'NVLink is the key': Analysts ponder on probably the biggest tech deal of the decade - Intel + Nvidia and what it means for TSMC, AMD and others (www.techradar.com)

Nvidia’s surprise $5 billion equity investment in Intel — paired with plans for Intel to build Nvidia-custom x86 CPUs and x86 system-on-chips that integrate Nvidia RTX GPU chiplets — shocked markets and analysts. Nvidia’s market value jumped roughly $150 billion on the news and Intel shares surged to their best day since 1987. The technical roadmap calls out NVLink as a linchpin: integrating Nvidia’s high-speed interconnect into x86 designs could enable NVL72-style systems, data‑center x86 APUs combining Intel CPUs (including Xeons) with Nvidia GPUs, and tightly coupled AI platforms. For the AI/ML community this could redraw competitive lines. NVLink-enabled integration would let Nvidia extend beyond discrete GPUs into integrated graphics and system-level AI infrastructure, potentially squeezing AMD’s CPU/GPU position. It also raises strategic questions about foundry dynamics: some analysts see a path to bolster Intel’s manufacturing (e.g., 14A) if joint products drive volume, while others note the deal doesn’t solve Intel’s broader foundry troubles and that TSMC remains central — a point Jensen Huang emphasized. Overall, the partnership signals a move from discrete accelerators toward vendor-controlled, NVLink-centric systems that could accelerate AI deployment but leave major uncertainties around fab choices, performance tradeoffs, and competitive fallout for AMD and other foundries.