Gulf AI buildout still hinges on Nvidia chips despite new deals

The Gulf’s AI ambitions are enormous—tens of billions in sovereign capital earmarked for compute—but this report is a reminder that hardware geopolitics still run through Nvidia’s stack. Even as Saudi Arabia and the UAE sign splashy agreements with AMD, Groq and Qualcomm, the biggest disclosed projects remain anchored on Nvidia GPUs and U.S.-controlled supply chains. That means that, in practice, much of the world’s new AI capacity is contingent on a single vendor and a handful of fabs.

For the race to AGI, this concentration cuts both ways. On one hand, massive, state-backed demand from the Gulf accelerates the buildout of frontier‑class clusters that labs everywhere can potentially tap, directly or via cloud partners. On the other, dependence on Nvidia and the U.S. export regime makes that capacity structurally fragile: a new round of controls, a supply shock in HBM, or a TSMC disruption would ripple through Saudi and Emirati plans just as they’re trying to position themselves as alternative AI hubs.

The effort to diversify suppliers is strategically rational but technically nontrivial. Non‑Nvidia hardware has to plug into existing software stacks, tooling and developer expectations. Until alternative ecosystems mature, the Gulf’s AI strategy will be less about true independence and more about hedging within a still‑Nvidia‑centric world.