Nvidia CEO Jensen Huang announced a projected $200 billion [1] market for CPUs designed to power agentic AI, including demand from China [2].
This projection signals Nvidia's intent to expand its dominance beyond graphics processing units (GPUs) into the central processing unit market. By targeting the infrastructure required for AI agents, the company is positioning itself to capture a broader share of the artificial intelligence hardware ecosystem, even as geopolitical tensions complicate trade with East Asia.
Huang made the announcement May 23, 2024 [3], during an event in Taipei, Taiwan [4]. He said, "We've found a brand new $200 billion market for Nvidia" [1].
The company is introducing the Vera CPU architecture to meet the needs of agentic AI, systems capable of acting autonomously to complete complex tasks. Huang said the company anticipates approximately $20 billion [5] in standalone CPU revenue from the Vera processor [6].
Despite ongoing U.S. export restrictions on high-end technology, Huang said that the $200 billion [2] opportunity specifically includes the Chinese market [2]. This suggests Nvidia believes long-term demand for AI-capable CPUs will persist in China regardless of current regulatory hurdles.
However, the impact of these trade restrictions remains a point of contention. Some reports indicate that export controls cloud near-term revenue prospects, while other forecasts focus on the long-term revenue potential of the Vera architecture [7].
Huang said, "We see a $200 billion opportunity in the CPU market, and that includes China" [2].
“"We've found a brand new $200 billion market for Nvidia."”
Nvidia's move into the CPU market via the Vera architecture represents a strategic shift toward providing a full-stack hardware solution for 'agentic AI.' By explicitly including China in its $200 billion valuation, Nvidia is betting that the global transition to autonomous AI agents will create a market demand so significant that it will outweigh the risks posed by U.S. trade sanctions and export controls.
