The geopolitics of compute
The new U.S. approval for Nvidia chip exports to the UAE marks more than a policy exception. By greenlighting billions of dollars in AI hardware shipments to American firms operating in the UAE, the Trump administration is threading a narrow line between containment and cooperation: enabling growth for domestic champions like Oracle and OpenAI abroad, while maintaining restrictions on local entities such as G42.
The move underscores how chip policy has become a key instrument in U.S. foreign strategy, determining who gets to scale intelligence infrastructure. As AI workloads converge with national security and energy diplomacy, each export license carries geopolitical weight, testing Washington’s ability to balance commercial ambition against concerns of technology leakage to China. The decision reflects a broader tension: the U.S. wants to lead the AI buildout globally, but only on trusted terrain.
Elsewhere, Qualcomm is redrawing the edge-AI map, integrating its recent acquisitions into a full-stack ecosystem for devices and industrial systems. And in Brazil, Elea’s $419 million Petrobras contract cements the country’s emergence as Latin America’s green AI infrastructure leader. Let’s dive in.
Juan Pedro Tomas
Editor
RCRTech
AI Infrastructure Top Stories
Nvidia gets UAE export nod: The U.S. approved multi-billion-dollar Nvidia chip exports for AI projects in the UAE, deepening a strategic tech alliance as Washington weighs security risks and Gulf nations invest heavily in AI infrastructure.
Qualcomm’s edge AI story: Through a series of acquisitions, Qualcomm has gradually developed an approach to edge AI spanning devices, data and developers. As the company executes on a diversification strategy, these moves appear key to bringing AIoT to industry.
Petrobras picks Elea for AI hub: Elea Data Centers won a Petrobras contract to build Latin America’s largest AI-ready, renewable-powered data center — a key step in Brazil’s rise as a green infrastructure hub.
AI-Powered Telecom Infrastructure
Supermicro, in collaboration with NVIDIA, delivers AI-powered infrastructure tailored for telcos, enhancing operational efficiency, network management, and customer experiences. Explore now
AI Today: What You Need to Know
Galaxy pivots to AI: Galaxy Digital will invest $460 million to transform its Texas Bitcoin mine into a large-scale AI data center, signaling a broader trend of crypto infrastructure repurposing for AI workloads.
Reflection AI scales up: Reflection AI raised $2 billion at an $8 billion valuation to expand global infrastructure and research on autonomous coding agents—underscoring investor demand for open-source frontier AI systems.
Taiwan bets on AI buildout: President Lai announced 10 new AI infrastructure projects to secure Taiwan’s global tech leadership, focusing on quantum computing, silicon photonics, and robotics as strategic national priorities.
Vertiv unveils OCP lineup: Vertiv introduced OCP-compliant rack, power, and cooling systems to accelerate high-density AI deployments, featuring modular architectures and 142 kW racks designed for next-gen compute environments.
Graphcore expands in India: SoftBank’s Graphcore will invest $1 billion in India over a decade, opening a new AI engineering campus in Bengaluru while doubling its UK headcount to boost semiconductor R&D.
TCS commits $7B to data centers: TCS plans to build 1 GW of data center capacity across India, investing $6–7 billion over seven years as part of its strategy to become the world’s top AI-led tech firm.
TikTok builds in Brazil: Brazil’s government confirmed construction of TikTok’s first data center in the country will begin in six months, a R$50 billion project powered by renewables and federal tax incentives.
Industry calls for standards: Leading hyperscalers, colocation providers, and manufacturers issued an open letter urging collaboration on AI data center infrastructure standards, targeting alignment in power, cooling, and mechanical interfaces.