Microsoft stock has a new number for investors to digest: 38 gigawatts. Microsoft is reportedly planning to expand its global data-center capacity to roughly that level by 2032, more than tripling its current footprint as the company races to secure enough computing power for artificial intelligence and traditional cloud workloads. The target, reported by Bloomberg and subsequently confirmed in reporting by Reuters, would represent an extraordinary infrastructure expansion over the next six years and underscores how aggressively Microsoft believes demand for AI computing will continue growing.
The scale sounds almost industrial rather than technological, and that is increasingly the point. Microsoft is no longer simply selling software and renting server capacity. Its future growth is becoming tied to enormous physical assets: land, data centers, networking equipment, GPUs, CPUs and, above all, electricity. Microsoft already expects approximately $175 billion in calendar-year 2026 capital expenditures, and management has said fiscal 2027 spending will rise again as it attempts to bring more capacity online. Yet despite those extraordinary investments, customer demand continues to exceed available Azure capacity.
That tension is what makes the reported 38-gigawatt target so important for Microsoft stock. There is little doubt that customers currently want more compute than Microsoft can provide. The much harder question is whether demand will remain strong enough through 2032 to justify an infrastructure footprint more than three times its current size — and whether the economics remain attractive after hundreds of billions of dollars have been committed to building it.
Microsoft Is Building Because Azure Still Does Not Have Enough Capacity
A company does not normally triple infrastructure capacity unless management believes it can fill it, and Microsoft has spent several quarters telling investors exactly why it thinks it can. Azure and other cloud-services revenue increased 43% year over year in Microsoft’s fiscal fourth quarter, significantly faster than the overall company, and management said customer demand continued to exceed the capacity available to serve it. Microsoft Cloud revenue reached $59.3 billion for the quarter and $214.4 billion for fiscal 2026, while commercial remaining performance obligations climbed to $678 billion, giving the company a massive pool of contracted business that will be recognized over time.
The capacity shortage is not new. During the fiscal third quarter, Microsoft said strong demand across workloads, customer types and geographic regions was already exceeding supply, even after it brought infrastructure online faster than anticipated. Management expected the company to remain constrained through at least calendar 2026, despite a dramatic increase in spending on GPUs, CPUs, storage and data-center construction. That makes the reported 2032 plan easier to understand. Microsoft is not merely betting that AI will create demand someday; it is attempting to solve a capacity problem that exists now while simultaneously preparing for what it believes could be a much larger computing market later.
Related articles
The challenge is that today’s shortage does not guarantee a shortage six years from now. AI models could become more efficient, customers could shift workloads toward specialized chips, competitors could add supply faster than expected, or the industry could eventually discover that some of today’s projected computing requirements were too optimistic. Microsoft therefore has to build ahead of demand without building so far ahead that billions of dollars of expensive infrastructure sit underutilized.
The 38-Gigawatt Number Shows How Physical the AI Boom Has Become
Thirty-eight gigawatts is not merely a data-center statistic. It is a power requirement large enough to highlight how thoroughly the AI boom has transformed the economics of Big Tech. Modern AI infrastructure combines huge clusters of accelerators with sophisticated cooling systems, high-speed networking and enormous electricity demand, meaning the competitive battle between Microsoft, Amazon, Alphabet and other cloud providers increasingly depends on their ability to obtain power and build physical infrastructure as much as their ability to write software.
Microsoft had already said in late 2025 that it intended to increase total AI capacity by more than 80% in one year and roughly double its overall data-center footprint within two years. At the time, management pointed to projects such as its Fairwater AI data center in Wisconsin, which was designed to scale toward two gigawatts on its own. The newly reported 38-gigawatt target suggests the company is now thinking far beyond that shorter-term expansion and planning for an AI infrastructure footprint that extends well into the next decade.
Finding enough electricity may become as important as finding enough chips. U.S. lawmakers are already responding to concerns that growing data-center demand could raise electricity prices for households, with legislation aimed at preventing infrastructure costs from being shifted onto ordinary ratepayers. At the same time, hyperscalers are increasingly expanding abroad because U.S. projects face permitting delays, disappearing tax incentives, local opposition and difficulties obtaining enough power quickly. Microsoft itself is pursuing overseas expansion alongside its U.S. buildout, illustrating how access to energy and construction capacity is becoming a strategic constraint on AI growth.
For investors, this means Microsoft’s competitive advantage may increasingly depend on something rarely associated with software companies: the ability to secure vast quantities of reliable electricity at attractive prices.
Microsoft Is Already Spending at a Scale That Would Have Looked Unthinkable Years Ago
The reported 2032 target becomes more striking when set beside Microsoft’s current capital expenditures. The company spent $41 billion in capital expenditures during its fiscal fourth quarter alone, with roughly two-thirds directed toward shorter-lived assets such as CPUs and GPUs. Microsoft now expects approximately $175 billion of calendar-year 2026 capex after accounting for changes in the classification of data-center leases, while fiscal 2027 spending is expected to increase again.
Those numbers illustrate how dramatically Microsoft’s financial model is changing. The company built its reputation on software economics, where an additional copy of Windows or Office could be sold at exceptionally high incremental margins. Cloud computing already required far more infrastructure, but generative AI has accelerated the capital intensity further because leading models need enormous numbers of expensive chips and the facilities required to operate them.
