Microsoft’s courtroom win over a patent challenge to its Azure networking stack has been short‑lived: the U.S. Court of Appeals for the Federal Circuit has reversed a Western District of Texas ruling that had struck Exafer Ltd.’s damages testimony and then entered summary judgment of no damages for Microsoft, returning the dispute to the lower court for further proceedings. The appeals court concluded the district court had misapplied the gatekeeping standards when excluding Exafer’s damages evidence and improperly entered judgment on the ground that Exafer could not present a non‑zero reasonable‑royalty theory. This reversal revives a years‑old patent fight over Azure’s Virtual Filtering Platform and SmartNIC features and signals another battleground moment in how courts assess expert damages opinions in high‑stakes tech litigation. (govinfo.gov)
Exafer Ltd., a small technology company, sued Microsoft in 2019 asserting that features of Microsoft Azure—specifically Azure SmartNICs and the Virtual Filtering Platform (VFP) FastPath packet‑processing software—infringed U.S. Patent Nos. 8,325,733 and 8,971,335. Exafer’s damages case relied on expert testimony that attempted to translate the technical benefit Microsoft obtained from those accused features into a monetary remedy. Microsoft moved to exclude Exafer’s damages expert and, after the district court sustained the exclusion, moved for summary judgment on the ground that Exafer had no admissible evidence supporting a non‑zero reasonable royalty. The district court granted Microsoft’s motion and entered final judgment for Microsoft. (govinfo.gov)
On appeal, the Federal Circuit—sitting to review the exclusion and the summary‑judgment decision—reversed the lower court’s rulings, finding that the court below had abused its gatekeeping discretion by excluding the damages expert and that the remaining record supported at least a triable issue about whether Exafer was entitled to a non‑zero royalty. The panel opinion—authored by Chief Judge Kimberly A. Moore—explained that Exafer’s damages theory possessed the requisite methodological soundness and was tied to the disputed technology in a manner that made the opinion admissible and suitable for jury consideration. The appeals court therefore vacated the summary judgment and remanded for further proceedings.
From the district court’s perspective, excluding the expert and granting summary judgment was rigorous gatekeeping: if the plaintiff’s sole damages path is built on a legally inadequate base, there is nothing for a jury to decide. That approach, in practical terms, converts the Daubert gatekeeping function into a near‑dispositive procedural weapon when damages evidence is limited and discovery choices narrow the record. The Western District concluded that Exafer had not supplemented its Rule 26 disclosures and had failed to develop alternative damages evidence; the court therefore treated exclusion as dispositive. (govinfo.gov)
Key elements of the Federal Circuit’s reasoning included:
The district court struck one key expert (Justin Blok) because he appeared to rely on an unaccused royalty base (virtual machine revenues for features Exafer had not expressly accused). That exclusion left Exafer without the primary numerical valuation it had planned to present at trial, which led the district court to find no triable issue on damages. The Federal Circuit found that the expert’s methodology was not so divorced from the accused features as to warrant wholesale exclusion. (govinfo.gov)
This stance reinforces scholarly and doctrinal caution against turning Daubert into a tool for resolving core factual disputes that juries are meant to resolve. It also incentivizes careful expert work: when experts explicitly connect their models to the accused technology and disclose assumptions and sources, their opinions are more likely to survive heightened scrutiny.
There is also a policy concern: reversing exclusions can prolong expensive litigation, especially where the excluded testimony was the patentee’s only numerical path to damages. Defendants will argue that allowing such testimony back in unfairly burdens them with protracted expert battles and potentially opens the door to speculative awards. Practitioners should therefore treat the Exafer decision as a narrow win for plaintiffs’ expert evidence, not a carte blanche to submit tenuous damages models. (govinfo.gov)
For Microsoft and other cloud vendors, the practical exposure remains conditional. The ultimate damages risk will hinge on whether Exafer ultimately proves infringement and on how a jury or the district judge resolves apportionment and the translation of technical benefits into monetary compensation. For the broader IP community, the ruling is another incremental pivot in the evolving jurisprudence on damages experts: a signal that methodological soundness and factual fit can, in close cases, defeat wholesale exclusion and premature summary disposition.
