NVIDIA's Blackwell AI Servers: Navigating US Regulations and High Costs in Global Markets

NVIDIA Expected To Offer "US-Compliant" GB20 Blackwell AI Servers To China

NVIDIA now looks to offer "US-compliant" Blackwell GB20 AI servers to China in an attempt to capitalize on rising demands. Well, it looks like Team Green is determined to consolidate its position in the Chinese markets, and the firm intends to achieve this by offering its next generation of Blackwell products in the region. Recently, we reported on how NVIDIA is preparing for the "B20 AI GPU", which will be a successor to the firm's H20 AI GPU, which was originally designed in an attempt to comply with US export regulations. This shows that NVIDIA is seeing enormous interest in its Blackwell architecture from China, which is why they have now decided to design a China-specific AI server as well. The news comes from The Information, citing supply chain sources, that NVIDIA has decided to offer Blackwell AI servers to Chinese customers, which are compliant with the regulations imposed by the Biden administration. The AI server is likely going to be called the "GB20", which will be offered alongside the B20 AI GPUs to provide a complete package to customers. While figures related to performance haven't been disclosed yet, it is said that the Blackwell generation will provide decent capabilities to Chinese customers despite them getting "downgraded variants." It looks like while the Hopper generation was all about AI accelerators, the Blackwell architecture will see a much wider adoption for its server offerings, given that companies are now focusing on creating giant AI clusters, to achieve maximum compute power. Similar is the case with Chinese AI-focused firms as well, and a recent report disclosed that the region's server markets are expected to grow tremendously, reaching up to $16 billion in valuation by 2027, and by the looks of it, Blackwell's servers will play a vital role. The restrictions from the US haven't bothered NVIDIA & its Chinese customers at all, since Team Green has come up with alternatives, by providing regulation-compliant products to satisfy both parties. With the GB20 expected to debut in Chinese markets soon, we will see a new wave of AI advancement in the country, and for NVIDIA, well, a new and wide revenue stream.

Nvidia and partners could charge up to $3 million per Blackwell server cabinet -- analysts project over $200 billion in revenue for Nvidia

According to a report from Morgan Stanley cited by United Daily News, Nvidia and its partners will charge roughly $2 million to $3 million per AI server cabinet equipped with Nvidia's upcoming Blackwell GPUs. The industry will need tens of thousands of AI servers in 2025, and their aggregate cost will exceed $200 billion. So far, Nvidia has introduced two 'reference' AI server cabinets based on its Blackwell architecture: the NVL36, equipped with 36 B200 GPUs, which is expected to cost from $2 million ($1.8 million, according to previous reports), and the NVL72, with 72 B200 GPUs, which is projected to start at $3 million. NVL36 and NVL72 server cabinets (or PODs) will be available not only from Nvidia itself as well as its traditional partners, such as Foxconn, the world's largest supplier of AI servers, Quanta, and Wistron but also from newcomers, such as Asus. Assuming that Nvidia's partner TSMC can produce enough B100 and B200 GPUs using its 4nm-class lithography process technology and packaged using its chip-on-wafer-on-substrate (CoWoS) technology, availability from newcomers should ease the tight supply of actual Blackwell-based machines. Based on the UDN report citing Morgan Stanley, Nvidia anticipates shipping between 60,000 and 70,000 B200 server cabinets, each priced between $2 million and $3 million next year. It translates to an estimated annual revenue of at least $210 billion from these machines, which means that companies like AWS and Microsoft will spend even more. It brings us to math by Sequoia Capital partner David Cahn, who believes that the AI industry has to earn around $600 billion to pay off machines and data centers. Demand for AI servers is setting records and will not slow down any time soon, which will benefit both makers of AI servers and developers of AI GPUs. Despite the influx of competitors, Nvidia's GPUs are set to remain the de facto standard for training and many inference workloads, which benefits the company. Based on talks with industry sources, the same Morgan Stanley report revealed that international giants such as Amazon Web Services, Dell, Google, Meta, and Microsoft will all adopt Nvidia's Blackwell GPU for AI servers. As a result, demand is now projected to exceed expectations, so Nvidia is prompted to increase its orders with TSMC by approximately 25%, the report claims.