establish public accessibility. Each are discussed below.
1. JCT-VC Meetings
JCT-VC members met to discuss the H.265 standard, then under development, during a July 2011 meeting in Torino, Italy ("the Torino meeting"). This meeting included about 250 participants, ranging from academics to representatives from various technology companies. One of the "primary goals" for the Torino meeting was developing the WD4 reference. J.A. 5327. In November 2011, at the next JCT-VC meeting in Geneva, Switzerland ("the Geneva meeting"), the WD4 reference was discussed and ultimately "approved." J.A. 5587.
2. JCT-VC and MPEG Websites
During this same time, the JCT-VC maintained a website allowing users to access various JCT-VC materials. The WD4 reference was uploaded to the JCT-VC's website on October 4, 2011. Final Written Decision , 2018 WL 1940480 at *5. To access the WD4 reference, users needed to follow at least four steps. First, they had to navigate to the JCT-VC website. J.A. 7026. Next, they had to select a menu option to view information about "All meetings" held by the JCT-VC. J.A. 7029. Then a user would need to select "Torino" from the list of available meeting options, which were not identified by subject matter. Id. At this point, the user would see a list of "hundreds" of documents organized by an identifying number rather than subject matter. Final Written Decision , 2018 WL 1940480 at *4 ; see also J.A. 7034-7081 (listing the documents). From this list, a user would need to select the WD4 reference, titled "WD4: Working Draft of High-Efficiency Video Coding." J.A. 7080.
The WD4 reference was also uploaded to a website maintained by the Moving Picture Expert Group ("MPEG"), a parent organization of the JCT-VC, on October 4, 2011. Final Written Decision , 2018 WL 1940480 at *2-3, *5. The MPEG website was arranged in a manner similar to the JCT-VC website, i.e. , a user would need to navigate to the relevant meeting and then select the relevant document. Id. at *8-9. But, unlike the JCT-VC website, the MPEG website required users to have a login and password to access materials. Id. at *7.
3. JCT-VC Listserv
On October 4, 2011, the same day it was uploaded to the Internet, Benjamin Bross-the lead author of the WD4 reference-emailed the reference to a JCT-VC listserv. J.A. 6377. According to Mr. Bross, the listserv included JCT-VC members who had attended the Torino meeting as well as other "interested individuals." J.A. 7947. Mr. Bross's email included a download link for the WD4 reference. Final Written Decision , 2018 WL 1940480 at *9.
C. The Board's Decisions
Samsung filed two inter partes review petitions challenging claims 1-7 (IPR2017-00099) and claims 8-9 (IPR2017-00100) of the '772 patent on October 17, 2016. The Board instituted as to all challenged claims. In both proceedings, the Board's institution decisions relied on prior art combinations including the WD4 reference. In its final written decisions, however, the Board concluded that the WD4 reference was not publicly accessible prior to the '772 patent's critical date and therefore could not be prior art. Id. at *10.
As to the JCT-VC website, the Board concluded that there was "insufficient competent evidence" to show that a person of ordinary skill in the art "would [have] know[n] to check the JCT-VC site for information of relevance to the art." Id. at *5. Although Samsung offered testimony from Mr. Bross on this point, the Board *1367rejected his testimony as "conclusory and insufficiently factually supported" because Mr. Bross could not testify about whether others would have navigated to the JCT-VC website to learn about developments in video coding. Id. at *6. The Board therefore found that Samsung could not "establish that WD4 was accessible to anyone other than members of JCT-VC" during the relevant time period. Id. The Board went on to explain that, assuming a person of ordinary skill might have known about the JCT-VC website, there was no evidence that such a person would have located the WD4 reference on the website by exercising reasonable diligence. Id. at *7 ("Although the JCT-VC site was organized in a hierarchical manner, the evidence does not establish WD4 was indexed in a manner that one ordinarily skilled in the art, exercising reasonable diligence, could locate it.").
The Board reached the same conclusion with respect to the MPEG website for essentially the same reasons. Id. at *7-9. For example, the Board again faulted Samsung for offering only "conclusory and factually unsupported" assertions that those skilled in the art were aware of the MPEG website. Id. at *8. The Board also found that, even if someone could have found the MPEG website, they would not have been able to reasonably find the WD4 reference. Id. ("[W]e also find Samsung has failed to present evidence that a person interested and skilled in the art could reasonably have found WD4 on the MPEG site in October 2011, even if the user had access and knew the reference was there.").
