Broadcom, Apollo and Blackstone are turning AI infrastructure into a private-credit financing story at global scale.
The artificial-intelligence buildout is no longer just a contest between model developers and chipmakers. It is becoming a financing market large enough to pull in some of Wall Street's most powerful private-credit firms.
Broadcom, Apollo and Blackstone announced a new AI XPV Platform designed to support more than 20 gigawatts of compute capacity through 2028, using Broadcom's custom XPUs and networking systems for frontier AI labs including Anthropic and OpenAI. The platform begins with a $35 billion transaction led by Apollo, in partnership with Blackstone, to help fund more than 1 gigawatt of Anthropic compute capacity expected to start deployment at Fluidstack-based sites in mid-2026.
That structure matters because it shows how the economics of AI infrastructure are moving beyond ordinary cloud spending. Frontier models require huge blocks of chips, power, networking gear and data-center capacity. The bill is too large to be absorbed only through traditional corporate capital budgets or venture funding.
Private credit is stepping into that gap. For investors, the appeal is exposure to contracted infrastructure tied to the AI boom. For AI labs, the model can convert immense upfront hardware needs into a more manageable financing arrangement. For Broadcom, it deepens the company's role as a custom silicon and networking supplier at the center of the non-Nvidia AI stack.
The scale also raises the market's biggest question: whether AI demand can grow fast enough to justify the capital being committed. A 20-gigawatt target implies not just more chips, but more power generation, grid access, cooling, land, permits and long-duration customers willing to pay for inference and training capacity.
The deal is a signal that AI infrastructure is becoming a new asset class. The winners may not be only the companies with the best models, but also the firms that can secure energy, finance capacity and lower the cost of delivering each token.
If the boom continues, this platform could become a template. If demand disappoints, it may become a stress test for one of the largest private-credit trades of the AI era.
Image source: i.ibb.co