SpaceX's planned constellation of AI satellites can gain unassailable economies of scale by 2031, becoming cheaper than some current terrestrial data centers, as per fresh numbers crunched by Morgan Stanley. According to Morgan Stanley, SpaceX's compute cost in orbit can fall down to $6.5 per watt per year by 2031 vs. $6.8 per watt per year all-in costs for NVIDIA Blackwell GPU-based terrestrial data centers As we detailed recently, SpaceX has now unveiled its first dedicated satellite design for AI compute. Dubbed the AI1 satellite, it can support up to 150kW of peak compute payload, replete with liquid radiators, meteoride shielding, […]

Read full article at https://wccftech.com/spacexs-orbital-ai-compute-can-reach-6-5-per-watt-per-year-by-2031-which-is-cheaper-than-nvidia-blackwell-based-terrestrial-data-centers/

As compute requirements grow in AI datacenters, so do the power requirements, which are estimated to reach 17x higher with NVIDIA's Feynman. NVIDIA Feynman Racks Estimated To Feature 17x Higher Power Semi Costs Per Rack Versus Blackwell NVIDIA Feynman GPUs feature several groundbreaking features and will launch in 2028, after Rubin. The company has been working hard to deliver more efficient AI solutions, but as requirements grow, power requirements have increased tremendously. Morgan Stanley Research has published a chart that visualizes the total power semi content of three AI rack solutions from NVIDIA. Starting with the baseline Blackwell or B200, […]

Read full article at https://wccftech.com/nvidia-feynman-gpus-push-power-semi-content-17-times-higher-vs-blackwell/