Login to Continue Learning
Kress noted that NVIDIA has received licenses for several key customers in China but faces a small geopolitical situation that needs to be resolved between both governments. Chinese customers want assurance that the government will accept the H20 architecture. While uncertain, Kress believes this could add $2 billion to $5 billion in revenue.
NVIDIA’s shares have faced pressure due to Broadcom’s new contract with NVIDIA for custom AI chips worth up to $10 billion. However, Kress emphasized the importance of power efficiency in AI computing. She explained that large-scale systems require high performance per watt and dollar, which is crucial when considering long-term power usage.
Kress mentioned that NVIDIA’s next-generation Vera Rubin AI GPUs are on track for a one-year cadence, with all six chips taped out. As part of preparing data centers for Rubin, Kress highlighted the need for “gigawatts” of power, indicating significant demand for these new GPUs.
Overall, Kress emphasized strong demand for NVIDIA’s products and the importance of power efficiency in AI computing systems.
