Sources: AI startup Poolside held talks with Google and others to revive a Texas data center project after a CoreWeave deal and a $2B Nvidia-led round collapsed
Context & Ripple Effects
Poolside had been pursuing an unusually infrastructure-heavy path for an AI coding company: it previously planned a 500-acre, natural-gas-powered Texas facility with CoreWeave, alongside a proposed Nvidia-backed financing. The reported collapse of both arrangements turns that plan from a paired compute-and-capital package into a search for a new anchor partner.
The outreach to Google follows Poolside's earlier effort to raise substantial outside capital for its developer-tools business, including a prior $400M-plus fundraising effort. It also tests whether a hyperscaler can replace the role CoreWeave was to play in the original Texas data-center plan.
First-order effects
- Poolside must secure a new counterparty and financing structure to revive the Texas project after the CoreWeave agreement and Nvidia-led round fell through.
- CoreWeave and Nvidia lose their proposed direct roles in Poolside's planned buildout, while Google becomes a potential alternative partner rather than a confirmed replacement.
Second-order effects
- A revived project would likely require Poolside to rework the linkage among site development, power, compute supply, and startup financing—rather than rely on the prior bundled arrangement.
- The failed pair of deals makes execution credibility more important for other AI startups seeking dedicated capacity: access to GPUs or a high valuation alone does not complete a data-center project.
Third-order effects
- If similar projects repeatedly require new sponsors after initial announcements, dedicated AI infrastructure will increasingly be governed by counterparties able to underwrite both compute and physical-project risk, favoring large cloud and infrastructure providers.
- The episode reinforces that AI infrastructure finance is not interchangeable with AI-company fundraising: capital commitments, power-backed sites, and operating partners must close on compatible terms.
The trend: AI startups are moving from simple cloud procurement toward capital-intensive, utility-like compute strategies, but their ability to execute depends on aligning financing with infrastructure partners.