In the span of just a few quarters, data center demand has accelerated materially, as what was once a measured build-out of hyperscale capacity in support of cloud services has now become a full-throttle race, driven by the relentless demands of artificial intelligence (AI). From OpenAI’s near-doubling of its US data center requirements to Anthropic’s ambitious 9GW roadmap, the scale of this transformation is without precedent. The tectonic plates of cloud, compute, and connectivity are shifting in real time—and with them, the fortunes of the companies that power this ecosystem.
A DEMAND CURVE LIKE NO OTHER
At the center of this acceleration lies OpenAI. Only months ago, the company’s US AI training demand profile hovered around 5GW. Today, it is approaching an extraordinary 10GW, split across both training and inference workloads. Oracle remains its key partner, procuring vast blocks of capacity to support model training, with Vantage Data Centers securing a 1.4GW deal in Shackleford, Texas. Yet OpenAI is not placing all its bets on a single operator. Channel checks reveal additional multi-gigawatt negotiations directly with third-party providers, which underscores a strategic diversification of its footprint.
Inference workloads, too, are scaling aggressively. Oracle is scouting 200– 300MW per market across eight US metros, with similar expansions in 20 global locations. Taken together, these deployments translate into an incremental 1.6–2.4GW of capacity demand—on top of OpenAI’s US training requirements. Internationally, the picture is equally bold. Projects under the Stargate banner are already underway in the UAE and Norway, with reports of a 1GW site in India and joint OpenAI / NVIDIA ventures in the United Kingdom.
This story is part of a paid subscription. Please subscribe for immediate access.
ABOUT THE AUTHOR
Michael Elias is a Senior Equity Research Analyst at TD Cowen. He covers the communications infrastructure sector, including data centers and content delivery networks (CDNs) and has been a member of the communications infrastructure team since 2017. Prior to joining TD Cowen, Elias worked as an equity analyst at Xanthus Capital Management. He received his B.S. in Industrial Engineering and Operations Research: Engineering Management Systems at Columbia University’s School of Engineering and Applied Sciences.
