As artificial intelligence infrastructure expands at unprecedented scale, access to electricity is emerging as one of the industry’s most critical strategic challenges. What was once primarily a question of securing land, servers, and network capacity is increasingly becoming a race to secure sufficient power to support the next generation of AI factories.

That reality was reinforced this week after BloombergNEF projected that data centers could account for roughly 20% of total U.S. electricity consumption by 2035, up from 5.9% today. Driven largely by AI workloads, the forecast estimates data center power demand could reach 194 GW by 2035—an 83% increase from the firm’s previous outlook published just seven months ago. The revision reflects a rapidly expanding pipeline of AI facilities planned across the country, particularly in key markets such as Virginia and Texas.

The forecast highlights a broader shift underway across the digital infrastructure ecosystem. As AI clusters grow larger and model training requirements continue to escalate, power availability is increasingly shaping where infrastructure can be built, how quickly projects can move forward, and which markets are best positioned to capture future investment. In many cases, electricity access is becoming as important as compute itself.

The implications extend well beyond the data center sector. Utilities, grid operators, policymakers, and infrastructure developers are all being forced to adapt to a new era of demand growth after decades of relatively stable electricity consumption. As AI deployment accelerates, competition for power is emerging as one of the defining challenges of the industry’s next phase—transforming energy infrastructure into a strategic pillar of the global AI race.