NVIDIA's Vera CPU in full production — OpenAI, Anthropic, and SpaceX named as early adopters as the GPU+CPU integration thesis crystallizes
NVIDIA CEO Jensen Huang confirmed at COMPUTEX that the Vera CPU is in full production. Early-adopter names: OpenAI, Anthropic, and SpaceX. The triumvirate represents the three highest-revenue AI workloads in the world — frontier-model training, frontier-model inference, and the SPCX/xAI distribution stack — and they're all moving onto Vera-class CPU+Rubin GPU pods.
The substantive piece is the validation of the integrated-rack thesis. NVIDIA's GB200 NVL72 and the upcoming Vera Rubin Ultra positioning argued that AI workloads benefit from a tightly-integrated CPU+GPU rack rather than commodity x86 CPU + GPU. Having OpenAI, Anthropic, and SpaceX as early adopters confirms the procurement case: the highest-margin AI workloads in the world are now Vera-Rubin-bound by capex commitment.
The competitive frame for AMD is the structural response. AMD's Helios rack volume with Supermicro targets system-level parity with NVL72, but the Vera CPU integration is the dimension AMD doesn't have an equivalent for — AMD's CPU strength is in EPYC for general compute, not in AI-rack-integrated CPU. The integration moat may be the durable NVIDIA advantage even as AMD closes the GPU gap.
NVIDIA Blog — NVIDIA GTC Taipei at COMPUTEX: Live Updates on What's Next in AI → · CNBC — Nvidia's new PC chips represent CEO Huang's bid to win at every layer of AI stack → · Bloomberg — Nvidia Is Taking On Intel and AMD With AI Chip for Computers →