As to the listserv email sent by Mr. Bross, the Board found Mr. Bross's testimony that the email was sent to individuals outside the JCT-VC as "nothing more than conjecture and speculation." Id. at *9. The Board went on to conclude:
We find Mr. Bross's email to the JCT-VC [listserv] was, at best, a limited distribution of a link to the WD4 document information webpage to a select group, the members of JCT-VC and perhaps others who may have opted into the reflector. We also find Mr. Bross's testimony is insufficient to establish that the e-mail with the link to the WD4 document information webpage was generally disseminated to persons interested and ordinarily skilled in the art. "[T]his record does not evince that [WD4] was accessible to anyone other than the [JCT-VC], thus further suggesting an absence of actual public accessibility."
Id. at *9 (quoting SRI Int'l, Inc. v. Internet Sec. Sys., Inc. , 511 F.3d 1186, 1196 (Fed. Cir. 2008) ) (first alteration added).
The Board did not address whether the WD4 reference was publicly accessible because it had been disseminated at the Torino or Geneva meetings.
Samsung timely appealed the Board's final decisions. We have jurisdiction under 28 U.S.C. § 1295(a)(4)(A).
II. DISCUSSION
Samsung's sole argument on appeal is that the Board erred by finding that the WD4 reference was not publicly accessible prior to the '772 patent's critical date. We turn to this argument below after addressing whether Samsung has standing to bring this appeal in the first place.
A. Standing
Federal courts exercise their judicial power according to limits imposed by Congress and the Constitution. Spokeo, Inc. v. Robins , --- U.S. ----, 136 S. Ct. 1540, 1546-47, 194 L.Ed.2d 635 (2016). These constitutional limits establish "the category of litigants empowered to maintain a lawsuit in federal courts to seek redress for a legal wrong." Id. at 1547.
*1368And they apply at every stage of a case before a federal court, including on appeal. Phigenix, Inc. v. Immunogen, Inc. , 845 F.3d 1168, 1171-72 (Fed. Cir. 2017). To clear this constitutional threshold, the party seeking to invoke our judicial power must show that it has suffered a "concrete and particularized" injury that is either actual or imminent. Lujan v. Defs. of Wildlife , 504 U.S. 555, 560-61, 112 S.Ct. 2130, 119 L.Ed.2d 351 (1992). It must also show that this injury is fairly traceable to its adversary and likely to be redressed by a favorable ruling. Id.
We have "an independent obligation to assure that standing exists, regardless of whether it is challenged by any of the parties." Summers v. Earth Island Inst. , 555 U.S. 488, 499, 129 S.Ct. 1142, 173 L.Ed.2d 1 (2009). Here, Samsung raised the issue of standing in its opening brief. Infobridge did not contest that Samsung has standing in response, but we think it is necessary to address the question.
Samsung's standing argument turns on its relationship to Infobridge and the '772 patent. According to uncontroverted evidence provided by Samsung, the '772 patent is licensed as part of a "pool" of patents, including some owned by Samsung, that have been declared essential to the H.265 standard. Appellant's Br. 59. Licensees pay a fixed royalty for the pooled patents and then members who own patents in the pool divide that royalty based on the number of patents in the pool. By the express terms of the license, if a pool patent is declared invalid, it is removed from the pool and the other members thereafter receive a higher proportion of the fixed royalty. Members of the pool, like Samsung, therefore stand to gain if another pool patent is invalidated and removed from the pool.
Against this backdrop, Samsung argues that it is being "depriv[ed]" of royalty payments and that "[t]his deprivation of royalties is the kind of 'concrete and particularized' economic injury that satisfies the Article III requirement." Id. (quoting Phigenix , 845 F.3d at 1171 ). We agree with Samsung that, under the facts and circumstances of this case, its injury confers Article III standing. See, e.g. , Chou v. Univ. of Chicago , 254 F.3d 1347, 1359 (Fed. Cir. 2001) ("If Chou has indeed been deprived of an interest in proceeds from licensing the invention ... then she will have suffered an injury-in-fact, i.e. , the loss of those benefits"). This injury can be traced directly to the validity of Infobridge's patent and would be redressed by a favorable decision for Samsung. While other licensing and royalty structures might compel a different result where other standard-essential patents are involved, the unique pool license here satisfies us that Samsung has standing in this appeal.
