In a move that influences the economics of artificial intelligence, a heavyweight consortium led by BlackRock agreed to acquire Aligned Data Centers for roughly $40 billion. Backed by Microsoft, NVIDIA, Abu Dhabi’s MGX, Elon Musk’s xAI and other investors, the deal ranks among the largest transactions in the data center sector.
Aligned, founded in 2013, operates a sprawling footprint: more than 50 campuses and over 5 gigawatts of operating and planned capacity. The buyer group, organized through BlackRock’s Global Infrastructure Partners and the AI Infrastructure Partnership, says it expects to close in the first half of 2026 pending approvals, including a likely review by CFIUS (Committee on Foreign Investment in the United States).
A Tight Market
Enormous demand for data center compute has driven this deal. Vacancy in North American data center hubs has slipped to historic lows (Northern Virginia fell below 1%), and roughly three-quarters of capacity under construction is reportedly pre-leased, mostly by cloud and AI providers, years before launch. That pre-emptive contracting creates a two-tier market: hyperscalers reserve future megawatts while enterprises scramble for the remainder.
The price backdrop reflects this scarcity. Colocation rates climbed to about $217 per kilowatt per month globally in early 2025, with major markets up high teens year over year. Private equity has been a major consolidator, accounting for the vast majority of data center M&A since 2022 as transaction values surged from $26 billion in 2023 to more than $70 billion in 2024. Concentrated ownership is expected to translate to pricing leverage.
Aligned is designed for the workloads supporting the rapid AI buildout cycle. Densities that once hovered around 40 kW per rack are moving past 100 kW and pointing to 250 kW as GPU clusters swell. The company’s modular designs, power roadmaps and patented cooling have made it a favored builder for rapid, high-density deployments. And with Microsoft and NVIDIA in the buyer mix, some of the investors are also end users.
The Aligned purchase also redefines what “infrastructure” means in the AI era. Investors increasingly view AI-ready campuses as utility-like assets: long-dated, contracted cash flows with upside from density upgrades (liquid cooling, optical interconnects) and multi-gigawatt expansion. Morgan Stanley pegs 2025 AI infrastructure spend at the high hundreds of billions. Supporting that analysis, OpenAI, Meta and others are investing aggressively to secure chips and megawatts at levels that have never been seen before. Against that backdrop, owning compute facilities becomes a strategic hedge.
New Challenges for AI Buyers
For enterprise managers, the news offers challenges. Data center capacity is increasingly allocated by long-term planning, not spot-market bids. Planning horizons stretch to three to five years. Analysts urge enterprise executives to expand beyond Tier-1 metros into secondary markets, negotiate rights of first offer/refusal on future build phases, and lock in SLAs tied to launch milestones.
Power has become the limiting factor, so strategies that secure substations, on-site generation, or preferential interconnects matter as much as square footage. There’s also now a greater reward for in-house compute planning: optimizing under-architected workloads can free meaningful compute capacity and delay expensive migrations.
For enterprise executives who need to support their AI deployments, the counter-strategy relies on major foresight and flexibility. The BlackRock deal means that it’s more important than ever to secure options early, diversify vendors and markets, and treat power supply like the first dependency in any AI roadmap.
Bottom line: the world’s AI software increasingly now runs on long-dated, less accessible contracts for data centers. If executives were on the fence about investing in AI, there’s no more time for hesitation—the available facilities are becoming far too scarce.
