STARK, Circuit Judge.

Bitmanagement Software GmBH (“Bitmanagement”) appeals the damages judgment it received in the United States Court of Federal Claims as compensation for copyright infringement committed by the United States Navy (“Navy”). The Court of Federal Claims determined that a hypothetical negotiation between the parties would have resulted in a license to Bitmanagement’s software at a cost of $154,400, which it ordered the Navy to pay Bitmanagement. Finding no abuse of discretion, we affirm.

I Bitmanagement “develops software for rendering three-dimensional graphics.” Bitmanagement Software GmBH v. United States, 989 F.3d 938, 941 (Fed. Cir. 2021) (“Bit II”). At issue in this case is Bitmanagement’s “BS Contact Geo” software, which is useful in conjunction with a Navy software platform called SPIDERS 3D. Used together, BS Contact Geo and SPIDERS 3D create a virtual reality environment in which Navy employees can view images of Naval installations, bases, and facilities in three dimensions.

In 2008, Bitmanagement, through a third-party reseller, provided the Navy with 100 “seat licenses” (also known as “PC licenses”) for BS Contact Geo. A “seat license” allows only “one installation of the software onto one computer per license.” Id. Hence, with 100 seat licenses, the Navy was permitted to install BS Contact Geo on 100 specific computers.

In 2012, the parties switched the Navy from seat licenses to a “floating license.” A “floating license” allows the licensed user to access the software on a wide range of computers, so long as measures are taken to limit the simultaneous usage of the program to only the number of individual users permitted by the license. See id. at 943. The Navy’s first floating license with Bitmanagement allowed the Navy to install BS Contact Geo on as many computers as it wished but capped the maximum usage at 20 users at any one time. To ensure the Navy adhered to the cap in the floating license, Bitmanagement and the Navy agreed that the Navy would use a tracking application, called “Flexera.” As we explained when this case was before us previously:

Flexera is a server-based program used to limit the number of simultaneous users of a “Flexera enabled” – or “FlexWrapped” – software based on the number of available licenses. When a user opens a FlexWrapped program, the program alerts the Flexera tracking server that the program is in use. The FlexWrapped program sends a similar alert when the program is no longer in use. The Flexera license manager thus limits the number of users of FlexWrapped software to the number of licenses that a user owns.

Id. In other words, Flexera “would limit the number of simultaneous users” by “allowing the program [i.e., BS Contact Geo] to run only if the number of persons using the program is less than the number of available licenses.” J.A. 3-4.

In 2013, the Navy began to deploy BS Contact Geo widely throughout the organization. Eventually, it was accessible on more than 429,000 Navy computers. Flexera, however, failed to operate as intended; it did not restrict the number of simultaneous users of BS Contact Geo to the number of licenses the Navy had purchased from Bitmanagement.

In July 2016, Bitmanagement sued the United States (“government”) in the Court of Federal Claims, alleging that the Navy had infringed its copyright on BS Contact Geo. In preparation for trial, Bitmanagement moved in limine to exclude the government’s damages expert, David Kennedy, arguing he had used the wrong legal test by calculating damages based on the amount of usage of BS Contact Geo by Navy personnel rather than the number of copies the Navy had made of the software. The Court of Federal Claims granted the motion and excluded Mr. Kennedy’s testimony. Following trial, the court granted judgment for the government, finding that the Navy had no liability for copyright infringement.

Bitmanagement appealed. We agreed with the Court of Federal Claims that the Navy had an implied license to make copies of BS Contact Geo. But we further held that the Navy’s agreement to use Flexera to limit the number of simultaneous users of BS Contact Geo was a material condition of the implied license, a condition the Navy had breached. We explained that “the Navy’s failure to abide by the Flexera condition of that license renders its copying of the program copyright infringement.” Bit II, 989 F.3d at 951.