Accordingly, we proceed to the merits of Samsung's appeal.
B. Public Accessibility
A person is not entitled to a patent if their invention was "described in a printed publication ... more than one year prior to the date of the[ir patent] application." 35 U.S.C. § 102(b) (2006).2 We have read this statutory text as imposing two requirements: (1) that a putative prior art reference be printed and (2) that the reference be published, i.e. , accessible to the public. In re Wyer , 655 F.2d 221, 225 (CCPA 1981). We have emphasized, however, *1369that what constitutes a "printed publication" must be determined in light of the technology employed, id. at 226, and that it is public accessibility that is the "touchstone," Jazz Pharm., Inc. v. Amneal Pharm., LLC , 895 F.3d 1347, 1355 (Fed. Cir. 2018) (internal quotation marks omitted). A reference is considered publicly accessible if "persons interested and ordinarily skilled in the subject matter or art, exercising reasonable diligence, can locate it." Acceleration Bay, LLC v. Activision Blizzard Inc. , 908 F.3d 765, 772 (Fed. Cir. 2018). In some cases, this inquiry is straightforward. But in other cases, public accessibility depends on a careful, case-by-case examination of how a particular reference was disseminated, to whom, for how long, and under what circumstances. Jazz , 895 F.3d at 1357.
When a reference is uploaded to a website or deposited in a library, the fact that the reference is indexed or cataloged in some way can indicate that it is publicly accessible. In Acceleration Bay , for example, the reference at issue was uploaded to a university website. 908 F.3d at 773. While this made the work technically accessible-someone could theoretically find it on the Internet-we explained that "public accessibility requires more than technical accessibility." Id. (internal quotation marks omitted). And while the website indexed references "by author and year," we agreed with the Board that this did not mean the reference at issue had been "meaningfully indexed such that an interested artisan exercising reasonable diligence would have found it." Id. at 774 ; see also SRI , 511 F.3d at 1196 ("[T]he FTP server did not contain an index or catalogue or other tools for customary and meaningful research."). On this record, we affirmed the Board's finding that the reference was not publicly accessible even though it had been uploaded to a website.
Similarly, in Voter Verified, Inc. v. Premier Election Solutions, Inc. , 698 F.3d 1374 (Fed. Cir. 2012), there was also no evidence that the reference had been meaningfully indexed or cataloged. But there was "uncontested evidence" that "a person of ordinary skill interested in [the relevant subject matter] would have been independently aware of [the website]" and that "such an interested researcher would have found [the reference] using that website's own search functions and applying reasonable diligence." 698 F.3d at 1381 ; see also Jazz , 895 F.3d at 1359 (noting "unchallenged findings" that a person of ordinary skill in the relevant art "would have had reason to" find the website at issue). In view of these facts, among others, we affirmed a district court's conclusion that the reference was publicly accessible. Voter Verified , 698 F.3d at 1381. Thus, while indexing is not required to show that a work is publicly accessible, some evidence that a person of ordinary skill could have reasonably found the website and then found the reference on that website is critical. Acceleration Bay , 908 F.3d at 773.
To the extent the Board resolves factual issues in determining that a reference is or is not prior art under § 102, its factual findings are reviewed for substantial evidence. Blue Calypso, LLC v. Groupon, Inc. , 815 F.3d 1331, 1339, 1348 (Fed. Cir. 2016). "[Substantial evidence] means-and means only-such relevant evidence as a reasonable mind might accept as adequate to support a conclusion." Biestek v. Berryhill , --- U.S. ----, 139 S. Ct. 1148, 1154, 203 L.Ed.2d 504 (2019) (internal quotation marks omitted). The ultimate question of whether a reference is prior art under § 102(b), however, is a legal question we review de novo. Blue Calypso , 815 F.3d at 1339, 1348.
