Nvidia's financial report ignited its stock price on Wednesday, with revenue and profits exceeding expectations, driving further market expectations for the concept of artificial intelligence.
After the impressive data, Nvidia quietly revealed its "hardcore" capabilities for updates and upgrades. Nvidia CEO Huang Renxun pointed out during the earnings conference call that Nvidia is developing a new chip following Blackwell, with a release schedule of once a year.
Previously, Nvidia's chip design frequency remained stable at once every two years, from the release of the Ampere in 2020, to the industry darling H100 chip (Hopper series) in 2022, and then to the highly anticipated Blackwell in 2024. But obviously, two years seems too long for Nvidia.
Earlier this month, renowned analyst Guo Minggui revealed that Nvidia's next-generation AI chip architecture Rubin will be launched in 2025, and R100 AI GPUs can be obtained in the market as early as next year. And currently, the credibility of this news seems to have risen sharply.
Huang Renxun stated that Nvidia will move forward at a very fast speed, with new CPUs, GPUs, network cards, switches... a large number of chips coming soon.
The exchange is fast, so the sales are also fast
Earlier this week, Amazon was reported to have suspended orders from Nvidia due to the fast iteration speed of AI chips, in anticipation of the upcoming release of Blackwell chips. Later, Amazon denied the suspension of the contract, but admitted to shifting towards purchasing Blackwell chips in the supercomputer project.
This has also raised concerns in the market that Nvidia's chip update speed may lead to unsold contemporary products. Regarding this, Huang Renxun responded that NVIDIA's new generation GPU is backwards compatible in both electrical and mechanical aspects, and runs the same software. Customers can easily transition from H100 to H200 and then to B100 in their existing data centers.
He believes that as the market transitions towards H200 and Blackwell, we expect demand to exceed supply for a period of time. Everyone is eager to put their infrastructure online to make profits as soon as possible, so their orders for Nvidia GPUs will not stagnate.
Huang Renxun raised an interesting question: Do companies want to be the first company to release the next major milestone in the artificial intelligence competition, or the second company to announce a 0.3% increase in effectiveness just a few days later?
In addition, Colette Kress, the chief financial officer of Nvidia, revealed that cars will become the largest vertical industry in Nvidia's data center this year. For example, Tesla has purchased 35000 H100 chips to train its full auto drive system. Outside of cars, Consumer Internet companies such as Meta are also growing strongly.