In short: the Federal Circuit restored a contested damages dispute to the trial‑level arena, reaffirming that in complex, technical patent cases the line between admissible expert opinion and inadmissible speculation is a quintessentially factual question best resolved with a full record—not by abbreviated pretrial shortcut. The Azure patent fight is back in play, and the outcome will test both the durability of Exafer’s damages approach and Microsoft’s ability to confine damages to legally appropriate bounds. (govinfo.gov)
Source: Bloomberg Law News Microsoft Azure Patent Victory Reversed by Federal Appeals Court
Background / Overview
Exafer Ltd., a small technology company, sued Microsoft in 2019 asserting that features of Microsoft Azure—specifically Azure SmartNICs and the Virtual Filtering Platform (VFP) FastPath packet‑processing software—infringed U.S. Patent Nos. 8,325,733 and 8,971,335. Exafer’s damages case relied on expert testimony that attempted to translate the technical benefit Microsoft obtained from those accused features into a monetary remedy. Microsoft moved to exclude Exafer’s damages expert and, after the district court sustained the exclusion, moved for summary judgment on the ground that Exafer had no admissible evidence supporting a non‑zero reasonable royalty. The district court granted Microsoft’s motion and entered final judgment for Microsoft. (govinfo.gov)On appeal, the Federal Circuit—sitting to review the exclusion and the summary‑judgment decision—reversed the lower court’s rulings, finding that the court below had abused its gatekeeping discretion by excluding the damages expert and that the remaining record supported at least a triable issue about whether Exafer was entitled to a non‑zero royalty. The panel opinion—authored by Chief Judge Kimberly A. Moore—explained that Exafer’s damages theory possessed the requisite methodological soundness and was tied to the disputed technology in a manner that made the opinion admissible and suitable for jury consideration. The appeals court therefore vacated the summary judgment and remanded for further proceedings.
What the district court did — and why the Federal Circuit disagreed
The district‑court rulings in short
The Western District of Texas followed a familiar two‑step path in patent damages disputes: first it applied its gatekeeper role under Daubert and Rule 702 to test the admissibility of Exafer’s damages expert reports; then, after excluding the plaintiff’s damages expert in whole (or in relevant part), it concluded the patentee lacked any admissible evidence that would support a non‑zero reasonable royalty and entered summary judgment for Microsoft. The court’s central concern was that Exafer’s damages model had used an improper royalty base—relying on revenues tied to virtual machines and other features that the court deemed not properly accused—so the expert’s numbers could not be connected to the legally compensable acts of infringement. (govinfo.gov)From the district court’s perspective, excluding the expert and granting summary judgment was rigorous gatekeeping: if the plaintiff’s sole damages path is built on a legally inadequate base, there is nothing for a jury to decide. That approach, in practical terms, converts the Daubert gatekeeping function into a near‑dispositive procedural weapon when damages evidence is limited and discovery choices narrow the record. The Western District concluded that Exafer had not supplemented its Rule 26 disclosures and had failed to develop alternative damages evidence; the court therefore treated exclusion as dispositive. (govinfo.gov)
Why the Federal Circuit reversed
The Federal Circuit’s reversal is significant because it rejected the district court’s use of exclusion and summary judgment as a shortcut to dispose of the case where disputed factual and methodological issues remained. Chief Judge Moore’s opinion emphasized that an appellate court must be able to meaningfully review why the district court acted as it did and whether the expert’s methodology actually failed basic reliability tests.Key elements of the Federal Circuit’s reasoning included:
- The court found that Exafer’s damages theory was sufficiently tethered to the accused technology (VFP FastPath and SmartNIC acceleration) to pass the admissibility threshold; that is, the expert did not base a royalty on unrelated features in a way that made the opinion irrelevant as a matter of law. The opinion stressed that fit—the requirement that expert evidence “fit” the facts of the case and assist the trier of fact—was present on the record.
- The appeals court concluded that the district court had not sufficiently considered the methodological underpinnings of the damages model—what the Federal Circuit described as the expert’s reliance on company documents, technical testimony about CPU‑cycle savings, and other record evidence tying the economic benefit to the accused features. Excluding an expert outright requires showing the model is not only disputed but fundamentally unsound; the Federal Circuit found that showing was not made.
- Finally, the panel stressed that exclusion of primary damages evidence often should not be treated as an automatic route to summary judgment. If plausible alternative evidence about damages remains in the record or if reasonable jurors could differ on the expert’s assumptions, the proper remedy is remand for trial or for further evidentiary development, not immediate dismissal. The Federal Circuit vacated the summary judgment and sent the case back.
The technical and damages issues at the center of the dispute
The technology: VFP FastPath and SmartNIC acceleration
At the heart of the case are components used by Microsoft to speed packet processing in Azure:- Virtual Filtering Platform (VFP) FastPath — a software path inside Azure’s virtual networking stack designed to accelerate packet processing without consuming the same CPU cycles as the general‑purpose path.
