OpenAI is making a major shift in how it powers its massive technology systems. The creator of ChatGPT has officially pledged to purchase specialized processors directly from its biggest rival, Google. This surprising new agreement marks the first time OpenAI is actively moving away from relying completely on hardware made by NVIDIA. The move highlights the intense and growing need for computing power across the tech industry today.
High infrastructure costs push competitors to share resources
Building and training massive language models takes an incredible amount of expensive computing power. As of now, almost everyone relies on NVIDIA to supply the graphics processors needed for this heavy lifting. By striking a deal to buy TPUs from a major rival, OpenAI is looking to secure enough hardware to support its future projects without getting stuck waiting in long supply chain lines.
Interestingly, this is not the first time the two companies have shared resources. Back in 2025, OpenAI started renting out processors from Google to keep its software products running smoothly. While the two brands are fighting hard for dominance in the artificial intelligence software market, they are perfectly willing to sign commercial partnerships when it comes to the physical computing hardware they both desperately need.
Don’t miss the best of The Mac Observer
Set us as a preferred source and our Apple reporting ranks higher in your Google Search results and Discover feed — one tap, no account changes.
OpenAI develops a custom processor for future hardware independence
While buying these alternative chips offers a great short term fix, OpenAI is also planning ahead for long term independence. Reports suggest that OpenAI is actively developing its own custom designed AI chip, which is expected to launch sometime before the end of 2027. This internal hardware project, reportedly code named Jalapeño, recently showed early benchmarks that claim to beat the latest hardware offerings from NVIDIA.
Other major technology players are following very similar paths right now. Brands like Microsoft and Amazon are investing heavily in their own physical data centers and custom hardware to reduce their dependence on a single supplier. As the demand for computing infrastructure continues to skyrocket, companies have to find creative ways to secure the processors they need.
Ultimately, this new purchase agreement shows that practicality beats rivalry when it comes to hardware limits. By securing alternative processors now while building custom hardware in the background, OpenAI is making sure it has enough power to train its next generation of software models. It is a smart strategy that keeps OpenAI moving forward, no matter who supplies the physical parts.
Discussion