Egenera, Inc. v. Cisco Systems, Inc.: Federal Circuit Affirms Noninfringement Verdict on All Asserted Claims

In a resounding victory for Cisco Systems, the Federal Circuit affirmed a district court’s ruling that Cisco’s Unified Computing System (“UCS”) does not infringe U.S. Patent No. 7,231,430, held by Egenera, Inc. The decision, issued on July 7, 2025, capped nearly a decade of litigation over Egenera’s patented “virtual processing area network” technology.

The ruling affirms summary judgment of noninfringement on claims 1 and 5, and a jury verdict of noninfringement on claims 3 and 7. Egenera’s post-trial motions for judgment as a matter of law (JMOL) and a new trial were also rejected. Writing for a unanimous panel, Judge Stark carefully dissected each of Egenera’s arguments and concluded that they either failed on the facts or had been waived procedurally.

As Charles Gideon Korrell notes, this case exemplifies the perils of failing to preserve claim construction arguments and demonstrates the rigorous evidentiary standards patent plaintiffs must meet at both the summary judgment and trial stages.


Background: The ’430 Patent and Cisco’s UCS

Egenera’s ’430 patent claims a platform for deploying virtual networks by configuring processors and internal communications via software, rather than physically rewiring connections. Claims 1 and 3 are system claims, while claims 5 and 7 are method claims. Central to all four is the concept of virtual deployment and the emulation of Ethernet functionality.

Cisco’s UCS, according to Egenera, performed similar virtualized management through software-defined networking, using NICs and virtual NICs (vNICs) to group servers into VLANs. But Cisco countered that the claimed Ethernet emulation and processor behavior were not met.


Summary Judgment on Claims 1 and 5: Failure to Emulate Ethernet Functionality at the CPU Level

The district court granted Cisco’s motion for summary judgment on claims 1 and 5 after construing “computer processor” to mean “CPU” and determining that Egenera had not shown the CPUs themselves “emulate Ethernet functionality.”

Egenera argued that use of Ethernet by CPUs satisfied the claim limitation. But the court held that use ≠ emulation and that Egenera failed to offer evidence showing that UCS’s CPUs performed the required emulation. Instead, the Ethernet functionality resided in the NICs—separate from the CPUs.

On appeal, Egenera stuck to the argument that the evidence supported a finding of emulation but notably did not challenge the underlying claim construction of “emulate Ethernet functionality.” As the Federal Circuit emphasized:

“Egenera has not, even on appeal, argued that claim construction is necessary, let alone presented a meaningful claim construction analysis of ‘emulate.’”

This procedural misstep proved fatal. The court reiterated that a party cannot raise new claim construction issues for the first time on appeal, especially when it consistently characterized the dispute as factual in both the district court and its appellate briefing.

This part of the ruling is a cautionary tale. As Charles Gideon Korrell emphasizes, practitioners should be wary of framing claim scope disputes as factual disagreements. The failure to preserve and present a proper construction argument can doom an entire case.


Trial on Claims 3 and 7: Topology Limitation Undermines Infringement

Claims 3 and 7 required that the processors be “programmed” to establish a specific “virtual local area network topology.” At trial, Cisco presented three alternative noninfringement theories: lack of (1) the topology limitation, (2) modifying messages, and (3) extracting and identifying messages. The jury returned a general verdict of noninfringement.

Under black-letter law (see i4i Ltd. P’ship v. Microsoft Corp.), a general verdict must be affirmed if supported by substantial evidence on any theory. The Federal Circuit thus focused on the “topology” issue.

Cisco’s evidence showed that its UCS established network topologies at the NIC level—not at the processor level as required by the claims. Egenera countered with arguments that software loaded into the CPU satisfied the programming requirement, but the jury was free to disbelieve this theory. As the panel explained, “drawing all reasonable inferences in favor of Cisco,” substantial evidence supported the verdict.

Egenera also argued that the district court erred by reading an “intent” requirement into the claim, but the Federal Circuit rejected this characterization, finding that the court merely summarized what Cisco witnesses had testified to: that CPUs were not programmed to establish topology.

Again, Charles Gideon Korrell observes that the outcome turned on careful adherence to the burden of proof and deference to the jury’s prerogative to weigh competing expert testimony.


Denial of New Trial: Procedural Missteps and Lack of Prejudice

Egenera’s motion for a new trial fared no better. The district court had broad discretion under First Circuit law, and Egenera could not demonstrate that any alleged error rose to the level of a miscarriage of justice.

Among Egenera’s arguments:

  1. Verdict Against the Weight of Evidence: Rejected due to relevance of Cisco’s evidence and proper jury instructions.
  2. Curative Instruction: Egenera claimed that a voir dire statement by the judge equating infringement with copying required correction. But it waited 10 days and failed to object contemporaneously, and the court’s final instructions correctly outlined the law.
  3. Omitted Jury Instruction on Later Patents: Egenera requested an instruction that a product can infringe even if it is covered by a subsequent patent. While substantively correct, the omission was found not to be prejudicial given the adequacy of the overall instructions.
  4. Improper Expert Testimony from Lay Witnesses: The district court found that Egenera had waived this argument by failing to make timely and specific objections at trial. The Federal Circuit upheld this decision, citing Fusco v. GM Corp. and First Circuit precedent allowing trial judges to enforce procedural forfeiture.
  5. Improper Closing Arguments: Cisco was accused of making veiled references to Egenera’s funding and characterizing it as a “bad bet by venture capitalists.” But these arguments did not literally violate pretrial orders, and Egenera failed to object contemporaneously. The court found no plain error.

Charles Gideon Korrell emphasizes that the procedural posture of these objections matters deeply. Litigants who fail to timely object cannot later complain that errors warrant a new trial, especially where the court provides accurate and complete jury instructions.


Key Takeaways

  1. Preserve Claim Construction Arguments Early and Clearly: Egenera’s failure to raise and brief the meaning of “emulate Ethernet functionality” effectively waived the issue. The Federal Circuit will not rescue an underdeveloped claim construction theory after the fact.
  2. Substantial Evidence Deference is High: On a general verdict, a plaintiff must show the absence of support for every alternative noninfringement theory. Egenera failed to overcome the substantial evidence supporting Cisco’s topology argument.
  3. Procedural Discipline Matters: Courts require timely objections to preserve error. Pretrial motions in limine and vague references to disputed issues are not enough.
  4. Appeals Framed as Factual Disputes Are Difficult to Win: Egenera characterized its appeal as a sufficiency-of-evidence question, not a legal error in construction or instruction, which narrowed its ability to persuade the appellate court.

As Charles Gideon Korrell observes, this decision reinforces the Federal Circuit’s institutional emphasis on claim construction clarity, trial discipline, and respect for jury fact-finding. Cisco’s clean win on all claims and all fronts is a blueprint for defending against complex system and method patents where the accused functionality is modular and distributed.

By Charles Gideon Korrell