Background
PACT XPP Schweiz AG holds U.S. Patent No. 9,250,908, which relates to how processors in multiprocessor computer systems access data stored in cache memory. Modern processors rely on hierarchical memory systems — main memory holds all data, while smaller cache memories sit closer to the processor cores and provide faster access to frequently used data. The ‘908 patent claimed a specific approach to managing data consistency across multiple processor caches.
Intel petitioned for inter partes review (IPR), arguing that the ‘908 patent’s claims were obvious in view of two prior art references: Kabemoto, which described a multiprocessor cache architecture, and Bauman, which addressed the same cache-coherency problem using a known memory management technique. The Patent Trial and Appeal Board (PTAB) rejected Intel’s obviousness challenge, finding that Intel had not shown a sufficient motivation to combine the two references. Intel appealed.
The Court’s Holding
The Federal Circuit reversed. The court focused on the “known-technique” rationale for motivation to combine — a principle derived from the Supreme Court’s 2007 KSR decision. Under KSR, a motivation to combine two prior art references exists when one reference discloses a known technique that would predictably solve a problem identified in another reference, and both references address the same general problem.
The court found that the PTAB’s own factual findings established exactly this scenario: Kabemoto and Bauman both addressed the same cache-coherency problem in multiprocessor systems, and Bauman disclosed a technique — a specific type of cache structure — that was a known, established solution to that problem. Given those facts, the court held that the PTAB’s conclusion of no motivation to combine was internally inconsistent. The whole point of the known-technique rationale is that when a reference provides a known solution to a problem identified in another reference, a skilled artisan would naturally look to apply it. The Board’s reasoning effectively contradicted the logic of KSR. The case was remanded for the PTAB to complete the obviousness analysis.
Key Takeaways
- When the PTAB’s own findings establish that two prior art references address the same technical problem and one provides a known technique for solving it, those findings compel a conclusion of motivation to combine under the KSR known-technique rationale.
- A PTAB decision finding no motivation to combine while simultaneously acknowledging that the prior art references share the same problem and that a known solution exists is legally inconsistent and will not survive Federal Circuit review.
- The known-technique rationale does not require that the combination be the only or best solution — it requires only that the technique be known and predictably applicable to the problem.
- IPR petitioners challenging patents on cache memory, processor architecture, or similar computing technologies should frame their motivation-to-combine argument explicitly in known-technique terms when the prior art addresses the same problem.
Why It Matters
Intel v. PACT XPP Schweiz reinforces that the PTAB cannot find “no motivation to combine” when its own factual findings satisfy the known-technique standard established by KSR. The decision is a reminder that the Federal Circuit will scrutinize whether PTAB obviousness conclusions are internally consistent — not merely whether the PTAB went through the right analytical steps. For technology companies defending IPR proceedings involving processor, memory, or system-architecture patents, the case highlights the risk that findings on record can be used against the patent owner even when the PTAB reached a favorable conclusion on the ultimate question.
More broadly, the ruling illustrates the ongoing tension between PTAB deference and the Federal Circuit’s obligation to ensure that obviousness determinations are legally coherent. Where a tribunal’s factual findings and its legal conclusion point in opposite directions, the Federal Circuit will not simply defer to the result.