NVIDIA has partially suspended the production of the H200 chip, redirecting related capacity at Taiwan Semiconductor to its next-generation architecture, Vera Rubin. This decision reflects a pragmatic response to the prolonged delay in U.S. export approval. Approximately 250,000 units of H200 inventory remain on hold, awaiting policy clearance, while new-generation computing power products are being fiercely contested by OpenAI and Google.
$NVIDIA (NVDA.US)$ is shifting its manufacturing capacity at Taiwan Semiconductor from the H200 chip to the next-generation architecture Vera Rubin, a strategic move that reflects the company's latest assessment of the current regulatory environment.
On Thursday, the Financial Times reported, citing two informed sources, that NVIDIA had suspended production of the H200 chip for specific markets and reallocated Taiwan Semiconductor’s related capacity originally designated for the H200 to Vera Rubin.
NVIDIA Chief Financial Officer Colette Kress stated last week during an earnings call that although the U.S. government has approved limited exports of the H200 product, "we have not generated any revenue to date and do not know whether any export approvals will be granted."
This adjustment indicates that NVIDIA no longer anticipates large-scale sales of the H200 in the near term. Informed sources suggest that this move may, to some extent, accelerate the delivery and launch process of Vera Rubin, which is currently experiencing strong demand from tech giants such as OpenAI and Google.
H200 Production Suspended, Vera Rubin Takes Over
The H200 is an earlier generation of NVIDIA’s AI processor, previously positioned as a compliant product meeting U.S. export control requirements. Vera Rubin, on the other hand, is NVIDIA’s latest chip architecture unveiled earlier this year, designed for more complex AI systems and seeing robust demand in the high-end computing market.
According to internal NVIDIA sources cited by the Financial Times, NVIDIA has already produced approximately 250,000 H200 chips. If only limited orders are ultimately approved, existing inventory would be sufficient to meet the corresponding demand.
"Rather than waiting amid uncertainty, NVIDIA must pivot toward achievable goals, especially when advanced products are in short supply," one informed source noted.
Approval Process Stalled, Supply Chain Preparations Fall Through
NVIDIA had previously made significant efforts to secure sales approval for the H200. Following signals last December from President Trump indicating possible permission for sales, NVIDIA promptly ramped up production. According to a previous report by the Financial Times, the company anticipated customer orders exceeding one million units, with suppliers working around the clock to stockpile inventory, originally planning deliveries to begin as early as March this year.
NVIDIA CEO Jensen Huang stated in early January this year that demand was 'very strong, and we have activated the supply chain with the H200 currently in production.' However, the approval process subsequently came to a standstill.
Despite the shift in production capacity, NVIDIA has not completely ruled out the possibility of restarting the H200. According to sources familiar with the matter, if there is a policy reversal, NVIDIA could reallocate or increase the H200 supply chain capacity within a maximum of three months, while existing inventory would also be sufficient to meet demand and fulfill deliveries during this period.
Editor/Lambor