Intel CEO Discusses Potential AMD Collaboration as 14A PDK 0.9 Nears Release
Intel CEO Lip Bu Tan has revealed that he recently discussed potential collaboration opportunities with AMD CEO Lisa Su, suggesting that the 2 longtime semiconductor rivals could explore new areas of cooperation. Speaking during an October 24, 2026, fireside chat hosted by the Chinese American Semiconductor Professional Association (CASPA), Tan also confirmed significant progress on Intel's advanced 14A manufacturing process, with its important 0.9 Process Design Kit (PDK) approaching release.
During the discussion with CASPA President Jimmy Cheng, Tan emphasized his determination to restore Intel's CPU leadership while transforming the company into a more customer focused semiconductor manufacturer. The reported conversation with Lisa Su attracted particular attention because AMD remains Intel's strongest competitor in desktop and server processors. However, Tan did not disclose which technologies or business opportunities were discussed, and neither company has announced a new partnership.
The possibility of cooperation is not entirely unexpected. Intel and AMD already collaborate through the x86 Ecosystem Advisory Group, established in 2024 to improve architectural compatibility and simplify software development. More recently, both companies have worked toward integrating AMD's ZenDNN optimizations into Intel's oneDNN library for AI inference. Expanding cooperation into advanced packaging, manufacturing or other technologies remains possible, but no specific foundry agreement involving AMD processors has been confirmed.
Beyond the potential partnership, Tan provided an update on Intel Foundry's manufacturing roadmap. He confirmed that Intel 18A has entered volume production, while the more advanced 14A process has completed its initial PDK 0.5 milestone. Intel is preparing PDK 0.9, which will provide customers with more mature design rules and transistor models necessary for developing chips intended for the new manufacturing technology.
According to Intel Foundry's published specifications, 14A is designed to deliver 15% to 20% higher performance at equivalent power, or 25% to 35% lower power consumption at equivalent performance, compared with Intel 18A. The process also targets up to 30% greater chip density through second generation RibbonFET transistors, PowerDirect backside power delivery and advanced lithography.
These developments strengthen Intel's efforts to compete with TSMC through advanced manufacturing and High NA EUV technology. However, attracting external customers will require more than competitive process specifications. Intel must demonstrate consistent yields, predictable manufacturing schedules and sufficient production capacity.
Tan acknowledged that Intel's previous approach was often too rigid, explaining that the company is now placing greater emphasis on customized silicon, particularly Xeon processors, application specific chips, advanced packaging and emerging AI infrastructure opportunities. He recalled a customer who identified 15 areas where Intel needed improvement, reinforcing the importance of rebuilding trust with potential foundry partners.
With Intel 14A already demonstrating improving defect density and growing customer interest, the upcoming PDK milestone could become an important step toward securing additional manufacturing contracts. Nevertheless, no new AMD manufacturing agreement or major external 14A contract was disclosed during the discussion.
The most interesting development is the possibility of Intel and AMD expanding cooperation beyond their existing work on x86 architecture and software compatibility. While both companies will remain fierce competitors in gaming CPUs and data center processors, collaboration in selected technologies could benefit the broader semiconductor industry.
However, the more immediate opportunity lies in Intel Foundry's 14A progress. Delivering PDK 0.9 will give potential customers greater confidence in the process and allow them to advance their chip designs toward production readiness.
If Intel can combine competitive manufacturing technology with a more flexible customer strategy, it could become a stronger alternative to TSMC. A future manufacturing agreement with AMD would be particularly significant, but for now, the priority remains proving that Intel can execute its roadmap and deliver reliable capacity at scale.
Would you support AMD manufacturing future Ryzen or EPYC processors through Intel Foundry if it improved pricing and availability, or should AMD continue relying primarily on TSMC?
