KKR and Nvidia Launch Helix Digital Infrastructure With $10 Billion to Solve the AI Data Centre Power Crisis

KKR, together with Nvidia, the Kuwait Investment Authority and power company Vistra, launched Helix Digital Infrastructure on June 11, a new company designed to serve as a single coordination point for hyperscalers' data centre, power, and connectivity needs. As Business Wire announced, the venture launched with more than $10 billion in long-duration committed capital and is led by Adam Selipsky, who served as CEO of Amazon Web Services until late 2024 and brings direct operational experience at the precise scale of customer Helix intends to serve. The company will invest across four verticals: hyperscale data centre development and operations, baseload and flexible power generation, transmission and distribution infrastructure, and fibre and connectivity networks. Nvidia will serve as a cornerstone strategic partner, aligning Helix's data centre projects with its DSX AI factory architecture — an approach explicitly designed to maximise tokens per watt and minimise total cost of ownership for AI workloads.

The founding thesis of Helix rests on a stark diagnosis of what is actually limiting AI infrastructure growth. As Selipsky himself stated at launch, more than 25 percent of announced data centre projects are currently failing to deliver on schedule. Research firm Sightline Climate has put the figure even higher, estimating that 30 to 50 percent of data centres scheduled to open in 2026 will be delayed or cancelled entirely. According to analysis published by MarketWise, transformer delivery times now stretch three to five years in the United States, switchgear is sold out through 2028, and grid connection queues — not capital or demand — represent the primary bottleneck constraining the AI buildout. More than 75 data centre projects worth $130 billion faced community opposition in the first quarter of 2026 alone. Helix is structured explicitly to confront these constraints by bundling capital, energy contracts, and construction capabilities under a single platform rather than requiring hyperscalers to assemble those elements themselves.

The investor lineup is itself a signal of how profoundly AI infrastructure has attracted sovereign and strategic capital. KKR already manages more than $100 billion in infrastructure assets globally. The Kuwait Investment Authority brings sovereign wealth firepower and a long investment horizon suited to decade-scale infrastructure cycles. Nvidia's involvement is particularly striking: its DSX AI factory architecture will effectively serve as the technical blueprint for every data centre Helix builds, tightening the link between chip design and physical facility design in ways that could create durable competitive advantages for hyperscalers who adopt the platform. Vistra, which operates generation capacity approaching 50 gigawatts across 18 U.S. states, becomes Helix's preferred power provider — a relationship that Data Center Frontier described as central to Helix's strategy of pre-securing power in a market where grid connections have become the scarcest resource.

Helix's launch is part of a broader pattern of private capital moving decisively into AI infrastructure at a scale that blurs the line between venture investment and sovereign wealth deployment. According to Crunchbase data, global venture capital hit $297 billion in the first quarter of 2026 alone, with Amazon, Nvidia, Microsoft and SoftBank collectively contributing more than $140 billion to the quarter's largest rounds. The Information reported that a separate AI infrastructure firm, also backed by KKR and Nvidia and led by Selipsky, was simultaneously exploring acquisition of an existing data centre company to accelerate its build-out of custom facilities for large cloud customers. Analysts have noted that when the same firms that manufacture GPUs, build cloud infrastructure, and hold equity stakes in the leading AI labs also co-own the physical facilities those labs run on, the potential for vertical integration raises complex questions that antitrust regulators in both the United States and the European Union have signalled interest in examining.

For enterprise technology buyers, the implications of Helix's model are significant beyond the headline capital figure. If AI infrastructure truly cannot be assembled piecemeal at the pace and scale the industry requires — as Selipsky has argued — then platforms that offer integrated data centre, power, and connectivity solutions will increasingly command premium pricing from hyperscalers desperate to accelerate deployment timelines. That dynamic in turn concentrates market power among a small number of heavily capitalised players, reducing the competitive diversity that has historically kept cloud infrastructure costs in check. Nvidia CEO Jensen Huang described AI as driving "the largest infrastructure buildout in modern history," and Helix is positioning itself as the instrument through which that buildout is actually delivered. Whether the bet pays off will depend not only on whether Helix can navigate the grid queues and regulatory hurdles its founders acknowledge, but on whether the extraordinary pace of AI demand growth continues to justify the extraordinary scale of capital being committed to serve it.