Intel’s effort to rebuild its chip business is widening beyond its own processor roadmaps. CEO Lip-Bu Tan says the company needs to tailor silicon to customer workloads and expand its foundry services; he has also reportedly discussed potential cooperation with AMD CEO Lisa Su. The talks could add to existing work between the rivals, though neither company has announced a new agreement.
Customer needs shape Intel’s product strategy
At an October 24 fireside chat with Chinese American Semiconductor Professional Association (CASPA) President Jimmy Cheng, Tan described plans to restore Intel’s CPU position, develop chips for specific workloads and attract more customers to its manufacturing business. He didn’t publicly describe what he and Su discussed.
Tan said Intel must adapt its products to customer requirements, including in its Xeon server CPU line, while continuing to develop GPUs and accelerators. “We have to reclaim the leadership in CPU. That is something that I am very determined to do,” he said.
He framed customization as a change in Intel’s approach. “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.”
The scope goes beyond CPUs. Tan named application-specific integrated circuits (ASICs), co-packaged optics (CPO), physical AI and advanced packaging as potential growth areas. His target is for two or three of these businesses to each reach about 10% of Intel’s revenue.
Foundry progress depends on customer confidence
Intel’s manufacturing roadmap is central to that plan. Tan said the 18A process has reached volume production and development of 14A is progressing well. Intel has published version 0.5 of the 14A process design kit (PDK), giving chip designers the technical information and tools to develop products for the node.
Version 0.9 is next, Tan said, with version 1.0 planned for volume production in early 2027. He said multiple customers want Intel to help scale their businesses, but didn’t name them or disclose potential order sizes.
The interest remains subject to evaluation and customer commitments. In January, Intel said two prospective customers were evaluating 14A test chips, with firm supplier decisions expected in the second half of 2026 and first half of 2027.
Tan cast trust as a core part of foundry work: customers entrust Intel with wafers that will become products tied to their own revenue. “The moment we take on the foundry, take on the wafer is a lot of responsibility to make them successful,” he said.
He also recalled a customer that waited six to eight months for him to become CEO before laying out 15 areas where Intel had fallen short. Tan said he addressed the concerns one by one, and the customer then agreed to deepen its relationship with Intel.
Rivals already share some x86 and AI work
Potential cooperation with AMD would build on existing ties. In October 2024, Intel and AMD joined other technology companies in forming the x86 Ecosystem Advisory Group, which coordinates improvements to the instruction set architecture and aims to make it easier for hardware and software developers to support both companies’ processors.
Their collaboration has also reached AI software. Last week, AMD’s ZenDNN optimizations began moving into Intel’s open-source oneDNN library, bringing CPU inference work for PyTorch under a shared software path. The companies have also worked on Advanced Matrix Extensions and other x86 capabilities through the advisory group.
That history shows where collaboration can serve both companies while they continue competing for customers. Tan’s reported discussions with Su may point to further opportunities, but for now the companies have announced no new agreement.