We remanded the case to the Court of Federal Claims for it to calculate the damages the government owed Bitmanagement. In doing so, we stated the following in footnote 5 of our opinion:

Because Bitmanagement’s action is against the government, it is entitled only to “reasonable and entire compensation as damages . . . , including the minimum statutory damages as set forth in section 504(c) of title 17, United States Code.” 28 U.S.C. § 1498(b). . . . Contrary to Bitmanagement’s argument, see J.A. 10002 ¶ 5, it is not entitled to recover the cost of a seat license for each installation. If Bitmanagement chooses not to pursue statutory damages, the proper measure of damages shall be determined by the Navy’s actual usage of BS Contact Geo in excess of the limited usage contemplated by the parties’ implied license. That analysis should take the form of a hypothetical negotiation. See Gaylord v. United States, 777 F.3d 1363, 1368-72 (Fed. Cir. 2015) [(“Gaylord III”)]; Gaylord [v. United States], 678 F.3d [1339,] 1342-45 [(Fed. Cir. 2012) (“Gaylord II”) 1]. As the party who breached the Flexera requirement in the implied license, the Navy bears the burden of proving its actual usage of the BS Contact Geo software and the extent to which any of it fell within the bounds of any existing license. Id. at 951 n.5.

On remand, the Court of Federal Claims determined that footnote 5 was part of our mandate, leaving the trial court no choice but to follow it. J.A. 20-21 (“[T]he Federal Circuit has mandated that Plaintiff is not entitled to damages based on the ‘cost of a seat license for each installation.’”). The court ordered the parties to submit supplemental briefing on damages, granted the government’s motion to reconsider its prior exclusion of Mr. Kennedy, and reopened the record to permit him to testify. The court also offered Bitmanagement the opportunity to present additional damages-related evidence of its own, though Bitmanagement declined to do so.

Bitmanagement argues that the Court of Federal Claims erred by awarding damages based on the number of copies of BS Contact Geo that were used by the Navy, rather than for every copy of the software the Navy made. See, e.g., Open. Br. 2-3 (arguing that Copyright Act “requires compensation for each infringing copy, as the other circuits to consider the issue have uniformly concluded”); id. at 28 (“The Court of Federal Claims erroneously failed to award damages for every infringing copy that the Navy made of Bitmanagement’s software.”); id. at 32 (“The Court of Federal Claims misapplied copyright law by awarding damages only for BS Contact Geo copies that were accessed, rather than for every infringing copy that the Navy made.”). Bitmanagement contends it should receive a per-copy payment of $200 for each of the 429,567 copies made by the Navy, for a total amount of damages of $85,913,400. We disagree. The law does not compel such a result where the hypothetical negotiation would have proceeded on a primarily per-use (floating license) basis, and the trial court’s rejection of a per-copy approach based on the record developed in this case is not clearly erroneous.

The law does not require that every award of copyright damages be on a per-copy basis. Title 28 of the United States Code, § 1498(b) provides:

[W]henever the copyright in any work protected under the copyright laws of the United States shall be infringed by the United States, . . . the exclusive action which may be brought for such infringement shall be an action by the copyright owner against the United States in the Court of Federal Claims for the recovery of his reasonable and entire compensation as damages for such infringement, including the minimal statutory damages as set forth in [17 U.S.C. § 504(c)] . . . . 28 U.S.C. § 1498(b) (emphasis added). 2 We held in Gaylord II, 678 F.3d at 1343, that when the government is the infringer, and a claim is brought under § 1498(b), “the methods used to determine ‘actual damages’ under the copyright damages statute, 17 U.S.C. § 504, are appropriate for measuring the copyright owner’s loss.” Notably, § 504(b) requires the copyright owner to prove “the actual damages suffered by him or her as a result of the infringement” (emphasis added).

