Nvidia Halts China-Bound H200 Chip Output, Caps AI Investments at $30B
Nvidia halted production of China-bound H200 AI chips and is redirecting TSMC capacity to its new Vera Rubin platform. CEO Jensen Huang confirmed a $30 billion OpenAI investment instead of the planned $100 billion as Anthropic resumes Pentagon contract negotiations.
1. H200 Chip Production Halt
Export restrictions in Washington and Beijing have forced Nvidia to stop manufacturing H200 AI accelerators destined for China, cutting anticipated revenue from that market segment and freeing up wafer capacity at its TSMC fabs.
2. Capacity Shift to Vera Rubin
Nvidia is reallocating its TSMC output toward the upcoming Vera Rubin AI platform, aiming to accelerate development timelines and meet surging demand for next-generation data-center accelerators.
3. Adjusted AI Investments
CEO Jensen Huang indicated that Nvidia will finalize a $30 billion investment in OpenAI but will not proceed with the previously discussed $100 billion commitment, signaling a more cautious capital deployment amid valuations.
4. Anthropic Pentagon Talks Resume
Anthropic has reopened negotiations with the Pentagon on a prospective military AI contract, a development that could influence Nvidia’s defense-related AI sales if approved.