Why Microsoft's AI Chip Supply Chain Awaits Power

Microsoft has all the AI chips it needs. The problem is, it can’t switch them on.

At the heart of its global supply chain, Microsoft faces a bottleneck — not in semiconductors, but in electricity.

Chief Executive Officer Satya Nadella sums up the dilemma: “The biggest issue we are now having is not a compute glut, but it’s power – it's sort of the ability to get the builds done fast enough close to power.

“So, if you can’t do that, you may actually have a bunch of chips sitting in inventory that I can't plug in. In fact, that is my problem today. It’s not a supply issue of chips; it's actually the fact that I don't have warm shells to plug into.”

In supply chain terms, Microsoft has inventory and components ready. What it lacks is the capacity to bring them online. This challenge is not about sourcing or stockpiling but distribution and infrastructure — specifically the physical facilities and energy resources needed to operate advanced AI systems at scale.

Infrastructure outpaced by ambition

Microsoft’s AI growth targets are backed by billions in investment. The company confirms plans to spend US$80bn on AI data centres across the 2025 fiscal year. This is part of a broader strategy supported by more than 400 sites across 70 regions, which act as critical hubs for global cloud and AI services.

Yet the pace of deployment isn’t matching the ambition. Although Microsoft continues to fund expansion, delays around grid connections, construction lead times and power procurement slow the actual delivery of usable capacity. In short, the data centre buildout cannot keep up with its supply chain or its demand forecast.

Earlier in 2025, Microsoft scaled back a planned 1.5GW self-build effort, opting instead to lease existing capacity. It commits US$11.1bn in Q1 2026 to secure ready-made facilities, shifting emphasis from self-construction to speed-to-market. This pivot reflects the growing urgency to overcome power-related constraints and streamline supply chain timelines.

The UK’s role in AI expansion

Microsoft’s long-term strategy includes major global investment, with the UK emerging as a key location. A US$30bn commitment to UK operations, spread between 2025 and 2028, aims to scale the country’s AI and cloud capabilities. Of that total, US$15bn is allocated specifically for capital expenditure.

Working with Nscale, Microsoft plans to deliver the UK’s largest supercomputer, powered by more than 23,000 GPUs produced by Nvidia. This system depends not just on chips but on the facilities and energy required to support them, further emphasising the need for dependable infrastructure across national supply chains.

Prime Minister Keir Starmer calls the deal “a powerful vote of confidence in the UK’s leadership in AI and cutting-edge technology”.

“This commitment," he says "will not only strengthen our digital infrastructure and support thousands of highly skilled jobs, but also ensure Britain remains at the forefront of global innovation as we deliver on our Plan for Change.

“We are proud to partner with world-leading companies like Microsoft to build a future powered by British ingenuity and ambition.”

Power supply issues dominate future capacity

Research from Bain & Company suggests utility connection timelines are now the biggest constraint on data centre growth, as some projects face five-year delays just to secure electricity access.

The same research forecasts a 163GW rise in global data centre electricity demand by 2030. Much of this growth is linked to generative AI, which requires intensive compute power and sustained infrastructure upgrades. In the US alone, demand could double to 409 terawatt-hours (TWh).

Meanwhile, TD Cowen reports that US hyperscalers, the largest cloud providers, have already leased more data centre capacity in Q3 2025 than the total leased across the entire year. Microsoft features prominently in that data, showing the market pressure to secure sites even before power availability is guaranteed.

As Bob Johnson, Vice President Analyst at Gartner, explains: “Significant power users are working with major producers to secure long-term guaranteed sources of power independent of other grid demands.

“In the meantime, the cost of power to operate data centres will increase significantly as operators use economic leverage to secure needed power. These costs will be passed on to AI/Gen AI product and service providers as well.”

In supply chain terms, this means longer lead times, higher costs and greater complexity. It also raises new questions for energy procurement, logistics and infrastructure planning — especially for AI and cloud providers operating at Microsoft’s scale.