Nvidia has reportedly postponed the release of its next major AI chip, the “Blackwell” B200, due to an unexpected design flaw discovered late in the production process. According to sources, this delay will push the chip’s availability back by at least three months, affecting its planned large-scale distribution until the following year.
The delay was communicated to key clients such as Microsoft and another undisclosed cloud provider. The B200 chips are the anticipated successors to the immensely popular H100 chips, which have become a cornerstone in the AI cloud industry and have significantly boosted Nvidia’s market valuation.
Nvidia spokesperson John Rizzo mentioned that the company expects to ramp up production in the second half of the year but did not provide further comments on the rumors.
To address the design issues, Nvidia is conducting a new series of test runs with Taiwan Semiconductor Manufacturing Company (TSMC). As a result, substantial shipments of the Blackwell chips are now expected to commence only in the first quarter of the following year. Major tech companies like Microsoft, Google, and Meta have already placed orders worth tens of billions of dollars for these chips.
This development comes just months after Nvidia’s announcement that Blackwell-based products would be available from partners starting in 2024. The B200 chips are intended to initiate a new annual cycle of AI chip releases from Nvidia, as competitors like AMD are also striving to develop their own AI chip solutions.