The biggest infrastructure deal in digital history just closed. A consortium led by BlackRock’s Global Infrastructure Partners, MGX of Abu Dhabi, and the AI Infrastructure Partnership acquired Aligned Data Centers for $40 billion — more than double the previous record for a data center acquisition. The deal reshapes who controls the physical layer of AI and introduces a new class of power broker into the AI value chain.

The Consortium: A New Kind of AI Power Player

The buyer group reads like a who’s who of global capital and AI compute: Nvidia, Microsoft, BlackRock, MGX, Elon Musk’s xAI, and the Kuwait Investment Authority. This is not a typical private equity buyout. It is a strategic infrastructure play where the largest AI consumers are securing their own supply chain.

Aligned Data Centers controls more than 5 gigawatts of data center capacity across 50 campuses in the US and Latin America. At current AI training power requirements, that capacity can support dozens of frontier model training runs simultaneously. The consortium is not buying a real estate portfolio — it is buying the physical bottleneck of AI progress.

Why BlackRock Is the Real Story

BlackRock managing $11.5 trillion in assets entering the AI infrastructure layer signals a fundamental shift. Until now, AI infrastructure has been funded primarily by hyperscaler capital expenditure — Microsoft’s $190 billion, Google’s $75 billion, Meta’s $65 billion annual commitments. BlackRock’s entry means institutional capital is now competing with hyperscaler balance sheets for control of AI’s physical layer.

The implications are significant:

• Capital advantage: BlackRock can deploy patient, long-duration capital that hyperscalers — under quarterly earnings pressure — cannot always match.
• Neutrality premium: A BlackRock-owned data center can serve any AI lab. A Microsoft-owned facility primarily serves Azure. Neutrality commands pricing power.
• Asset class creation: BlackRock is effectively creating AI infrastructure as a distinct asset class, separate from traditional real estate or utilities. This opens the door for pension funds, sovereign wealth funds, and insurance companies to fund AI buildout.

The Hyperscaler Dilemma

Microsoft, Google, and Meta have collectively committed $690 billion in AI capital expenditure for 2026. Yet they face a structural problem: building data centers from scratch takes 18 to 36 months. The AI race does not wait. Acquiring existing capacity — even at premium valuations — is faster than building new capacity.

This creates a paradox. Microsoft is both a consortium member in the Aligned deal and a competitor building its own capacity. The company is hedging: securing third-party capacity for immediate needs while constructing proprietary facilities for long-term advantage. It is an expensive strategy, but the alternative — running out of compute during a competitive sprint — is worse.

The Vertical Integration Trend

The Aligned acquisition follows Alphabet’s purchase of clean-power developer Intersect Power earlier this year. Hyperscalers are no longer content to buy compute and electricity from third parties. They are vertically integrating into power generation, data center ownership, and chip design.

This vertical integration creates a new competitive moat: the companies that control power, cooling, and physical compute will have structural cost advantages over those renting it. The $40 billion Aligned deal is the clearest signal yet that AI’s competitive advantage is moving from algorithms to atoms.

What This Means for AI Startups

For AI startups, the infrastructure consolidation is a double-edged sword. On one hand, more data center capacity means more available compute. On the other hand, the consortium members — Nvidia, Microsoft, xAI — are also competitors. Startups training frontier models will increasingly depend on infrastructure owned by their rivals.

The startup that cannot secure its own compute capacity will not survive the next phase of AI competition. Anthropic’s deep Amazon partnership and xAI’s Memphis supercluster are not vanity projects — they are survival strategies.

The Infrastructure Thesis

The $40 billion Aligned deal crystallizes a thesis that has been building for two years: AI’s real bottleneck is not intelligence — it is infrastructure. The companies and investors that control data centers, power supply, and chip manufacturing will extract more long-term value than the companies building models on top of that infrastructure. BlackRock just placed the biggest bet in history on that thesis.

Read the full deep dive on Business Engineer →

Free strategic tools: Map of AI | Try the Framework | Defensibility Audit

Subscribe free → businessengineer.ai