- Azure SmartNICs — specialized network interface cards that offload networking tasks from host CPUs, further reducing per‑packet CPU work.
The damages model: cost‑savings and per‑hour royalty approaches
Exafer’s damages expert proposed an economic model that (in summary) calculated a running royalty based on unitized usage metrics tied to virtual machines and internal revenue figures or proxies for the economic benefits Microsoft derived from freeing CPU cycles. The expert’s approach was, in part, a cost‑savings methodology: tying alleged CPU efficiencies to reduced infrastructure cost or opportunity value that Microsoft realized by operating Azure with the accused features.The district court struck one key expert (Justin Blok) because he appeared to rely on an unaccused royalty base (virtual machine revenues for features Exafer had not expressly accused). That exclusion left Exafer without the primary numerical valuation it had planned to present at trial, which led the district court to find no triable issue on damages. The Federal Circuit found that the expert’s methodology was not so divorced from the accused features as to warrant wholesale exclusion. (govinfo.gov)
Legal and industry implications
For patent damages practice: a recalibration of gatekeeping
This decision is a reminder that courts must apply Rule 702 and Daubert as gatekeeping, not gate‑shutting. The Federal Circuit’s reversal signals that appellate oversight remains alive on questions where courts exclude complex economic testimony in patent cases. Practitioners should take away three central lessons:- Methodology matters—but so does context. An expert’s reliance on company documents, technical testimony, and reasonable bridging assumptions can be enough to survive exclusion even if the precise monetary valuation is contested.
- Discovery and disclosure are critical. While the Federal Circuit reversed here, the district court’s frustration with the plaintiff’s limited non‑expert damages disclosures remains a caution: failure to produce a robust Rule 26 damages computation invites tactical risk.
- Courts should be wary of deciding damages questions prematurely. If alternative record evidence or contested factual issues persist, exclusion plus summary judgment may be an abuse of discretion. Expect more appeals where courts try to dispose of complex damages theories without a full trial record. (govinfo.gov)
For Microsoft and the cloud industry
The practical stakes for Microsoft are straightforward but not trivial. A reversal and remand mean:- Microsoft may face a new damages determination or a future trial where Exafer’s experts (and Microsoft’s rebuttal experts) square off in front of a jury or the district judge at trial.
- Although the Federal Circuit did not rule on infringement or validity, the decision reintroduces the remedy question as live: if Exafer ultimately prevails on infringement, Microsoft may be liable for meaningful damages tied to Azure’s vast scale—unless later proceedings limit the royalty base or apportionment. That creates potential downstream exposure that could materially affect settlements, insurance considerations, and public‑company risk disclosures.
- For the cloud industry broadly, the ruling underlines that networking‑level innovations—including offload features like SmartNICs—can be the subject of high‑value patent disputes. Cloud providers must therefore maintain robust freedom‑to‑operate analyses and defenses, and customers should be mindful of litigation risk in critical infrastructure components.
The precedent context: where this fits in the Federal Circuit’s recent damages jurisprudence
The Federal Circuit has been deeply engaged with how damages experts should be vetted. Recent Federal Circuit decisions (including en banc rulings in other high‑profile cases) have pushed and pulled on the admissibility standard, sometimes tightening the scrutiny on how experts use prior licenses and how experts apportion revenue bases. The Exafer reversal shows a countervailing impulse: while courts must exclude unreliable methods, they should not substitute their own view of the facts for the jury’s when the expert demonstrates a coherent methodology tied to the asserted invention. Expect further doctrinal refinement as panels grapple with the line between reliability and factual dispute.Practical consequences and next steps in the case
What happens procedurally now
- The Federal Circuit’s mandate vacates the summary judgment of no damages and sends the case back to the district court for proceedings consistent with the appellate opinion. That will likely include permitting the previously excluded expert testimony (subject to any narrowing the district court may try to craft on remand), allowing Microsoft to present rebuttal evidence, and potentially setting a schedule for trial or a damages‑only hearing.
- The appeals court’s reversal does not decide liability (infringement) or patent validity. Those issues may still be contested at trial or could themselves be the subject of separate motions or appeals. The outcome of liability and validity will frame any future damages calculus. (govinfo.gov)
Financial and strategic considerations for Microsoft
- The reversal increases uncertainty: even if Exafer is commercially small, a finding of a non‑zero reasonable royalty applied to massive Azure usage could create substantive monetary exposure or prompt a settlement.