That pressure is already visible in Microsoft’s margins. Fiscal fourth-quarter company gross margin was 67%, down from the prior year partly because of the shift toward Azure and continued investment in AI infrastructure and higher product usage. Microsoft has nevertheless managed to offset some of that pressure through efficiency improvements, and management expects full-year fiscal 2027 operating margins to decline by less than one percentage point even as investment continues.
That may be one of the strongest arguments supporting the Microsoft stock bull case. The company is spending enormous sums, but it is not currently sacrificing its entire profit structure to do so. If Microsoft can keep revenue and operating income expanding while building toward 38 gigawatts, investors may tolerate a capital expenditure bill that would be alarming for almost any other company.
Azure Is Giving Investors a Reason to Accept the Spending — for Now
The willingness to fund this buildout ultimately rests on Azure. Microsoft’s fiscal fourth-quarter Azure and other cloud-services revenue rose 43%, comfortably exceeding analyst expectations of roughly 40%, while the company projected another strong start to fiscal 2027. Microsoft’s overall quarterly revenue reached $90 billion, up 18%, operating income rose 18% to $40.6 billion, and adjusted earnings increased 23% to $4.74 per share.
Those results matter because the biggest fear surrounding Big Tech’s AI spending spree has been that investment could grow much faster than monetization. Microsoft is currently offering evidence of the opposite: infrastructure spending is exploding, but cloud revenue is accelerating too. Microsoft 365 Copilot has surpassed 30 million paid seats, Azure growth remains above 40%, and commercial contracted obligations continue climbing. That gives management a stronger argument that the company is responding to real customer demand rather than constructing speculative capacity.
Still, the burden of proof becomes larger with every new data-center announcement. A few years of strong Azure growth can justify today’s investments. A 38-gigawatt footprint by 2032 requires Microsoft to make assumptions about demand much further into the future, when AI architectures, chip economics and competitive dynamics could look very different from today.
That is where the investment case becomes less comfortable.
The Biggest Risk Is Not That AI Fails — It Is That Microsoft Overbuilds
The danger does not require artificial intelligence to disappear. AI could become enormously important and Microsoft could still build too much capacity, too early, or at the wrong cost. Technology history is filled with periods when a legitimate long-term trend inspired infrastructure spending that temporarily outran demand. Reuters Breakingviews recently compared the current data-center expansion to the late-1990s broadband boom, when enormous investment eventually produced valuable infrastructure but also destroyed capital for companies whose growth assumptions proved too aggressive.
The parallel is not perfect. Microsoft has advantages that speculative infrastructure companies lack: huge recurring revenue, a diversified customer base, powerful software franchises and one of the strongest balance sheets in corporate America. It is also generating substantial cash while investing. Nevertheless, the underlying lesson remains relevant. A technology can transform the world while individual infrastructure investments still produce poor returns if too much capacity is built.
Financing conditions add another layer. Microsoft, Alphabet, Amazon, Meta and Oracle have collectively helped drive a surge in bond issuance as the technology industry funds increasingly expensive AI infrastructure. Reuters reported that major hyperscalers issued around $220 billion in bonds over the previous year, showing how the AI race is spreading from semiconductor and equity markets into credit markets.
For Microsoft, the balance sheet is not yet the central concern. Return on invested capital is.
What Microsoft Stock Investors Should Watch as the 2032 Buildout Begins
The reported 38-gigawatt plan makes several metrics increasingly important. Azure growth must remain strong enough to justify continual capacity additions, while Microsoft Cloud margins need to show that efficiency improvements can offset the higher depreciation and operating costs associated with AI infrastructure. Investors should also watch capital expenditures, free cash flow and commercial remaining performance obligations because together they reveal whether future customer commitments are expanding quickly enough to support the construction pipeline.
Power availability may become equally important. Data-center projects can be delayed even when customer demand and financing are available simply because grids cannot supply enough electricity or transmission capacity. Rising political resistance to data-center energy use could complicate expansion further, particularly if governments require hyperscalers to shoulder a larger portion of generation and grid-upgrade costs. Microsoft’s international expansion may therefore become not just a geographic growth strategy but a necessity for securing enough power to reach its reported target.
Microsoft stock finished September 10 at roughly $492 after edging higher in an otherwise weak market session, suggesting the 38-gigawatt report itself did not trigger a dramatic reassessment. But the longer-term significance is much larger than one day’s price move.
Microsoft is effectively betting that computing demand in the AI era will become so enormous that even today’s record infrastructure spending will eventually look insufficient.
For now, Azure’s growth and Microsoft’s $678 billion commercial backlog give investors reasons to believe that demand is real. The company is already capacity constrained, cloud revenue is accelerating, and management says it can continue investing while remaining strongly profitable and free-cash-flow positive.
But a plan to more than triple data-center capacity by 2032 moves the argument beyond whether AI is popular today.
The question for Microsoft stock is whether six years from now the company will look brilliant for securing 38 gigawatts of computing capacity before everyone else — or whether investors will discover that the most expensive part of the AI boom was building enough infrastructure for demand that never became quite as large as expected.
Disclaimer
This article is for informational purposes only and does not constitute financial or investment advice. Readers should conduct their own research and consider consulting a qualified financial advisor before making investment decisions. This article was researched and drafted with the support of AI, then reviewed, fact-checked, and edited by the editorial team before publication.