Samsung argues that the WD4 reference was publicly accessible because it was discussed *1370at various JCT-VC meetings, uploaded to the JCT-VC and MPEG websites, and emailed to the JCT-VC listserv. We address each argument below.
1. JCT-VC Meetings
Samsung argues that the WD4 reference became publicly accessible because it was developed at a "prominent international conference," Appellant's Reply Br. 15, and was distributed to attendees, id. at 12-15; Appellant's Br. 32-36. Infobridge responds that Samsung actually argues distribution of the WD4 reference at the Torino and Geneva meetings, an argument that Infobridge contends Samsung has waived. To the extent Samsung can be understood to be arguing distribution of the WD4 reference at the Torino and Geneva meetings, we agree the argument was waived.
Before the Board, Samsung did not argue that the WD4 reference was publicly accessible because it was discussed at the Torino meeting or disseminated at the Geneva meeting. In fact, Samsung never mentioned the Geneva meeting. To the extent Samsung mentioned the Torino meeting, moreover, it was related to the JCT-VC website and Mr. Bross's email. J.A. 132-39. Samsung conceded as much at oral argument. See Oral Arg. at 1:23-54 ("Q. I thought it was pretty clear from your reply brief that you had backed away [from relying on the Torino meeting]? A. We have never taken the position before the Board or here that there was a distribution of WD4 at the Torino meeting .... Q. But even as to the Geneva meeting .... you never relied on that alone? A. We are not relying on the Geneva meeting ...."), available at http://oralarguments.cafc.uscourts.gov/default.aspx?fl=2018-2007.mp3. Thus, Samsung waived any separate argument on appeal that the WD4 reference was publicly accessible because it was disseminated at the Torino or Geneva meetings. See MCM Portfolio LLC v. Hewlett-Packard Co. , 812 F.3d 1284, 1294 n.3 (Fed. Cir. 2015) (deeming arguments not presented to the Board waived on appeal); Novartis AG v. Torrent Pharm. Ltd. , 853 F.3d 1316, 1330 (Fed. Cir. 2017) (same).
Alternately, we agree with Infobridge that, if not waived, any argument based upon distribution of the WD4 reference at the Torino and Geneva meetings is without merit. The WD4 reference was not created until after the Torino meeting. J.A. 7949. The reference therefore could not have been circulated to conference attendees at the meeting itself. Moreover, Samsung conceded at oral argument that the Geneva meeting occurred after the relevant critical date. Oral Arg. at 1:41-54 (noting that the Geneva meeting occurred "a few days after the one-year period"). Thus, even if the disclosure at that meeting could establish public accessibility, it would not establish that the WD4 reference could serve as prior art here.
2. JCT-VC and MPEG Websites
To establish that the WD4 reference was publicly accessible once it was uploaded to the JCT-VC website, Samsung provided a declaration from Mr. Bross where he testified as follows:
[B]ased on my knowledge and recollection, given the prominence of the JCT-VC in the video coding industry, persons interested in tracking the developments of the latest video coding standard would regularly visit the JCT-VC site to ensure that products and services they were developing were consistent with the HEVC Standard under development.
Final Written Decision , 2018 WL 1940480 at *5-6 (quoting J.A. 7942). The Board rejected this testimony as "conclusory and insufficiently factually supported." Id. at *6. According to the Board, Mr. Bross had *1371no basis to say whether others "regularly visit[ed]" the JCT-VC website or thought it was "prominen[t]." Id.
Samsung also argued that Infobridge's expert, Dr. Furht, supported Mr. Bross's testimony. Id. But the Board rejected this characterization of Dr. Furht's testimony. Id. As the Board explained, Dr. Furht acknowledged that he was "probably aware," before June 2013, that efforts were being made to develop the H.265 standard. Id. The Board concluded that "this testimony [was] not relevant or probative of what persons of skill in the video coding art in general knew about WD4" or the JCT-VC website. Id.
Having rejected Mr. Bross's testimony, and without additional support, the Board concluded that Samsung failed to show that ordinarily skilled artisans "who were not part of JCT-VC" would have known about or found the JCT-VC website. Id. On this basis, the Board found that Samsung "d[id] not establish that [the] WD4 [reference] was accessible to anyone other than members of JCT-VC." Id.
