Helix brings $10B to AI infrastructure buildouts

Helix is a $10 billion AI infrastructure company built to link data centers, power, and connectivity.

KKR’s new Helix venture is a signal that AI infrastructure is moving from piecemeal projects to integrated platforms. Backed by $10 billion in anchored commitments, the company is designed to help hyperscale operators get the capital, power, and site capacity they need faster.

1. A single platform for data center buildouts

Helix is not just another capital vehicle. It is meant to package data centers, power generation, transmission, distribution, fiber, and related assets under one roof so large AI customers do not have to stitch the pieces together themselves.

That matters because the bottleneck is no longer only chips or software. The hard part is getting enough land, grid access, interconnects, and construction coordination at the same time.

• Hyperscale data centers
• Power generation facilities
• Transmission and distribution infrastructure
• Fiber networks and connectivity assets

2. $10 billion of anchored capital

The company launches with US$10 billion in anchored investment and long-term capital commitments. For AI infrastructure, that kind of balance-sheet support can shorten the gap between a plan and a financed project.

It also gives Helix room to pursue assets that need patient capital rather than quick-turn venture funding. That is especially relevant for grid-heavy projects, where permitting, utility coordination, and construction timelines can stretch for years.

• Anchored investment: US$10 billion
• Backed by long-term capital commitments
• Built for infrastructure with multi-year payback horizons

3. NVIDIA adds an AI factory blueprint

NVIDIA is not just a name on the cap table. As a strategic partner, it will help align Helix deployments with its DSX AI factory platform, which gives the venture a technical framework for AI-ready infrastructure.

That should matter to cloud providers and enterprise buyers that want capacity designed around AI workloads rather than retrofitted after the fact. In practice, it means the infrastructure plan is tied to compute architecture, not only real estate and power.

• Strategic partner role
• DSX AI factory platform alignment
• Focus on AI cloud providers and hyperscalers

4. Vistra ties power directly to deployment

Vistra will act as Helix’s preferred power provider, which is a practical answer to one of AI’s biggest constraints: electricity. Power generation and grid interconnections are often the gating items that slow new data center projects.

By pairing infrastructure development with a power partner from the start, Helix tries to reduce the stop-start cycle that happens when data center plans outpace available utility capacity.

• Preferred power provider
• Supports generation and grid interconnections
• Targets large-load customers

5. A management team built for scale

The company will be led by Adam Selipsky, the former CEO of Amazon Web Services, which gives Helix a leader with direct experience running cloud infrastructure at global scale. That background fits a business aimed at hyperscalers, not small deployments.

KKR says its infrastructure platform manages more than US$100 billion in assets and has invested more than US$70 billion across digital and power infrastructure sectors. That track record suggests Helix is being built as a large institutional platform, not a one-off project company.

• CEO: Adam Selipsky
• KKR infrastructure platform: over US$100 billion managed
• KKR digital and power investing: over US$70 billion

How to decide

If you are a hyperscale cloud operator, Helix is the most relevant if you need integrated access to power, fiber, and data center capacity from a single partner. If you are an investor, the story is the scale of the capital stack and the long-duration demand behind AI infrastructure.

If you are tracking the sector more broadly, Helix is a useful marker for where the market is heading: fewer standalone projects, more bundled infrastructure platforms built around power availability and AI deployment speed.