We further explained in Gaylord II that the “reasonable and entire compensation” provided for by § 1498(b) “entitles copyright owners to compensatory damages, . . . but not to non-compensatory damages.” 678 F.3d at 1343. We went on to observe that the focus for calculating damages is on “the copyright owner’s loss,” as opposed to the value obtained by the government. Id.; see also id. (“[C]ourts must determine just compensation for[*~1373–1379] On remand, the Court of Federal Claims faithfully and carefully applied the Gaylord II framework, ultimately finding – based on the record before it – that the parties would have agreed to a primarily usage-based licensing scheme. J.A. ^[1] (“The Federal Circuit directed this Court to look at the Gaylord line of cases as a guide.”); see also J.A. 21 (“[T]he essence of the damages’ inquiry is how much would the Navy agree to a license covering this usage.”). In our view, whether or not footnote 5 in Bit II is binding, the facts found by the Court of Federal Claims – none of which are clearly erroneous – permit no other conclusion than that the parties would have adopted a primarily usage-based approach in a hypothetical negotiation. ^[3] As required by Gaylord II, the trial court evaluated the totality of the evidence before it and made findings as to the value of a license the parties would have hypothetically negotiated at the time the Navy started to infringe. J.A. ^[1] (trial court explaining it “was tasked with determining damages taking the form of a hypothetical negotiation”). The court “look[ed] at the evidence presented by both sides to determine the fair market value of a license to which the parties would have agreed.” Gaylord II, 678 F.3d at 1343; see also Gaylord III, 777 F.3d at 1367-68 (noting need to use “objective considerations in the determination of a copyrighted work’s market value”) (internal quotation marks and citations omitted). Its factual findings are not clearly erroneous and its damages award is not, in any respect, an abuse of its discretion.

Bitmanagement’s arguments against the findings made by Court of Federal Claims do not demonstrate any clear error or abuse of discretion by the trial court. The court rationally based its decision on the evidence in the record.

Bitmanagement asserts that the record is devoid of evidence that it ever agreed to a per-use license. See Oral Argument at 7:16-59, available at https://oralarguments. cafc.uscourts.gov/default.aspx?fl=23-1506_09032024.mp3. To the contrary, the floating license agreement negotiated with the Navy was, as we have explained, a form of a per-use license.

Bitmanagement also suggests that the Court of Federal Claims failed to adequately consider the convenience benefit to the Navy of having Bitmanagement’s software accessible on nearly half a million computers. This is incorrect. After noting that it was “attracted to the ‘convenience factor’ argument,” the trial court added that Bitmanagement had “yet again, . . . failed to provide the cost of convenience in its damages’ calculations.” J.A. 19 n.17. Because Bitmanagement had not attempted to quantitatively value the convenience factor and had instead insisted on a per-copy damages theory, it gave the court no reliable way to account for the value of convenience to the Navy.

Bitmanagement also attempts to dissect the royalty base the Court of Federal Claims used, which was “429,567 copies of BS Contact Geo 8.001 with 597 unique users and 100 additional simultaneous-use licenses.” J.A. 17. Bitmanagement emphasizes that the Navy only produced usage logs for years two and three of the damages period, and only for usage of BS Contact Geo with SPIDERS 3D and not with any other platform. Open. Br. 53-58; see also J.A. 10087 (government admitting: “The Navy has not tracked whether Navy personnel have used BS Contact Geo to view X3D files outside the context of SPIDERS 3D, and, thus, cannot identify whether such uses occurred or each such use.”). These holes in the record were plugged in a reasonable manner. For year one usage, “the missing year,” the Court of Federal Claims used “the highest year of users, 224 users for September 2014 – August 2015,” rejecting Mr. Kennedy’s proposal to take an average of the year two and year three figures (which would have resulted in an even smaller damages award). J.A. 17. With respect to non-SPIDERS 3D usage, multiple witnesses testified that, outside the SPIDERS 3D platform, they were unaware of any need for BS Contact Geo, and Bitmanagement did not pursue further discovery that might have allowed it to quantify any potential non-SPIDERS 3D usage. See, e.g., J.A. 1866 (government witness testifying that he was “not aware of any other demand for BS Contact Geo” outside of SPIDERS 3D); J.A. 1905 (another government witness testifying to same); see also J.A. 2132 (same government witness testifying that “[p]rior to SPIDERS 3D, it would be a handful [of users]” who were interested in this type of 3D capability).