The Board also found that a skilled artisan would not have been able to locate the WD4 reference on the JCT-VC website-even assuming the JCT-VC website itself was accessible-by exercising reasonable diligence. Id. at *7. To support this conclusion, the Board walked through the steps that a person would have to go through to find the WD4 reference on the JCT-VC website. Id. As the Board explained, "identifying a meeting location was key to navigating the JCT-VC site." Id. In other words, if a person did not know to search for the WD4 reference by looking under the "Torino" folder-named for the Torino meeting-then a person would not have found the WD4 reference. Id. But the Board noted that there was "no evidence" anyone, outside those participating in the JCT-VC meetings, would have found "cities ... helpful in any respect in locating a document on the site." Id. This difficulty was compounded by the fact that Samsung presented "no evidence that one could search for or locate [the] WD4 [reference] based on its subject matter." Id. The Board therefore concluded that Samsung failed to show that the WD4 reference was publicly accessible. Id.
Samsung argues that the Board failed to consider whether access by members of the JCT-VC could make the WD4 reference publicly accessible. Appellant's Br. 45 ("The Board improperly required Samsung to prove access outside the JCT-VC ... membership ...." (emphasis added)). This error is critical, Samsung insists, because the JCT-VC was composed of more than 250 members who were skilled artisans following the development of H.265 and video coding in general. Id. According to Samsung, sharing the WD4 reference among JCT-VC members is like an academic presenting a paper at a conference, which we have said can make a work publicly accessible. Id. at 44 (collecting cases).
We are not persuaded by Samsung's analogy. Like the Board, we find SRI is instructive. There, the reference at issue was emailed to the organizer of an upcoming symposium, Dr. Bishop. SRI , 511 F.3d at 1190. Based on that record, the district court concluded on summary judgment that the email made the work publicly accessible. Id. at 1194. On appeal, we remanded based on various disputed facts. As relevant here, we noted that the evidence suggested "only one non-SRI person, Dr. Bishop, knew about the availability of [the reference]." Id. at 1196 (citing Application of Bayer , 568 F.2d 1357 (CCPA 1978) ). This "militate[d] against a finding of public accessibility." Id. at 1197. We reached a similar result in Bayer , where the reference at issue, a student thesis, was only accessible to members of a faculty review committee.
*1372568 F.2d. at 1361 (concluding that "[a]ccessibility to appellant's thesis by the three members of the graduate committee under the circumstances of the present case" did not demonstrate that the work was publicly accessible); see also SRI , 511 F.3d at 1196 (noting the similarity between "only one non-SRI person" having access to a reference and "the knowledge of the thesis's availability by the three professors in Bayer ").
Taken together, these cases suggest that a work is not publicly accessible if the only people who know how to find it are the ones who created it. This is why SRI focused on the knowledge of those outside the authoring organization and why Bayer discounted the knowledge of various professors on a faculty committee reviewing student theses. To hold otherwise would disincentivize collaboration and depart from what it means to publish something.
The Board's analysis is consistent with this principle and its underlying conclusions are supportable. Unlike the conference attendees in GoPro, Inc. v. Contour IP Holding LLC , 908 F.3d 690 (Fed. Cir. 2018) or Medtronic, Inc. v. Barry , 891 F.3d 1368 (Fed. Cir. 2018), JCT-VC members were part of ongoing, collaborative efforts to draft the WD4 reference. For example, Mr. Bross testified that the WD4 reference was "developed by the Joint Collaborative Team on Video Coding (JCT-VC)." J.A. 7936; see also J.A. 5326 (noting that "[t]he JCT-VC produced ... the HEVC specification Working Draft 4 (WD4)"). Even if Mr. Bross was the lead author of the WD4 reference, we cannot say on this record that the Board erred in treating the other JCT-VC members who knew about this ongoing project like the faculty advisers in Bayer or in-house employees in SRI . The Board therefore properly focused on whether those outside of the JCT-VC knew about the JCT-VC website in considering whether posting the WD4 reference on the website made it publicly accessible.
