AI's $3tn Data Center Boom: Who's Funding the Future?
The global race to build out the foundational infrastructure for artificial intelligence has ignited an unprecedented construction boom, raising a crucial $3 trillion question: who will bear the colossal cost of powering the AI revolution? While tech giants initially spearheaded this massive undertaking, private capital is now aggressively joining the fray, eager to capture the rewards from this new digital gold rush.
The scale of investment is staggering. Global data center capital expenditure surged by an astonishing 51% to $455 billion in 2024, with projections indicating a further rise of over 30% in 2025, potentially nearing $600 billion. This exponential growth is driven primarily by the insatiable demands of generative AI models and the broader expansion of cloud computing. By 2025, an estimated 33% of global data center capacity will be dedicated solely to AI applications, a figure expected to skyrocket to 70% by 2030. Analysts predict approximately 10 gigawatts of new global data center capacity will break ground in 2025 alone, representing roughly $170 billion in asset value requiring immediate financing.
Unsurprisingly, Big Tech remains at the forefront of this buildout. Hyperscalers like Amazon Web Services (AWS), Google, Meta, and Microsoft continue to pour immense resources into AI infrastructure, collectively accounting for 44% of data center capital investments in the first quarter of 2025. Microsoft alone plans to invest $80 billion in AI data centers by 2025, while Amazon has allocated $86 billion to expand its AI infrastructure. Meta, too, is securing substantial financing, including a recent $29 billion package for a new data center project in Louisiana.
However, the sheer capital requirements and the perceived long-term stability of data center assets have attracted a diverse array of private capital. Private equity (PE) firms, infrastructure funds, and sovereign wealth funds (SWFs) are increasingly viewing data centers as “long-term safe havens” for investment, even amid economic uncertainties. In 2024, PE-backed data center mergers and acquisitions globally reached C$18.15 billion, marking the highest total in at least five years. Major private capital players are forging significant partnerships, such as KKR and Energy Capital Partners’ $50 billion AI infrastructure collaboration, and Blackstone’s extensive portfolio of investments in data center operators like QTS Realty Trust, Digital Realty, and AirTrunk, alongside its backing of specialized AI chip owners like CoreWeave. Global Infrastructure Partners, part of BlackRock, has also launched a $30 billion Global AI Infrastructure Partnership with tech giants Microsoft, Nvidia, and xAI, aiming to build data centers for all hyperscalers. Sovereign wealth funds from the UAE (Mubadala, ADIA, ADQ’s Zero Two) and Singapore (GIC), along with Canada Pension Plan Investments (CPP Investments), are actively deploying billions into data center projects across North America, Asia-Pacific, and Europe, underscoring the global nature of this investment wave.
Yet, this unprecedented expansion is not without significant hurdles. The most pressing challenge is the immense energy appetite of AI data centers. Generative AI models, in particular, consume exponentially more power for training and processing, with some estimates suggesting data centers could account for up to 21% of overall global energy demand by 2030. This surging demand is already straining existing power grids and contributing to higher electricity costs for consumers in various regions. Beyond power, managing the intense heat generated by high-density AI hardware is critical, making advanced liquid cooling solutions, which are 3,000 times more efficient than traditional air cooling, increasingly essential. Supply chain bottlenecks, particularly for key infrastructure components, and protracted power transmission delays, which can stretch up to four years in some areas, further complicate the rapid buildout. Addressing these challenges requires not just capital, but also innovative engineering solutions, from integrating renewable energy sources and exploring small modular nuclear reactors (SMRs) to developing new network architectures capable of handling the massive “east-west” data traffic within AI clusters.
Ultimately, the $3 trillion AI building boom signifies a fundamental reorientation of global capital towards the digital infrastructure that underpins the future of artificial intelligence. With Big Tech and a growing army of private investors betting big on this foundational layer, the question is no longer if the infrastructure will be built, but how these vast investments will overcome significant energy and logistical challenges to power the next wave of technological innovation.