Finally, Bitmanagement points to language in Bit II to suggest that the Court of Federal Claims understated the damages Bitmanagement suffered. Bitmanagement highlights two sentences from our prior opinion:

This condition [i.e., the requirement to use Flexera to track usage] rendered reasonable the otherwise objectively unreasonable decision of Bitmanagement to allow the Navy to make unlimited copies of its commercial product. . . . This is one of the rare circumstances where the record as a whole reflects that the only feasible explanation for Bitmanagement allowing mass copying of its software, free of charge, was the use of Flexera at the time of copying. Bit II, 989 F.3d at 950. These statements explain our conclusion that the Flexera provision in the license agreement is an enforceable condition of the agreement between the parties – rendering the Navy’s conduct copyright infringement – and not “merely a covenant such that any grievance raised by Bitmanagement necessarily sounds in contract.” Id. These statements concern liability, not the amount of damages needed to appropriately compensate Bitmanagement for the Navy’s infringement. Nothing in the sentences excerpted above, or any other part of Bit II, constrained the trial court’s freedom to find, on the facts before it, whatever amount of damages was supported by the evidence.

Thus, again, we conclude that the Court of Federal Claims’ damages analysis is not clearly erroneous and the court did not abuse its discretion.

B Bitmanagement’s further procedural attacks on the Court of Federal Claims’ judgment are largely dependent on its contention, which we have now rejected, that a proper damages calculation must be based on each unauthorized copy the Navy made of the software, rather than the use the Navy made of those copies. Therefore, we may dispose of these arguments with little discussion.

Bitmanagement insists that the Court of Federal Claims erred by requiring it, instead of the Navy, to prove the amount of use the Navy made of BS Contact Geo. Pointing to footnote 5 in our prior opinion, Bitmanagement contends that the trial court was required to place this burden on the Navy, but failed to do so. See Bit II, 989 F.3d at 951 n.5 (“As the party who breached the Flexera requirement in the implied license, the Navy bears the burden of proving its actual usage of the BS Contact Geo software and the extent to which any of it fell within the bounds of any existing license.”). The government counters that Bitmanagement, as the party seeking to prevail on an infringement claim, bears the burden of proving its damages, notwithstanding whatever this court may have meant by footnote 5.

For at least two reasons, we find it unnecessary to determine which party should have borne the burden of showing the Navy’s usage. First, the Court of Federal Claims did, in fact, place this burden on the Navy. J.A. 15 (“[T]he Defendant has the burden of proof (according to the Federal Circuit) . . . .”). Therefore, the basic premise of Bitmanagement’s argument is unsupported. The court never shifted this burden to Bitmanagement. At most, the court merely mentioned – in the course of rejecting Mr. Kennedy’s estimate of the Navy’s use of BS Contact Geo during the first year of infringement – that Bitmanagement, like the government, had failed to produce evidence on this point. J.A. 17 (noting Bitmanagement “has not provided any calculations that would capture this year”). Bitmanagement’s evidentiary failings on numerous points were worthy of note, and the trial court emphasized this reality. ^[5] The Court of Federal Claims was obligated to make findings based on the record the parties did, and did not, create, and that is precisely what the court did. But at no point did the Court of Federal Claims shift the burden of proof to Bitmanagement on the issue of the amount of infringing use.

Second, the Navy did produce evidence of its use, indeed the best available evidence, including whatever usage logs it had in its possession and witnesses to testify about that usage. See, e.g., J.A. 16-17. “[I]f actual damages can not be ascertained with precision because the evidence available from the infringer is inadequate, damages may be estimated on the best available evidence, taking cognizance of the reason for the inadequacy of proof and resolving doubt against the infringer.” Sensonics, Inc. v. Aerosonic Corp., 81 F.3d 1566, 1572 (Fed. Cir. 1996) (emphasis added). The trial court proceeded in a manner consistent with this directive, which permitted it, in its role as factfinder, to credit the government’s best available