NVIDIA has rebutted the claims On X post that the sale of H20 AI GPUs is impacting the supply of H100 and H200 models. The company has issued this statement for clarification.
NVIDIA's Official Statement Says No Delay in Orders for H100 and H200
The company's version says the allegations are based on some "erroneous chatter." It states that its cloud partners currently have full access to their available H100/H200 GPUs, and that no new orders for these high-demand chips are being delayed.
Role of H20 AI GPU
The NVIDIA H20 chip was specifically developed as an alternative for export to China subject to US trade restrictions. While it was made to be in compliance, the H20 is reportedly a much better AI performer than Chinese chips being manufactured locally, such as Huawei's Ascend 920 series.
Market Speculation Addressed
This is not the first time NVIDIA has had to deny market speculation. A month ago, the company refuted claims that its chips for the China market had possible back-doors or spyware embedded. With its high stature and influence in the market, NVIDIA is also a target when rumors are spread to counter its growing influence.
Confidence Supported by Performance and a Clear Roadmap
Confidence for NVIDIA is based on its recent strong performance, with $41.096 billion revenue posted by its AI segment. It is this strong confidence that led the company to inform clients to place orders now and assure them that their orders will be fulfilled as promised.
For the high-end AI GPUs though, NVIDIA will be moving to an annual launch schedule. The roadmap yet to be announced specifies the following:
- Rubin Launching in 2026
- Rubin Ultra Launching in 2027
- Feynman Launching in 2028
