AMD EPYC Venice Could Ship More CPUs Than NVIDIA Vera in 2027

AMD’s next EPYC server processor platform, codenamed Venice, could reach higher shipment volumes than NVIDIA’s Vera CPUs in 2027, according to a new industry forecast. The estimate points to 6.75 million Venice units for AMD, compared with 5.75 million Vera CPUs for NVIDIA.

Both companies are preparing for a period where CPUs become more important in AI infrastructure. GPUs remain central to training and running large AI models, but data centers also need powerful processors to manage memory, networking, storage, orchestration, and other workloads around those accelerators.

AMD’s Venice platform is expected to use the upcoming Zen 6 architecture and TSMC’s 2nm manufacturing process. NVIDIA’s Vera CPUs are expected to use a 3nm process and are designed to support its wider AI data center strategy.

AMD Could Gain Ground With Zen 6 EPYC Venice

The projected 6.75 million unit figure would represent a major increase for AMD’s server CPU business. The forecast suggests Venice could ship more than five times the estimated volume of AMD’s current generation EPYC products in 2026.

That would put AMD in a strong position as cloud companies, AI firms, and high performance computing customers expand their infrastructure. EPYC has already become a serious alternative to Intel Xeon in servers, and Venice could help AMD strengthen its role in AI related deployments.

The expected gap is not enormous, but it would be notable because NVIDIA is entering the server CPU market with strong momentum from its AI hardware ecosystem.

NVIDIA Will Still Be a Major TSMC Customer

NVIDIA is still expected to remain TSMC’s largest advanced packaging customer in 2027. The company relies on TSMC’s CoWoS packaging technology for its AI GPUs, including Blackwell and Rubin products, as well as for Vera CPUs.

The forecast suggests NVIDIA’s use of CoWoS L packaging for AI GPUs could reach around 910,000 units in 2027, up roughly 40% from the prior year. Its Vera CPU shipments are also expected to grow quickly as NVIDIA expands systems built around CPUs, GPUs, networking, and software.

NVIDIA’s advantage is its ability to sell a complete AI platform. A customer can buy GPUs, CPUs, networking hardware, data processing units, cables, software tools, and support through one ecosystem. That makes it easier for large data center operators to deploy AI clusters at scale.

CPUs Are Becoming More Important in AI Infrastructure

The AI market is often discussed through GPUs, but CPUs remain essential. They manage operating systems, coordinate workloads, feed data to accelerators, handle networking tasks, and support general purpose computing across large server fleets.

AMD’s EPYC Venice processors may appeal to customers that want flexible CPU performance for AI, cloud services, and high performance computing. NVIDIA’s Vera CPUs are expected to be more closely tied to AI focused systems, especially deployments that use NVIDIA accelerators.

The bigger challenge for both companies may come from custom silicon. Large technology firms are increasingly designing their own processors and AI chips to reduce reliance on external suppliers. Companies such as Google, Amazon, and OpenAI are already investing in custom hardware strategies.

That trend could create more competition for AMD and NVIDIA, even as demand for computing power continues to grow. For now, the forecast suggests that 2027 could be a major year for server CPUs, with AMD and NVIDIA both expanding their presence in the AI infrastructure market.