- Microsoft’s litigation posture—if it continues an aggressive defense—will likely emphasize technical distinctions, apportionment, and the limits of reasonable royalty theory. Microsoft will also continue to argue that any alleged benefits do not translate neatly into direct financial gains that a reasonable‑royalty award could capture.
- From a risk‑management standpoint, Microsoft may choose to renew settlement discussions, seek further interlocutory review on particular legal questions, or litigate to trial depending on the evolving risk profile.
Critical analysis: strengths and weaknesses of the Federal Circuit’s result
Strengths — restoring the factfinder’s role and emphasizing methodological review
The Federal Circuit’s reversal is defensible on sound judicial and evidentiary principles. When damages models rest on contested assumptions but are undergirded by real company data, technical testimony, and an explicable method for converting technical benefit into economic value, courts should be reluctant to categorically exclude the evidence. The appeals court’s approach vindicates the role of the jury (or trial court at a full evidentiary hearing) to weigh conflicting expert testimony rather than allowing a preliminary gatekeeping decision to become a dispositive merits judgment.This stance reinforces scholarly and doctrinal caution against turning Daubert into a tool for resolving core factual disputes that juries are meant to resolve. It also incentivizes careful expert work: when experts explicitly connect their models to the accused technology and disclose assumptions and sources, their opinions are more likely to survive heightened scrutiny.
Risks and weaknesses — the procedural friction and discovery shortfalls
That said, the Federal Circuit’s decision may be criticized for downplaying legitimate procedural concerns raised by the district court. The trial court faulted Exafer for failing to provide Rule 26 disclosures and for refusing to reopen discovery when the primary damages expert was excluded. If a patentee deliberately withholds a damages theory until late in the case, courts and defendants have reason to be frustrated—and remedies for that conduct (including exclusion) can be appropriate.There is also a policy concern: reversing exclusions can prolong expensive litigation, especially where the excluded testimony was the patentee’s only numerical path to damages. Defendants will argue that allowing such testimony back in unfairly burdens them with protracted expert battles and potentially opens the door to speculative awards. Practitioners should therefore treat the Exafer decision as a narrow win for plaintiffs’ expert evidence, not a carte blanche to submit tenuous damages models. (govinfo.gov)
Takeaways for practitioners, in‑house counsel, and enterprise buyers
- Patent plaintiffs: Make sure damages theories are disclosed early and transparently. Tie economic models explicitly to technical facts and to produced documents. Demonstrating fit early reduces the risk that exclusion will end the case.
- Patent defendants (and cloud providers): Preserve targeted Daubert and motion practice options, and prepare to litigate the assumptions underlying damages models at trial. Consider narrowing liability exposure through patent eligibility, claim construction, or targeted motions on apportionment topics.
- In‑house legal and procurement teams: Monitor litigation involving key infrastructure components (SmartNICs, offload engines, virtual networking paths). Even a small patentee can create outsized commercial leverage against hyperscalers via patent litigation.
- Trial counsel and experts: Draft damages reports that transparently show the economic chain linking a patented feature to calculable financial impact. Rely on contemporaneous internal documents and explicit apportionment to the smallest asserted inventive contribution. Avoid leaps from technical value to companywide revenue without intermediate, documented steps. (govinfo.gov)
Final assessment — what to watch next
This Federal Circuit reversal reshapes the immediate litigation landscape: expect renewed briefing, potential remand scheduling, and the reintroduction of expert testimony that the district court had previously excluded. Beyond the Exafer case itself, watch the Federal Circuit’s ongoing calibration of expert admissibility in patent cases—judges will continue to squint between two poles: guarding against unreliable, speculative testimony; and ensuring juries can hear expert‑driven economic narratives tethered to technical facts.For Microsoft and other cloud vendors, the practical exposure remains conditional. The ultimate damages risk will hinge on whether Exafer ultimately proves infringement and on how a jury or the district judge resolves apportionment and the translation of technical benefits into monetary compensation. For the broader IP community, the ruling is another incremental pivot in the evolving jurisprudence on damages experts: a signal that methodological soundness and factual fit can, in close cases, defeat wholesale exclusion and premature summary disposition.
In short: the Federal Circuit restored a contested damages dispute to the trial‑level arena, reaffirming that in complex, technical patent cases the line between admissible expert opinion and inadmissible speculation is a quintessentially factual question best resolved with a full record—not by abbreviated pretrial shortcut. The Azure patent fight is back in play, and the outcome will test both the durability of Exafer’s damages approach and Microsoft’s ability to confine damages to legally appropriate bounds. (govinfo.gov)
Source: Bloomberg Law News Microsoft Azure Patent Victory Reversed by Federal Appeals Court
