According to a Bloomberg News report cited by Reuters, Microsoft plans to expand its global data center capacity to around 38 gigawatts by 2032 — more than tripling its current infrastructure footprint.
The ambitious expansion highlights just how rapidly the world’s biggest technology companies are scaling their physical infrastructure to support the growing demand for artificial intelligence, cloud computing and increasingly power-hungry AI models.
But Microsoft’s massive AI bet could face a major obstacle: finding enough electricity to power it all.
Microsoft Reportedly Plans to Triple Data Center Capacity
Microsoft currently has about 12 gigawatts of data center capacity, according to reporting on the Bloomberg plan. Reuters reported that the company is targeting roughly 38GW by 2032, which would represent a more than threefold expansion.
The scale of the reported plan underscores the extraordinary infrastructure demands created by artificial intelligence.
AI services require enormous numbers of advanced chips, servers, networking systems and cooling equipment. Unlike traditional cloud computing, training and operating large AI models can require vast amounts of computing power operating continuously.
Microsoft has already publicly confirmed that it is rapidly expanding its infrastructure.
During its fiscal 2026 first-quarter earnings call, the company said it expected to increase total AI capacity by more than 80% during the year and roughly double its overall data center footprint within two years. Microsoft also said its Fairwater AI data center in Wisconsin is designed to scale to 2 gigawatts.
More recently, Microsoft said it added another gigawatt of capacity during its fiscal fourth quarter and remained on track to roughly double its overall capacity within two years.
AI Could Take a Bigger Share of Microsoft’s Infrastructure
The Bloomberg-reported expansion is not simply about building more conventional cloud capacity.
According to reporting on the plan, only about 2GW of Microsoft’s current 12GW capacity is dedicated to AI-specific chips, but AI infrastructure could account for approximately one-third of the company’s projected 38GW capacity by 2032.
That would represent a dramatic increase in the amount of infrastructure dedicated specifically to artificial intelligence.
The strategy reflects a growing problem across the technology industry: companies are increasingly competing not just for AI models and talent, but for the physical infrastructure required to run them.
That includes:
- Advanced AI chips
- Massive server fleets
- Data center land
- Electricity generation
- Grid connections
- Cooling infrastructure
- High-speed networking
In other words, the AI race is becoming an infrastructure race.
Microsoft Is Already Building at Breakneck Speed
Microsoft’s recent expansion demonstrates that the reported 2032 target is part of a much broader infrastructure push already underway.
Data Center Dynamics reported that Microsoft brought 88 data centers online during fiscal 2026. The company added 31 data centers across five continents in the final quarter alone, contributing approximately 1GW of additional capacity.
Microsoft has also been working to improve the efficiency of its existing infrastructure rather than relying solely on new construction.
The company said it has optimized systems across silicon, servers and software, while improving throughput for AI workloads. Microsoft is also expanding the use of its own AI chips alongside technology from Nvidia and AMD.
The Biggest Challenge May Not Be Chips — It Could Be Electricity
Microsoft’s reported 38GW target illustrates a growing challenge facing the global AI industry: power availability.
Building a data center is already expensive and complex. Connecting that facility to enough reliable electricity can be even more difficult.
Across the technology industry, companies are competing for access to power grids that were never designed for such rapid growth in electricity demand.
Reuters recently reported that AI-related data center construction is increasingly facing delays and financing pressure because of supply-chain problems, power constraints and political opposition.
The pressure is also becoming increasingly visible outside the United States.
In Malaysia, Reuters reported that data centers accounted for a record 9.3% of electricity consumption in mid-August 2026, amid soaring temperatures that increased cooling requirements. Officials warned that data center electricity demand could account for up to 31% of Peninsular Malaysia’s power consumption by 2035.
That highlights a crucial reality for Microsoft and other AI giants: building more data centers is only part of the challenge.
They also need enough electricity to keep them running.
Big Tech’s AI Spending Spree Is Getting Bigger
Microsoft is far from alone.
Amazon, Alphabet, Meta, Oracle and other major technology companies are pouring billions of dollars into AI infrastructure as they compete for cloud customers and AI leadership.
Reuters reported this week that hyperscale technology companies have issued approximately US$220 billion in bonds over the past year to help finance their expanding AI and data center ambitions.
The enormous spending has raised questions among investors about whether the industry is building infrastructure faster than future demand can justify.
At the same time, companies argue that demand for AI computing remains strong and that capacity shortages are already limiting growth.
Reporting on Microsoft’s reported 2032 plan suggests the company has faced capacity constraints that have affected its ability to meet demand for cloud servers and AI computing resources.
That makes the reported 38GW target about more than future growth.
It could also be Microsoft’s attempt to ensure it does not fall behind in the increasingly fierce battle for AI computing capacity.
Microsoft Faces a Global Race for AI Infrastructure
The competition is no longer limited to Silicon Valley.
Data center construction is accelerating across Europe, Asia and the Middle East as governments and technology companies seek to secure computing capacity closer to users and energy resources.
Microsoft and other hyperscalers are increasingly looking outside the United States as construction costs, regulatory hurdles and power shortages complicate new projects.
Recent reporting has highlighted major overseas investments by technology companies as they seek locations with faster access to electricity, renewable energy and cooler climates that can reduce cooling costs.
The trend could reshape the global technology landscape.
Countries with abundant power, strong grids and favorable infrastructure policies could become increasingly important hubs for the next generation of AI computing.
Is the AI Infrastructure Boom Becoming Too Big?
Microsoft’s reported 38GW plan also comes as concerns grow about the financial risks surrounding the AI infrastructure boom.
Reuters Breakingviews compared the rapid expansion of AI-focused data center companies to earlier infrastructure booms, warning that aggressive investment and high debt levels could create vulnerabilities if demand expectations change.
That does not mean demand for AI computing will disappear.
However, the industry faces significant questions:
Can power grids expand fast enough?
Can companies secure enough chips?
Will communities continue accepting giant data centers?
Can technology companies generate enough revenue to justify hundreds of billions of dollars in infrastructure spending?
And perhaps most importantly:
What happens if everyone builds massive AI capacity at the same time?
Microsoft’s 38GW Plan Shows How Serious the AI Race Has Become
For now, Microsoft’s reported 2032 target demonstrates the extraordinary scale of the technology industry’s AI ambitions.
Just a few years ago, artificial intelligence was largely discussed as a software revolution.
Today, it is increasingly becoming a battle over factories, chips, land, power plants, electricity grids and data centers.
Microsoft is betting that demand for AI computing will continue growing for years.
Its reported plan to reach approximately 38 gigawatts of data center capacity by 2032 would make it one of the most aggressive infrastructure expansions ever attempted by a technology company.
But the real question may not be whether Microsoft can afford to build the data centers.
The bigger question is whether the world’s energy systems can keep up with the AI revolution Microsoft and its rivals are trying to create.
Leave a Reply