Saudi Arabia, UAE Build $300 Billion Leverage Play While Washington Demands Loyalty

Southeast Asia and the Gulf are forging a multibillion-dollar AI partnership that neither China nor the United States saw coming. While Washington prepares to force 35 allied nations to choose sides in what a leaked State Department letter calls an AI loyalty test, two regions stuck in the middle have stopped waiting for permission and started writing their own playbook .

The numbers tell the story. Saudi Arabia is deploying $100 billion through Project Transcendence into AI infrastructure and startups. The United Arab Emirates has committed $148 billion to AI buildout over two years, with a single project, Stargate UAE, carrying a $30 billion price tag for a 1-gigawatt computing cluster built with G42, OpenAI, Oracle, NVIDIA, SoftBank, and Cisco . Across the South China Sea, Southeast Asia's digital economy is projected to hit $300 billion in gross merchandise value by 2025, while its data center market will more than double from $13.7 billion to $30.5 billion by 2030 .

Neither region produces frontier AI models or fabricates cutting-edge semiconductors. But together, they control what matters next: capital, energy, infrastructure velocity, demand, and crucially, optionality in a world where the U.S. and China together produced 80 of the notable AI models released last year and dominate the only two countries running public cloud AI on domestically manufactured chips .

This is not a hedge. It is a deliberate architecture for middle powers who refuse to be rule-takers.

The Leverage Play Hiding in Plain Sight

The United States operates 5,427 data centers, more than ten times any other country . China and the U.S. have built full-stack AI ecosystems, from silicon to inference. The rest of the world, including Europe's regulatory-first approach and the hodgepodge of national strategies elsewhere, largely consumes what the duopoly produces.

Southeast Asia and the Gulf have accepted they will not compete at the frontier. What they are building instead is structural leverage through complementary dependencies. The Gulf brings the trifecta of speed, scale, and cheap energy. Southeast Asia delivers demand density, digital-native populations, a rapidly growing fintech sector, and regulatory frameworks already in motion. Vietnam enacted a standalone AI law and positioned itself as a regional AI hub, hosting over a quarter of Southeast Asia's generative AI startups . Malaysia pulled in $23 billion in data center investment in 2024 and signed formal AI and digital economy agreements with the UAE .

The capital flows match the ambition. G42 signed a framework agreement with a Vietnamese consortium in February 2026 worth up to $1 billion to build national AI and cloud infrastructure spanning government systems, industrial applications, and academia . Masdar, Abu Dhabi's state-owned clean energy firm, inked a $15 billion renewables deal with the Philippines in January 2025 . Project mBridge, a cross-border digital currency initiative, links the UAE, Thailand, and Saudi Arabia . The Joint Declaration on Economic Cooperation between ASEAN and the GCC explicitly endorsed AI and digital economy collaboration, giving the trend formal political backing .

The structure is intentionally modular. No region-wide consensus is required. Each deal matches a specific capability to a specific need: UAE capital and infrastructure buildout against Vietnam's regulatory maturity and startup ecosystem, Abu Dhabi's energy capacity against Malaysia's data center appetite, Gulf sovereign wealth against Southeast Asia's underbanked populations ripe for AI-driven financial inclusion.

What makes this durable is what it avoids: dependency on a single geopolitical bloc.

The Loyalty Test and the Escape Hatch

Washington's leaked ultimatum to the 35 signatories of its June AI Opportunity Statement makes the stakes explicit. "To be part of everything is to be part of nothing," the draft State Department letter warns . The mechanism is Pax Silica, an initiative designed to lock down AI model supply chains, semiconductors, and critical minerals. Nations that join gain access to shared AI investment opportunities. Those that also participate in China's competing World Artificial Intelligence Cooperation Organization, launched by Xi Jinping in July with an explicit open-weight technology pitch, face exclusion .

One U.S. official told The Next Web: "It's difficult to see how a country can credibly position themselves as trusted partners in one technology ecosystem while simultaneously signing up for an initiative designed by China to advance a competing vision for AI" . The subtext is stark. Neutrality reads, in Washington, as defection.

Only Kazakhstan currently sits in both camps, a function of its critical mineral reserves . That position is now untenable. The letter is designed to end hedging. For Southeast Asia and the Gulf, both of whom have resisted binary alignment, the timing could not be worse, or more clarifying.

China is not sitting idle. Beijing is weighing curbs on overseas access to its best models, a sign of newfound leverage as cheap Chinese open-weight models close the gap with Anthropic and OpenAI on cost and capability . The pitch to the Global South is explicit: openness, access, and no political strings.

The Southeast Asia-Gulf partnership offers a third option. It does not require signing loyalty oaths to either camp. It builds on bilateral deals, sectoral pilots, and mutual economic interest. It treats AI as infrastructure, not ideology.

Sectoral Pilots as Strategic Scaffolding

The practical architecture already exists. Recent research from the AI Asia Pacific Institute documents the pattern across UAE-Singapore memoranda on digital government, UAE-Malaysia cooperation under the MADANI AI initiative, and the UAE-Vietnam framework . Each is structured as a sectoral pilot, testing AI applications in domains suited to each region's endowments.

Southeast Asia, with large underbanked populations and exploding fintech adoption, is the natural testbed for financial inclusion, small business credit, and payment infrastructure. The Gulf, with growing state capacity and legal digitization efforts, is positioned to pilot AI in courts, public administration, and regulatory automation .

