Login to Continue Learning
[Ad_1]
NVIDIA has just debunked reports claiming that its H20 sales have impacted the supply of H200 and H100 AI GPUs.
NVIDIA Has More Than Enough H100/H200 AI GPUs To Satisfy Every Order Without Delay, Company Debunks Reports Of H20 Constraining Supply
Recently, there were various reports suggesting that NVIDIA might prioritize the sale of H20 GPUs, leading to supply constraints for the H100 and H200 AI lineup. However, NVIDIA quickly addressed these claims in a statement.
According to NVIDIA, the reports are based on “erroneous chatter.” Their cloud partners can rent out every available H100/H200 GPU online, and the company can also fulfill new orders without any delays.
The NVIDIA H20 chip was designed as a US regulations-compliant variant for export to China after the US imposed trade restrictions on the country. Despite being a cut-down version, it still offers more AI horsepower and capabilities than Chinese-made Huawei chips such as the Ascend 920 series.
Reports from last month suggested that there might be backdoors or spyware in NVIDIA’s chips for China, but NVIDIA rejected these claims in its blog post.
Given NVIDIA’s popularity in the data center and AI space with its recent Hopper and Blackwell architectures, rumors are likely being spread to counter their growing influence. However, this hasn’t worked before and won’t work again, as demonstrated by NVIDIA’s record $41.096 billion in revenue from its AI segment during its recent financial earnings.
NVIDIA is confident in its supply and production capabilities. They want buyers to know that they can not only provide these AI chips but also do so without any delays. NVIDIA is transitioning to a yearly cadence where it will produce high-end AI GPUs each year, with Rubin launching in 2026, Rubin Ultra in 2027, and Feynman in 2028.
[Ad_2]
