Intel CEO Lip-Bu Tan may have just agreed with Nvidia CEO Jensen Huang and Meta's former chief AI scienti - The Times of India

Lip-Bu Tan's alignment with Jensen Huang is not a partnership announcement—it is a strategic retreat from an AI accelerator war Intel cannot win. Three years of Gaudi underperformance against Hopper and Blackwell, compounded by a foundry business hemorrhaging cash, made continued head-on competition a fiscal non-starter. The presence of Meta's former chief AI scientist on the same table shifts the negotiation from "who builds the chip" to "who defines the AI compute paradigm"—a far more consequential contest over heterogeneous interconnect standards (UCIe, CXL) and software-stack ownership. Technically, Intel will almost certainly trim its in-house accelerator team and redirect capital toward EMIB/Foveros advanced packaging and the 18A node, repositioning as the infrastructure layer of the AI era. AMD captures a near-term window as xDNA and MI300 absorb displaced demand, but if Intel's packaging capacity becomes a supplementary node for Nvidia, AMD's foundry leverage erodes structurally. On compliance, any architecture-level collaboration touching Meta's AI stack will intersect export-control regimes, triggering fresh scrutiny on the packaging ecosystem in Taiwan, China. Twelve to twenty-four months out: the AI silicon market consolidates into a two-horse race. Intel demotes itself to the substrate layer. The real battleground migrates to inference-side ASICs and the software moat.