Nvidia CFO Claims $20 Billion CPU Revenue Year Will Make It the World's Top CPU Supplier

24 May 2026

Nvidia CFO Colette Kress announced the company has visibility to nearly $20 billion in CPU revenue this year, positioning it to become the world's leading CPU supplier, driven by its new Vera CPU chip based on custom Arm cores. Major hyperscalers and AI labs including Anthropic, OpenAI, and Oracle have begun taking delivery of Vera-based systems, with Nvidia claiming the chip offers significant performance and efficiency advantages over x86 alternatives. This CPU push comes alongside a strong fiscal Q1 2027, in which Nvidia reported $81.6 billion in revenue — up 85% year-on-year — and forecast Q2 revenues of $91 billion.

Dominating the GPU market apparently wasn't enough. Nvidia now has its sights set on becoming the world's largest CPU supplier too, and according to CFO Colette Kress, it's not a distant ambition.

"We have visibility to nearly $20 billion in total CPU revenue this year, setting us up to become the world's leading CPU supplier," Kress said during Nvidia's Q1 FY2027 earnings call on Wednesday.

Nvidia isn't new to CPUs. The company announced its first Arm-based datacenter processor, Grace, back in 2021, though for years these chips mostly lived inside GPU-heavy systems destined for AI datacenters and supercomputers. The idea of Nvidia competing seriously in the standalone CPU space felt distant.

That started shifting in February when Nvidia revealed Meta was deploying standalone Grace CPU Superchips in its datacenters for general workloads, including AI agents. Then in March, at GTC, Nvidia formally introduced the Vera CPU, a chip built around 88 custom Olympus Arm cores with simultaneous multi-threading support and confidential computing features baked in.

The specs are notable. Each Vera chip can be configured with up to 1.5 TB of LPDDR5x SOCAMM memory offering up to 1.2 TB/s of bandwidth, the low-power memory type more commonly associated with laptops. Kress claims Vera delivers 1.5x better performance per core, double the performance per watt, and four times the rack density compared to x86 alternatives.

Reference designs put two Vera CPUs on a single board connected via NVLink. In Nvidia's highest-end rack-scale AI systems, Vera pairs with Rubin GPUs in a 2:1 GPU-to-CPU ratio.

Since the chip's spring announcement, Kress says virtually every major hyperscaler and system builder has signalled plans to deploy it. This week, Anthropic, OpenAI, Oracle, and SpaceX all received their first Vera-based systems.

"Vera CPU opens a brand new $200 billion TAM for Nvidia, a market we have never addressed before," Kress said.

Worth keeping in mind: like Nvidia's push into Ethernet networking, these CPUs are squarely aimed at AI and HPC workloads. They aren't drop-in replacements for x86 chips across general enterprise computing. Not yet, anyway.

The CPU ambitions land at the tail end of a very strong quarter. Nvidia posted $58.3 billion in profit on $81.6 billion in revenue, with revenue up 85 percent year-on-year and 20 percent sequentially. Kress credited an "inflection in inference demand" for the quarter-on-quarter jump.

Nvidia also reorganised how it reports revenue. There are now two segments: a datacenter group covering cloud, hyperscale, neocloud, and enterprise, and an edge group for everything else including gaming, robotics, automotive, and vRAN. Datacenter pulled in $75.2 billion, split almost evenly between hyperscalers and public cloud providers ($38 billion) and neoclouds, industrial, and enterprise customers ($37 billion). Edge contributed $6.4 billion, with Blackwell-based workstation hardware cited as a key driver.

For Q2, Nvidia is guiding for $91 billion in revenue, give or take two percent, with that forecast explicitly excluding any datacenter sales into China.

Nvidia has been trying to revive its Chinese GPU business since the Trump administration cleared it to sell H200 processors to Chinese customers in December, a first. Despite securing approval and racking up billions in orders, shipments are reportedly stuck in Beijing's bureaucratic machinery.