Analyst warns Meta cloud ambitions could squeeze GPU-cloud rivals CoreWeave and Nebius

Meta’s growing interest in AI infrastructure is raising new competitive questions for companies that sell compute capacity to the market. In a Yahoo Finance segment published July 1, D.A. Davidson head of technology research Gil Luria argued that a potential Meta cloud business would be “very bad” for two GPU-focused providers, CoreWeave and Nebius.

CoreWeave (CRWV) and Nebius (NBIS) are positioned as providers of cloud-based access to high-performance computing, particularly GPUs used for training and running AI models. Their business models rely on attracting customers that want computing capacity without having to procure and operate their own data center hardware. The segment’s premise is that Meta, given its scale and existing infrastructure, could become a direct source of that capacity.

Luria’s central point in the discussion was not that AI demand would slow, but that channel dynamics could shift. If Meta offered cloud services for workloads that today might be bought from specialized compute providers, it could change pricing power and customer choices for firms like CoreWeave and Nebius. In other words, the competitive threat is framed as coming from Meta as a potential alternative path to capacity.

The segment also indicates how investors are thinking about the “who supplies the pipes” question in AI. GPU-cloud providers have benefited as companies tried to scale quickly, but the market has also been watching whether the largest AI buyers and infrastructure builders will decide to monetize capacity more directly through cloud services.

Meta, for its part, has publicly emphasized building AI capabilities and related infrastructure, which has included large investments in data centers and computing resources over time. Meta has also continued to publish frequent updates about AI and infrastructure on its newsroom. While those updates do not, by themselves, confirm the existence or timing of a broad external cloud offering aimed at third-party customers, they underline why industry observers connect Meta’s internal compute buildout with the broader cloud ecosystem.

Still, the Yahoo Finance post does not appear to provide detailed disclosures in the excerpted context about what “Meta’s cloud business” would specifically include, which customer segments it would target first, or how pricing and service levels might be structured. It also does not spell out the regulatory or contracting hurdles that could affect how quickly Meta could expand access for outside customers.

For CoreWeave and Nebius, the practical takeaway is that their competitive landscape could widen if Meta’s compute scale translates into a more general cloud offering. The risk, as Luria framed it, would be a reduction in the relative advantage enjoyed by specialized providers that have built their reputations on supplying capacity to AI developers and enterprises.