IT House February 24th news, UBS analysts said greatly shortened the AI GPU delivery cycle, from 8-11 months to 3-4A month later, Yingweida officially stated: "We expect the next generation of products (B100) will be in short supply.
At the recent financial report, Nvidia Chief Financial Officer Colette Kress said: "Because the demand far exceeds our supply capacity, we estimate that the next generation of products will have tight supply."
IT House learned from the report that Nvidia used Multi-Chiplet design for the first time on Blackwell to simplify the production process basedAs a result, production capacity is improved.
The Blackwell architecture that people are eager to expect and the B100 GPU based on this architecture are expected to achieve a major leap in performance, which is likely to stimulate NivineExisting customers place orders in advance.
However, the scarcity of Nvidia products may bring great opportunities for important competitors such as AMD and .If these two companies can provide a product that can defeat Nvidias current H100 and provide suitable software stacks, then customers will be willing to buy their products without waiting for several months because of the expected high delivery cycle.
Intel is preparing the next generation of GAUDI 3 and is developing the Falcon Shores accelerator for artificial intelligence and high -performance computing.AMD is launching a competitive Instinct MI300 accelerator, and has already developed MI400 generation products.
