Intel CEO Lip-Bu Tan in Talks With AMD's Lisa Su Over Cooperation

Intel CEO Lip-Bu Tan has discussed potential areas of cooperation with AMD chief Lisa Su, according to a recording of an October 24 fireside chat — but he did not publicly detail what those areas are. The reported conversation adds a possible partnership between two longtime CPU rivals as Tan pushes a turnaround built on custom silicon and foundry growth.

The recording covers Tan's conversation with Jimmy Cheng, president of the Chinese American Semiconductor Professional Association (CASPA). In it, Tan laid out plans to rebuild Intel's CPU business, develop workload-specific chips, and attract foundry customers. The AMD cooperation itself remained unspecified. Neither company has announced a new agreement.

What did Tan promise on CPUs and customization?

Intel has historically designed processors around its own roadmaps. Tan said that model no longer fits customers who want silicon tuned to particular workloads. He cited Intel's Xeon server processors as one line where the approach is shifting, alongside continued GPU and accelerator development.

"We have to reclaim the leadership in CPU. That is something that I am very determined to do," Tan said.

He was blunt about the cultural change required. "Intel is so proud they never want to modify for customer needs," he said. "I'm modifying it for customer needs, and so that, I think, to me, is a fast-growing area, and I'm double, triple down on it."

Tan named four growth candidates beyond conventional CPUs:

• Application-specific integrated circuits (ASICs, chips designed for a single task)
• Co-packaged optics (CPO, networking components integrated with the silicon)
• Physical AI
• Advanced packaging

He wants two or three of these businesses to eventually contribute roughly 10% of Intel's revenue each. No timeline was given, and the targets are aspirational rather than guided.

Where does the foundry business stand?

Intel's 18A manufacturing process has reached volume production, Tan said. Development of the next-generation 14A node is progressing. Intel has released version 0.5 of the 14A process design kit (PDK), the bundle of technical data and tools designers need to build chips for the node.

Version 0.9 comes next, followed by version 1.0 for volume production in early 2027. "We have multiple customers [who] really want us to really scale their business," Tan said, referring to foundry interest.

He did not name those customers or disclose order sizes. In January, Intel had two prospective customers evaluating 14A test chips, with firm supplier decisions expected in the second half of 2026 and the first half of 2027. That decision window still lies ahead.

Tan framed foundry work as a trust business as much as a technology business. Customers hand over wafers that become products their own revenue depends on, so missed deadlines carry real cost. "The moment we take on the foundry, take on the wafer is a lot of responsibility to make them successful," he said.

He recounted one customer who waited six to eight months, hoping Tan would become CEO, before presenting 15 areas where Intel had fallen short. Tan said he addressed each problem individually, after which the customer agreed to deepen the relationship.

Is Intel-AMD cooperation actually new?

No. The two companies already cooperate through the x86 Ecosystem Advisory Group, formed in October 2024 with other technology companies to coordinate improvements to the x86 instruction set architecture (the foundation of both firms' processors) and ease developer support across both.

That cooperation reached AI software last week, when it emerged that AMD is integrating its ZenDNN optimizations into Intel's open-source oneDNN library, unifying the companies' CPU optimization work for PyTorch AI inference (a machine-learning framework) under a common software path. The companies have also collaborated on Advanced Matrix Extensions and other x86 capabilities.

The pattern is established: cooperate on shared technologies, compete for customers. Any new agreement from the Tan-Su discussions remains unannounced, and until Intel names foundry customers or ships 14A in volume, the turnaround plan rests on targets without disclosed figures.