This week’s massive deal between Blackstone and Google is meant to bypass Nvidia’s hardware monopoly, with Blackstone committing an initial $5 billion in equity to launch a new, dedicated AI cloud company that could purportedly mushroom to as much as $25 billion. With the announcement, Blackstone moves up the stack to finance the actual compute capacity and silicon hardware itself.

That has the potential to disrupt the status quo for how AI hardware is sold and financed. In this case, Blackstone is leveraging its capital and energy resources to fund the Google TPU-based “compute-as-a-service” that could offer a lower-cost alternative to Nvidia GPUs. 

Hyperscalers typically finance their own data centers, but with this deal, some experts believe Blackstone can possibly convert highly volatile tech risks into a more asymmetric financial upside. It might also signal trouble for smaller AI cloud providers like Nebius, Crusoe, CoreWeave, and Lamda, which have flourished with fast, flexible access to hardware. In an X post, Rittenhouse Research noted these neoclouds have largely aligned themselves with Nvidia hardware, creating an opening for the Google-backed TPU alternatives.

But, others argue that the deal is not “immediately problematic,” since enterprises, startups, and developers will continue to favor independent neoclouds because of, not in spite of, compatibility with Nvidia’s CUDA software platform, which is central to most AI software, developer tools, and ML models. Also, companies not looking for massive, multi-MW deployments will prefer the agility and speed of “GPUs in minutes,” so to speak – the bread and butter of neoclouds.

AI tale of two industries: Two recent reports capture opposite ends of the same AI infrastructure value chain wherehyperscalers are fearless, agile, borderless, cash-rich, and scalable; and telcos are fearful, entrenched, regulated, and yield-oriented.

Optical transport market forecast: After “standout” Q1 results, Dell’Oro Group’s Jimmy Yu says the market will surpass $18 billion in manufacturer revenue for the first time since 2000, driven AI data centers’ network infrastructure investments.

Dell Technologies World: CEO Michael Dell said agentic AI will shift AI Factory from a hardware supply story to an enterprise infrastructure stack for deskside systems, data orchestration, rack-scale AI infra, networking, cooling, and more.