NVIDIA's various Taiwan-based systems and AI server part suppliers (cabinets, assembly, liquid cooling components, and more) expects orders for GB200 NVL36 AI servers to start in mid-September, with the higher-end GB200 NVL72 AI server cabinets expected in the second half of October.
In a new report from Ctee, we're learning that the "market generally believes" the GB200 AI server orders will be "launched as expected" with NVL36 and NVL72 AI server cabinets in mid-September, and late-October, respectively. GB200 is crucial for NVIDIA, as it's expected that out of the 5 million Blackwell chips to be produced in 2025, around 80% of them will be used in GB200.
Each of the GB200 AI GPUs cost around $30,000 each, with the GB200 Superchip (CPU + GPU) costs upwards of $70,000. NVIDIA's new NVL72 AI server cabinet costs around $3 million per AI server, filled with 72 x B200 GPUs and 36 x Grace CPUs. The NVL36 AI server cabinet features 36 x B200 AI GPUs and 16 x Grace CPUs.
- Read more: NVIDIA hits major roadblocks with Blackwell AI GPU: revised B200A AI GPU coming
- Read more: NVIDIA's new Blackwell AI GPUs 'encountering major issues' which requires redesign
- Read more: NVIDIA's next-gen Blackwell AI GPUs delayed, 'design flaws' are to blame
- Read more: NVIDIA to make $210B revenue from Blackwell GB200 AI servers in 2025 alone
- Read more: NVIDIA places fresh new orders with TSMC for more Blackwell GB200, B100, B200 AI chips
- Read more: Foxconn is the sole supplier of NVLink switches for next-gen GB200 AI servers
- Read more: NVIDIA GB200 AI servers led by Foxconn with 40% and Quanta with 30%: ships in Q3 2024
- Read more: NVIDIA's next-gen GB200 AI server cabinets to ship in 'small quantities' in Q4 2024
- Read more: NVIDIA's new GB200 Superchip costs up to $70,000: full B200 NVL72 AI costs $3M
NVIDIA's new GB200 AI servers are moving into the world of liquid-cooling, which is increasing costs and the complexity of these already complex AI servers. Taiwan is home base for NVIDIA's production, with Taiwanese companies jumping to meet the AI demand for NVIDIA, with issues of liquid-cooling leakage in the last couple of months leading into GB200 shipping.
The demand for AI servers is stronger than ever, with volume shipments expected to ramp up in January 2025.