Source: G42 is talking to Google, AWS, Meta, Microsoft, and xAI to become tenants at its UAE-US AI Campus, and aims to diversify chip suppliers beyond Nvidia
THE SCOOP — Abu Dhabi-backed G42 is still negotiating with major American technology firms — with no agreements yet …
Context & Ripple Effects
G42’s campus outreach extends a deliberate pivot toward US technology partners after it moved away from Chinese hardware suppliers and Microsoft took a minority stake in the company in 2024. The talks are not agreements, but they test whether that alignment can translate into tenant demand for UAE-based capacity.
The supplier strategy also broadens an earlier Nvidia-centered path: G42 later described planned deliveries from Nvidia, AMD, and Cerebras for its UAE hub. Its US corporate build-out further indicates that the campus effort sits alongside a wider expansion of G42’s US presence.
First-order effects
- Google, AWS, Meta, Microsoft, and xAI are potential tenants rather than committed customers, leaving G42 with negotiations to convert into anchor demand for the campus.
- G42 gains a clearer basis to seek chips beyond Nvidia, potentially matching different suppliers to tenant requirements if it secures capacity and deployments.
Second-order effects
- Prospective tenants gain another possible route to regional AI infrastructure, while G42 can use multivendor sourcing to reduce its exposure to any single chip supplier’s availability or terms.
- Nvidia faces a more competitive account opportunity as AMD and Cerebras become part of G42’s stated supply mix; rival vendors have a stronger incentive to support deployments that can attract large US platforms.
Third-order effects
- If tenant agreements materialize, the model would shift G42 further from an AI buyer into an infrastructure intermediary: assembling capital, compliant supplier relationships, and compute access for global platforms.
- The pattern points to state-linked AI infrastructure becoming a vehicle for cross-border capacity allocation, where supplier eligibility and government alignment can shape where leading-model workloads run.
The trend: This is one data point in the rise of state-mediated AI infrastructure, where sovereign-backed operators package chips, power, compliance, and tenant access into a single strategic platform.