AI boom pushes data-center dealmaking to record $61B through November

The race for computing power has become the defining investment story of the year. Through November, global dealmakers have channeled nearly $61 billion into data centers—a figure that already eclipses the previous record set just twelve months earlier. More than 100 separate transactions moved through the market during this period, each one a bet that artificial intelligence will continue to demand ever more infrastructure to run on.

The surge reflects something fundamental about where the technology industry stands right now. Tech giants and the companies building AI systems at scale have committed to spending billions on new data-center capacity. They need the physical infrastructure—the servers, the cooling systems, the power supplies, the real estate—to support the computational demands of training and running large language models and other AI applications. Without that infrastructure, the software cannot exist. So the infrastructure itself has become the thing to own.

The numbers tell the story with clarity. Across mergers, acquisitions, asset sales, and equity investments, the data-center sector has absorbed $61 billion in capital through the end of November alone. This already surpasses what the entire year of 2024 produced: $60.81 billion. The velocity is striking. Since 2019, the United States and Canada combined have seen roughly $160 billion flow into data centers. Asia-Pacific has attracted nearly $40 billion over the same span. Europe has drawn $24.2 billion. But none of those regional totals are moving at the pace we're seeing now.

Who is doing the buying? Financial sponsors and private equity firms have emerged as the most aggressive players. They are drawn to data centers because the assets themselves—the buildings, the equipment, the long-term contracts with tenants—offer a particular kind of appeal: steady cash flows paired with the potential for significant upside as demand continues to climb. The risk-reward calculation, from their perspective, makes sense. But this appetite has created a peculiar market dynamic. Private equity buyers are eager to acquire, yet reluctant to sell. The best assets rarely come on the market. When they do, competition is fierce.

This scarcity of high-quality data-center assets available for purchase has tightened the market considerably. Sellers know what they have. Buyers know the supply is limited. The result is a seller's market where valuations climb and deal structures become more creative. For investors, it means opportunity—if you can find something to buy. For companies trying to build or expand their data-center footprint, it means costs are rising and options are constrained.

Yet beneath the dealmaking frenzy sits a question that has begun to nag at observers. The companies pouring money into AI infrastructure—the hyperscalers, the tech giants, the startups—are doing so on the assumption that they will eventually monetize these investments. They are betting that the AI applications they build will generate revenue sufficient to justify the capital expenditure. But valuations in the AI sector have climbed steeply, and much of the spending has been financed through debt. The concern is not whether data centers will be useful—they clearly will be—but whether the companies using them can turn those investments into profits quickly enough to service their obligations and reward their shareholders. That question remains unanswered, and it hangs over even the most bullish dealmaking.