AI infrastructure funding shifts to private equity and sovereign funds

In short: More of the money for AI data centers and large computer clusters is coming from private investors instead of public stock markets.

What's going on

Building AI systems now requires huge amounts of physical infrastructure, like data centers and large groups of GPUs (special chips that act like many fast calculators working at once). These projects also need steady electricity and networking gear to move data quickly. The price tag is getting so large that companies are increasingly looking beyond traditional public markets.

The Financial Times reports that private investors are stepping in, including private equity firms, sovereign wealth funds (government-owned investment funds), infrastructure specialists, and large asset managers. The idea is to finance AI build-outs through private deals, which can move quickly and handle very large checks.

Recent examples show the scale. Crusoe raised $1.375 billion in a private Series E round to build AI “factories,” meaning facilities designed to run AI workloads at high volume. Poolside AI is reportedly finalizing up to $2 billion in private funding, and NVIDIA is said to be committing $500 million to $1 billion, along with access to more than 40,000 GPUs through a CoreWeave data center project. Separately, KKR has committed $31.3 billion in equity to digital infrastructure over six years.

What to watch

Analysts project AI-related spending on infrastructure could approach $700 billion by the end of 2026. If that happens, expect more creative financing, and more competition for basics like power, land, and GPUs, since these are the “factories and fuel” behind modern AI.

Source: Financial Times