Susana Schwartz
With latest UAE deal, Microsoft continues aggressive evolution from primarily software to cloud and AI infrastructure
In sharp contrast to President Trump’s statement on 60 Minutes that only U.S. customers should have access to the top-end Blackwell chips offered by Nvidia, Microsoft the next day announced the U.S. Commerce Department approved a deal to export 60,000 Nvidia GB300 Grace Blackwell chips to the United Arab Emirates. The announcement would open the door to AI models created by companies like OpenAI and Anthropic, and it would give Microsoft access to the UAE’s ambitious AI and cloud services initiatives and markets.
As we reported here, Microsoft has laid the groundwork the past two years to become a trusted AI partner to the UAE: first, with a $7.3 billion investment between 2023 and 2025, and now with $15.2 billion over the next few years for AI and cloud infrastructure. That will include $10 billion in capital spending and major datacenter projects in Abu Dhabi and Dubai. In exchange for these types of U.S. investments in the region, the UAE has pledged to invest $1.4 trillion in U.S. energy and AI-related projects (although the UAE’s annual GDP is approximately $540 billion).
To put Microsoft’s aggressive AI infrastructure spend into perspective, check out yesterday’s RCR Wireless Reader Forum, in which high-profile CEOs Chetan Sharma and Jim Brisimitzis examine Microsoft’s transformation from a “perpetual software company into a cloud powerhouse.” With substantial investment in infrastructure and chips, they contend Microsoft not only supports its own AI-enabled applications (like Copilot), but also opens critical access to other advanced AI models down the road. In essence, by becoming its own cloud platform for Windows Server, Office, and Dynamics, Microsoft has made what they call “a bold decision” to extend its influence in AI-focused regions like the UAE. Sharma and Brisimitzis contend this can be a lesson for the telecom industry, with network operators holding tremendous power in the data they move in the 5G, cloud and AI evolution toward data-driven actionable intelligence. Check it out, and be sure to scroll below for the latest AI infrastructure headlines.
Susana Schwartz
Technology Editor
RCRTech
AI Infrastructure Top Stories
Industrial AI: Deutsche Telekom and Nvidia contend they’ve launched “the world’s first Industrial AI Cloud,” a sovereign, enterprise-grade platform to accelerate Germany’s industrial transformation.
Telecom AI opps in AI: Networks are the platform, so can telecoms take a page from the Microsoft playbook and become more influential in AI infrastructure? CEOs Chetan Sharma and Jim Brisimitzis take a closer look.
Nokia-Softbank deal: Nokia and SoftBank signed network deal covers 4G modernization and 5G expansion across Western Japan with Nokia’s Habrok Massive MIMO radios and AirScale baseband units.
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
Enterprise AI: New research from TencentAI and Tsinghua University proposes “training-free” methods for improving large language models (LLMs), suggesting that models can learn and improve from on-the-job experience.
AI-grade fiber in NY: Lightpath will add approximately 100 route miles of new fiber — high-capacity backhaul and aggregation for bandwidth-intensive applications will strengthen critical edge presence and interconnect corridors.
Kerala chip strategy: The Indian state of Kerala is charting a different path with its design-first semiconductor roadmap that focuses on chip design, testing and IP creation rather than competing to build multi-billion-dollar fabs.
Grid AI-LV Grid: Purporting 165% growth to 50 GW by 2030, Entero subsidiary Grid AI will accelerate expansion into hyperscaler AI data center market with LV Grid and Grid AI agreeing to develop an AI Data Center Controls division.
Siri to get Gemini boost: After evaluating competitors, Apple has decided to supercharge Siri with Alphabet’s Gemini, whose 1.2 trillion parameters will underpin the next-gen version of Siri.
Nvidia’s Huang clarifies: Nvidia’s Jensen Huang softens alleged remark that “China will win the AI race,” clarifying that the United States must also “race ahead.”