These are not vanity projects. They generate defensible use cases, real-world training data in languages and regulatory contexts the U.S. and China do not prioritize, and exportable learnings to other middle powers facing similar constraints. The same model extends to healthcare, logistics, energy management, and language models trained on Arabic and Southeast Asian languages. G42's Jais 2 model, trained on 600 billion Arabic tokens, is building a homegrown Arabic-language AI ecosystem . There is no equivalent effort at scale for Bahasa Indonesia, Vietnamese, or Thai, languages spoken by hundreds of millions.

The deal-by-deal structure also insulates against geopolitical shocks. When the Iran-U.S.-Israel conflict escalated, Gulf investors did not pause. They accelerated diversification, treating Southeast Asia as a hedge against overconcentration in any single market . Vietnam's energy transition, which saw solar and wind rise from 0.4 percent of electricity generation in 2018 to 13 percent in 2024, demonstrates the appetite for infrastructure-scale partnerships . The renewable energy buildout came at significant cost to state utility EVN, but it also created the foundation for energy-intensive AI infrastructure .

Capital Allocation in a Fragmented Order

For institutional capital, the Southeast Asia-Gulf corridor represents a structural reallocation opportunity disguised as a bilateral trade story. The total addressable market is not the $148 billion UAE commitment or the $300 billion Southeast Asia digital economy in isolation. It is the compounding effect of pairing capital surplus regions with demand surplus regions outside the U.S.-China duopoly.

Sovereign wealth funds, pension allocators, and late-stage venture capital have spent two years trying to underwrite AI exposure without direct exposure to U.S.-China decoupling risk. This corridor offers that. It is not frontier model risk. It is infrastructure, application layer, and services risk, with government balance sheets and state-backed entities like G42, Masdar, and national utilities on the other side of the term sheet.

The risk is execution, not access. Malaysia's $23 billion data center pipeline depends on energy capacity, regulatory stability, and grid resilience, all of which remain question marks in a region where EVN's balance sheet strain from renewable commitments shows the cost of moving fast without pricing discipline . Vietnam's AI law and startup density are strengths, but the $1 billion G42 framework is just that, a framework. Converting commitments into operating assets at scale is where capital gets tested.

The geopolitical overlay adds a timing dimension. If Washington's Pax Silica ultimatum forces Southeast Asian and Gulf nations to make binary choices, the corridor weakens. If it accelerates the search for non-aligned partnerships, it strengthens. The leaked State Department letter suggests the former is Washington's intent . But revealed preference, the actual flow of capital and signed MOUs, suggests the market is betting the other way.

The Plocamium View

The AI cold war narrative assumes two poles and a set of vassal states forced to choose. What Southeast Asia and the Gulf are building is proof that middle powers can construct a third option if they move with speed, specificity, and complementary factor endowments.

The conventional wisdom holds that full-stack AI capability is the only durable position. That frontier models, proprietary chips, and hyperscale cloud infrastructure are the choke points, and everyone else is downstream. We think that misreads where value accrues in a fragmented technology order.

The U.S. and China will continue to race on model capability and chip design. But application layer dominance, the ability to deploy AI at scale in specific verticals, languages, and regulatory regimes, does not automatically follow from frontier research. It requires local data, local partnerships, and patient capital willing to build for regional use cases, not global platform dreams.

The Southeast Asia-Gulf partnership is structurally positioned to capture that layer. The Gulf has the capital and energy to build infrastructure faster than anyone except China. Southeast Asia has the demand, the demographic tailwinds, and the regulatory maturity to absorb it. Neither has the geopolitical baggage of being a rule-maker, which in 2026 is an asset, not a liability.

The second-order play is the precedent this sets for other middle power corridors. If the UAE-Vietnam framework works, the same architecture applies to Gulf-Latin America, Southeast Asia-Africa, or any other pairing where capital surplus meets demand surplus outside the ideological camps. The modularity is the feature. No multilateral consensus required. No WTO-style negotiation rounds. Just bilateral deals that scale horizontally.

The risk we are watching is whether Washington's loyalty test fractures the corridor before it scales. If Pax Silica becomes a real constraint, forcing Gulf states or ASEAN members to choose access to U.S. chip supply chains or Chinese open models, the deal flow slows. Our base case is that economic gravity wins. The Gulf has too much capital to deploy, Southeast Asia has too much growth to finance, and both have too much to lose from subordination to either camp. But the next six months will test that thesis hard.

The Bottom Line

The AI economy is not converging on a single set of rules, platforms, or supply chains. It is fracturing into regional blocs with different capital structures, regulatory philosophies, and strategic objectives. Southeast Asia and the Gulf are building the first successful middle power alternative, one that does not chase self-sufficiency at the frontier but instead leverages complementary strengths into durable infrastructure and application layer positions.

For institutional allocators, the trade is clear: underweight U.S.-China binary exposure, overweight the corridors that benefit from fragmentation. The Southeast Asia-Gulf partnership is the template. Watch the deal flow, not the diplomatic cables. When G42 converts frameworks into gigawatts and Vietnam's AI startups start training models on Arabic and ASEAN language datasets, the architecture becomes undeniable. The next phase of AI will not be built only in Silicon Valley and Shenzhen. It will be financed in Abu Dhabi and deployed in Ho Chi Minh City, and the capital that recognizes that early will capture the asymmetry.