Cloud Computing & Data Centers

What's Happening in Cloud Computing & Data Centers

U.S. export controls on advanced semiconductors are failing in enforcement while simultaneously driving Chinese firms toward workarounds and alternative supply chains, fragmenting the global compute infrastructure market and undermining the stated purpose of restrictions. The Commerce Department has granted export licenses to Chinese companies with shell structures and problematic affiliations, while a short seller alleges a broader network of intermediaries facilitating restricted chip access; Chinese AI developers like Moonshot are leasing Nvidia chips through European cloud providers to circumvent U.S. restrictions, and DeepSeek and Huawei are developing competing software and chip capabilities to reduce dependence on American suppliers. Simultaneously, major U.S. cloud providers are diversifying away from concentrated reliance on Nvidia through internal chip development and partnerships, while new infrastructure models—underwater datacenters reducing cooling costs, space-based AI systems addressing thermal and power constraints, and orbital manufacturing for specialized applications—are accelerating buildout of compute capacity to meet surging AI demand. The net effect is a bifurcating market: U.S. and allied cloud providers investing in redundant, distributed, and alternative compute architectures to hedge geopolitical risk and operational constraints, while Chinese competitors are forced to innovate around restrictions rather than compete on cost or performance, raising the structural cost of compute infrastructure globally and compressing margins across semiconductor distribution, cloud services, and data center operations.