At the Cisco AI Summit in San Francisco, AWS CEO Matt Garman called the notion of space-based data centers “pretty far” from practical deployment, highlighting that the current cost to launch just one thousand-pound server rack into orbit remains prohibitively high. With average rocket payload capacity constrained and per-kilogram launch fees measured in tens of thousands of dollars, Garman emphasized that transport economics—not compute technology—are the primary barrier. He noted that global launch capacity would need to multiply by orders of magnitude before any meaningful volume of hardware could be deployed, ruling out orbital facilities as a near-term solution to surging AI compute demand.

Rapid expansion of AI workloads has driven AWS to consume record levels of power and cooling in its terrestrial data centers, where electricity usage can exceed 200 megawatts per campus and chillers circulate millions of gallons of water daily. Over the past year, AWS has added eight new hyperscale regions and increased its global server count by more than 15%, yet Garman warned that even this growth trajectory will struggle to keep pace with emerging generative AI models requiring 2–3× more compute per training run. AWS is now investing heavily in custom silicon and liquid-cooling solutions to improve energy efficiency by up to 30%, but executives concede that such upgrades are still more cost-effective than pioneering off-planet alternatives.

Investors watching AWS’s capital expenditures note that the division’s budget grew by nearly 25% year-over-year in the last quarter, reflecting a commitment to expand cloud capacity on Earth rather than chase speculative space projects. Industry analysts have revised cloud spending forecasts upward by an average of 10 percentage points for the coming fiscal year, citing continued demand from enterprise AI deployments in sectors such as autonomous vehicles and pharmaceutical research. With AWS growth rates still tracking above 20% annually and margins remaining in the mid-teens after infrastructure reinvestment, long-term investors are focusing on terrestrial scale-outs, custom hardware rollouts and regional network expansions as the key drivers of Amazon’s cloud segment profitability.