GCC Data Centers and AI Infrastructure.
The market is built. The entry is the mandate.
GCC compute infrastructure has crossed from project scale to programme scale. Stargate UAE: 1 GW cluster inside a 5 GW UAE-US AI Campus, first 200 MW Q3 2026, financed and partnered with OpenAI, Oracle, NVIDIA, Cisco, SoftBank. HUMAIN: 6.6 GW target by 2034, 1.9 GW by 2030, $10B AMD JV, $5B AWS partnership, 18,000 GB300 chips authorised. Qatar: $20B Brookfield/Qai joint venture launched December 2025. NEOM Oxagon: $5B / 1.5 GW DataVolt campus.
The market is real. The entry is structurally hard. International firms entering this market face four binding constraints in parallel: partnership architecture with sovereign platforms that expect equity rather than vendor relationships; a regulatory layer that includes chip export controls, data residency, and entity registration; a talent pool that cannot be staffed from a Dubai regional office; and a procurement structure where selling into G42 or HUMAIN is not the same as selling into AWS or Microsoft.
The hyperscalers are not the competition. They are part of the puzzle. AWS opens its $5.3B Saudi region in 2026 alongside its HUMAIN partnership. Microsoft opens three Eastern Province data centers in 2026. Oracle is anchored in Stargate UAE and Saudi. These hyperscalers compete with sovereign platforms on capacity but anchor them on workloads. Any entry strategy that treats hyperscalers as either pure rivals or pure customers will fail. They are both — simultaneously.
The argument continues across 2 further sections of the paper.
The depth behind the thesis.
The page above is the argument. The paper is the evidence behind it: named transactions, sector-specific data, the regulatory references in full, and the closing position with its implications for capital allocation.
Twelve pages. Sent by email within one business day. By introduction, under standing confidentiality. The paper is not redistributed without permission.