Samsung also argues that the Board's factual findings about the website, and the accessibility of the WD4 reference on the website, were not supported by substantial evidence. We disagree. Samsung did not offer evidence, apart from the speculation of Mr. Bross, that a person of ordinary skill, exercising reasonable diligence, would have located the JCT-VC website or even known to look for it.3 Final Written Decision , 2018 WL 1940480 at *5-6. That sets this case apart from Voter Verified and Jazz . Cf. Voter Verified , 698 F.3d at 1381 (noting uncontested evidence that persons of ordinary skill in the art knew about the website); Jazz , 895 F.3d at 1359 (same). And, as the Board explained in detail, even a person who found the JCT-VC website lacked a reasonable way of locating the WD4 reference unless they already knew what to look for and where to look for it. Id. at *7. This supports the Board's conclusion that the WD4 reference would not have been reasonably accessible even if a skilled artisan knew about the JCT-VC website. See Bayer , 568 F.2d at 1361 ("[S]ince appellant's thesis could have been located in the university library only by one having been informed of its existence by the faculty committee, and not by means of the customary research aids available in the library, the 'probability of *1373public knowledge of the contents of the [thesis],' was virtually nil." (internal citation omitted)); In re Cronyn , 890 F.2d 1158, 1161 (Fed. Cir. 1989) ; Acceleration Bay , 908 F.3d at 773. Samsung's contrary argument simply asks us to reweigh the evidence in its favor. We decline to do so. Regents of Univ. of California v. Broad Inst., Inc. , 903 F.3d 1286, 1294 (Fed. Cir. 2018) ("It is not our role to ask whether substantial evidence supports fact-findings not made by the Board, but instead whether such evidence supports the findings that were in fact made. Here, we conclude that it does.").
Samsung's remaining arguments regarding the website are similarly unavailing. For example, Samsung insists that the Board required it to show that the JCT-VC website was "advertised outside of" JCT-VC. Appellant's Br. 47. But the Board imposed no such requirement. Instead, it pointed to the lack of advertising as an example, among others, of Samsung's failure to show that those outside of the JCT-VC knew about the JCT-VC website. Final Written Decision , 2018 WL 1940480 at *6. Samsung also argues that the Board required it to show that the WD4 reference was indexed. Again, the Board did no such thing. Rather, it noted that the structure of the website, which organized content by meeting and lacked a way to search by subject matter, meant that a person would only find the WD4 reference if they knew where to look. The Board's ensuing analysis, as explained above, is consistent with our precedent and factually supported. Samsung also faults the Board for changing its view from its institution decisions to its final written decisions. We have repeatedly said, however, that "[t]he Board is free to change its view of the merits after further development of the record, and should do so if convinced its initial inclinations were wrong." Trivascular, Inc. v. Samuels , 812 F.3d 1056, 1068 (Fed. Cir. 2016) ; see also In re Magnum Oil Tools Int'l, Ltd. , 829 F.3d 1364, 1377 (Fed. Cir. 2016) ("[T]he Board has an obligation to assess the [validity] question anew after trial based on the totality of the record."). The Board did just that.
We reach the same conclusion with respect to the MPEG website for the same reasons stated above. Appellant's Br. 56 (noting the similarities between the JCT-VC website and the MPEG website). In particular, the Board's conclusion that a person of ordinary skill in the art would not have been able to find the WD4 reference on the MPEG website, even after exercising reasonable diligence, is supported by substantial evidence. Final Written Decision , 2018 WL 1940480 at *7-9.
3. JCT-VC Listserv
Samsung last argues that the WD4 reference was publicly accessible because Mr. Bross emailed it to the JCT-VC listserv. On this point, Mr. Bross testified that the listserv included 254 JCT-VC members and other "interested individuals." J.A. 7947-48. Mr. Bross further testified that "any person could subscribe to the JCT-VC reflector by requesting a subscription at the JCT-VC reflector management site" and "anyone with a valid e-mail address requesting subscription was typically approved." J.A. 7948.
The Board credited this evidence, at least in part. For example, the Board found that Mr. Bross emailed a link to the WD4 reference to the listserv in October 2011. Final Written Decision , 2018 WL 1940480 at *9. It also agreed that the listserv may have included "others who may have opted into the [listserv]." Id. The Board, however, found this evidence insufficient to establish public accessibility because it could not determine that "those 254 individuals represented a significant portion of those interested and skilled in *1374the art." Id. Instead, the Board viewed the email as a "limited distribution" that did not show the work was "generally disseminated to persons interested and ordinarily skilled in the art." Id.
