Former AWS CEO Adam Selipsky will lead a new $10 billion AI data center venture, GeekWire reported. The move signals the escalating capital race for AI compute infrastructure.
Key facts
- $10 billion capital commitment for AI data center venture.
- Adam Selipsky, former AWS CEO, to lead the company.
- Selipsky left AWS in May 2024.
- Venture targets AI compute infrastructure gap.
- No name, locations, or investors disclosed yet.
Adam Selipsky, who stepped down as CEO of Amazon Web Services in May 2024 after a tenure marked by aggressive expansion, is returning to lead a new venture focused on building AI data centers. According to GeekWire The venture is capitalized at $10 billion, a figure that underscores the immense capital required to build out AI compute infrastructure.
The company has not yet disclosed its name, specific locations, or investor backing. The $10 billion figure, however, places it among the largest dedicated AI data center plays globally, rivaling projects from Crusoe (which claims a 5 GW pipeline) and Google's $11B/year commitment to SpaceX for compute. Selipsky's leadership lends immediate credibility and operational expertise, given his track record scaling AWS's infrastructure to serve millions of customers.
Why This Matters More Than the Press Release Suggests
Selipsky's move is less about one startup and more about the structural shift in AI infrastructure. The hyperscalers—Google Cloud, AWS, Microsoft Azure—are all building their own capacity, but the demand from AI startups and enterprises is outstripping supply. New ventures like this one, backed by veteran operators, aim to fill the gap. The $10 billion figure also signals that capital markets are willing to fund independent infrastructure plays, not just the Big Tech incumbents.
This is the second high-profile departure from AWS's executive ranks for an AI infrastructure bet. The trend suggests that the bottleneck for AI progress is no longer models or algorithms—it's the physical infrastructure to run them. Selipsky's venture will compete directly with offerings from his former employer, as well as Google Cloud and Microsoft, for the same pool of enterprise AI customers.
The venture's success will hinge on its ability to secure power, land, and chips—the three scarce resources for AI data centers. With Google booking Intel for 3M+ TPUs in 2028 and Nvidia's Blackwell GPUs in high demand, supply chain access will be critical.
What to watch
Watch for the venture's official name, first site location, and anchor investor reveal, likely within the next 90 days. Also watch whether Selipsky's venture secures a multi-year GPU reservation deal with Nvidia or another chipmaker—that would signal supply chain access.
Source: news.google.com
[Updated 12 Jun via gn_dc_power]
The Information reports that OpenAI is in talks to lease a 10-gigawatt data center campus in Ohio, with potential backing from Nvidia. The project could cost up to $500 billion and would be located on federal land. This development is separate from Selipsky's venture but underscores the same AI infrastructure race. Nvidia's involvement as a backer signals a deepening of its role beyond chip supplier to project financier [per The Information].
