Nvidia isn’t just a GPU company anymore. Jensen Huang wants you to know that, and he’s putting a $20 billion price tag on the message.
The Nvidia CEO has publicly sized the revenue opportunity for standalone Vera CPUs at roughly $20 billion, adding that the chips will be supply constrained for the long term.
What Vera actually is, and why it matters
Vera is a standalone processor designed for data centers that doesn’t need to be paired with an Nvidia GPU to justify its existence.
Introduced at GTC 2026, Vera promises 2x efficiency improvements and 50% faster performance compared to traditional x86 rack systems. Nvidia is claiming its ARM-based CPU can do what Intel’s Xeon and AMD’s EPYC do, but meaningfully better on both speed and power consumption.
The broader picture ties into Nvidia’s Vera/Rubin platform, which the company says has generated guidance suggesting $1 trillion in cumulative orders by the late 2020s. The platform reportedly offers a 35x improvement for AI inference workloads.
Meta is already on board
Meta has committed to deploying Vera CPU-only servers as part of a dual-vendor strategy alongside AMD, with deployment targeted for 2027. Meta isn’t just buying Vera as a companion to Nvidia GPUs. It’s deploying Vera in configurations where the CPU is the star of the show.
A dual-vendor strategy with AMD also reveals that Meta views Nvidia as a credible alternative to AMD’s EPYC rather than a GPU-only partner.
The competitive landscape just got more complicated
Nvidia already owns the GPU relationship with virtually every major cloud and enterprise customer. Vera is built on ARM, not x86. Amazon’s Graviton processors already proved that ARM-based data center CPUs can win on price-performance.
Investors watching this space should pay attention to Vera’s actual production timelines and volume shipments, and whether Meta’s 2027 deployment timeline holds and whether other hyperscalers follow with similar commitments.
The $20 billion opportunity Huang is describing would represent a massive new revenue stream for a company that already generates enormous returns from GPUs.