Samsung argues that the Board erred by confusing access with accessibility. We agree. Our cases have consistently held that the standard for public accessibility is whether a person of ordinary skill in the art could , after exercising reasonable diligence, access a reference. Jazz , 895 F.3d at 1355-56 ("A reference is considered publicly accessible upon a satisfactory showing that such document has been disseminated or otherwise made available to the extent that persons interested and ordinarily skilled in the subject matter or art, exercising reasonable diligence, can locate it. If accessibility is proved, there is no requirement to show that particular members of the public actually received the information." (internal quotation marks and citations omitted)); In re Lister , 583 F.3d 1307, 1314 (Fed. Cir. 2009) ("[O]ur cases have held that once accessibility is shown, it is unnecessary to show that anyone actually inspected the reference."); Constant v. Advanced Micro-Devices, Inc. , 848 F.2d 1560, 1569 (Fed. Cir. 1988) ("If accessibility is proved, there is no requirement to show that particular members of the public actually received the information."). Thus, a petitioner need not establish that specific persons actually accessed or received a work to show that the work was publicly accessible.
The Board departed from this well-established principle by repeatedly faulting Samsung for not proving that the WD4 reference was "generally" or "widely" disseminated. Final Written Decision , 2018 WL 1940480 at *9. In fact, a limited distribution can make a work publicly accessible under certain circumstances. See, e.g. , GoPro , 908 F.3d at 694. But the Board's analysis stopped short of considering whether those circumstances were present here. The Board also faulted Samsung for failing to show that the email recipients "represented a significant portion of those interested and skilled in the art." Final Written Decision , 2018 WL 1940480 at *9. That was not Samsung's burden. See, e.g. , Mass. Inst. of Tech. v. AB Fortia , 774 F.2d 1104 (Fed. Cir. 1985) (distribution to six conference attendees made work publicly accessible). The Board's decision to reject Samsung's evidence because it did not establish that enough interested and ordinarily skilled artisans actually obtained the WD4 reference was therefore erroneous.
Infobridge offers no substantive response on this point. Infobridge actually agrees that the relevant question here is "the extent that [the WD4 reference] could b[e] located by an interested person exercising reasonable diligence." Appellee's Br. 28 (emphasis added). Infobridge simply argues that the Board applied this test here. We are not persuaded. As noted above, the only rationales offered by the Board on this issue were legally erroneous.4 In re NuVasive , 842 F.3d 1376, 1383 (Fed. Cir. 2016) ("The PTAB's own explanation must suffice for us to see that the agency has done its job ....").
Rather than requiring Samsung to prove that persons of ordinary skill actually received the listserv email, the Board should have considered whether Samsung's evidence established that an ordinarily skilled *1375artisan could have accessed the WD4 reference, after exercising reasonable diligence, based on the listserv email. This might include examining whether a person of ordinary skill, exercising reasonable diligence, would have joined the listserv. See, e.g. , GoPro , 908 F.3d at 694. It also might include considering the circumstances of the email itself, for example why the email was sent and whether it was covered by an expectation of confidentiality. See, e.g. , Mass. Inst. of Tech. , 774 F.2d at 1109. Because the record is not clear on these factual questions, we decline to resolve them in the first instance on appeal. See, e.g. , Medtronic , 891 F.3d at 1381-83 (remanding for the Board to consider the public accessibility question). In particular, we are reluctant to assume that an email among potential collaborators should be treated the same as a public disclosure without clear findings by the Board. Accordingly, we vacate the Board's finding that Mr. Bross's email did not make the WD4 reference publicly accessible and remand so that the Board can consider this issue after applying the correct legal standard.
III. CONCLUSION
We have considered the parties' remaining arguments and find them unpersuasive. Ultimately, the Board applied an erroneous legal standard in concluding that the listserv email did not make the WD4 reference publicly accessible. We therefore vacate the Board's findings on this point and remand for further proceedings consistent with this opinion.
VACATED AND REMANDED
COSTS
No costs.