Nvidia opens CPU campaign at Computex while Intel defends turf

Nvidia CEO Jensen Huang told press at GTC during Computex week in Taipei that the company has manufacturing capacity to supply "robust CPU and GPU growth" simultaneously. The statement arrived one day after Nvidia unveiled a chip that processes AI workloads on-device, positioning the company to address the $200 billion CPU market with AI agent PCs shipping from Microsoft, Dell, and HP, according to TechCrunch.

Intel CEO Lip-Bu Tan presented at the same Computex forum on June 2. The timing is not coincidental. Nvidia is pricing into a market Intel has held for two decades, and the deployment vector is enterprise client refresh plus the agent-PC category that both companies now claim as addressable.

Huang's capacity statement matters because it signals no tradeoff between datacenter GPU shipments and the CPU ramp. If Nvidia can ship both without constrained substrate or CoWoS allocation, the margin structure of the x86 installed base comes under pressure in 2027 when enterprise budgets turn over. Intel's Computex remarks did not disclose shipment guidance, but the defensive posture was clear.

The agent-PC play depends on whether local inference at acceptable cost and latency moves from roadmap to shipping product. TechCrunch framed it as conditional: "If Nvidia has cracked a way to bring AI agents easily, safely and usefully to the masses, it could — and should — be big." The word *if* carries the entire business case. On-device inference has been two quarters away for three years. This time the SKU exists and the OEMs have committed, which changes the prior.

Nvidia is no longer a GPU-only provider. It is a compute provider with two product lines competing for the same enterprise dollar, one in the rack and one on the desk. Intel now defends both.