Apollo announced a $35 billion capital solution with Blackstone and Broadcom tied to AI compute infrastructure. That is a very different kind of AI story than a model launch or a product demo. This is AI being financed like infrastructure.
I think that is the part worth paying attention to. Once compute starts getting funded through huge debt structures, special purpose vehicles, and long-term leasing arrangements, the market is telling you that AI is not being treated like a normal software cycle anymore. It is being treated like something closer to railroads, telecom, cloud, or energy.
That does not automatically mean the economics are good. It means the stakes are getting bigger. If these companies need this much capital just to stay competitive, then the winners are not only going to be the companies with the best models. The winners are going to be the ones that can secure compute, structure financing intelligently, and keep customers growing fast enough to justify the buildout.
Sam C BarthThe bill shows up in your stack eventuallyI help operators keep HubSpot and RevOps simple enough that bigger shifts do not break the basics.Visit samcbarth.comMy hesitation is that this is where hype can get dressed up as financial engineering. Off-balance-sheet structures can make the story look cleaner than it feels underneath. The demand might be real, but the risk is also real. If customers slow down, if margins compress, or if compute gets cheaper faster than expected, these deals start to look a lot less elegant.
For business operators, the lesson is not to copy the scale. It is to understand the direction. AI is moving from a tool budget into an infrastructure decision. That means the teams adopting it need to think harder about where the cost shows up, who owns the risk, and whether the workflow value is real enough to justify the expense.
