Microsoft has shut down its Internet of Things (IoT) and Artificial Intelligence (AI) lab in Shanghai's Zhangjiang Hi-Tech Park, marking another step in the tech giant's retreat from China. According to the South China Morning Post, the facility has been completely cleared out, with signage removed and no official explanation yet provided by Microsoft.

The lab, launched in May 2019, was one of only seven Microsoft AI & IoT labs worldwide. Over five years, it supported 258 projects from 100 companies, trained nearly 10,000 professionals, and helped over 50 firms secure more than RMB 9.4 billion in external investment. Despite its contributions, Microsoft has continued to scale back operations in China, including layoffs, offering relocation to AI staff, and closing all physical retail stores on the mainland.

In a 2024 U.S. congressional hearing, Microsoft President Brad Smith noted that China accounts for just 1.5% of the company's global revenue. Following the lab's closure, Microsoft will continue operating its other AI & IoT labs in the U.S., Germany, Uruguay, and Japan.

Microsoft isn't alone in shifting away from China. HP has relocated its laptop production lines to Mexico and Thailand, shutting down its Shanghai operations, while IBM has closed several R&D units in China, including its China Development Lab (CDL) and Systems Lab (CSL), affecting over a thousand engineers. Analysts cite rising operational costs, geopolitical tensions, and increasing local competition as key drivers of this trend. However, many see it as part of China's broader tech industry upgrade, creating more room for domestic firms to grow.

Beyond China, Microsoft is also reevaluating its global infrastructure strategy. The company has paused or scaled back several data center projects in regions including the U.S., U.K., Australia, and Indonesia. Projects in Illinois, Wisconsin, North Dakota, and central England have been delayed or shelved. A $262 million project in Mount Pleasant, Wisconsin has been halted, while talks for a 210 MW facility near London's Docklands have stalled.

Insiders say Microsoft is reassessing whether the projected demand for AI services justifies its aggressive data center expansion. A spokesperson stated that Microsoft is adjusting capacity planning to reflect evolving needs: “As demand for AI continues to grow, we are making strategic adjustments to ensure we have the right infrastructure in the right locations.”

In contrast, other tech giants are ramping up infrastructure investment. OpenAI, Oracle, and SoftBank recently announced a $500 billion joint venture, Stargate, to build AI infrastructure, including a $100 billion phase-one project in Texas.

While Microsoft still plans to spend about $80 billion on data centers this fiscal year, the focus may shift next year toward upgrading existing facilities rather than building new ones. Regulatory concerns about power consumption and environmental impact in cities like Dublin and Amsterdam have also added pressure, leading to tighter controls on data center expansion.

Industry analysts suggest this is a broader trend among hyperscalers, with a growing preference for faster, more cost-effective infrastructure deployment. As Microsoft recalibrates, competitors like OpenAI may be redirecting workloads elsewhere—potentially to Oracle. Meanwhile, advances in AI efficiency, such as from China's DeepSeek, are also prompting a reassessment of future compute requirements.