A New Kind of Infrastructure Deal

It wasn’t the number that caught my eye first—it was who was behind it. When Nvidia and BlackRock came together in October 2025 to acquire Aligned Data Centres for $40 billion, the names said more than the price tag.

Nvidia, known globally for powering the artificial intelligence revolution with its GPUs, isn’t in the data centre business by accident. And BlackRock, with its long view on global asset management, doesn’t chase trends. Together, they’ve built a coalition that’s looking decades ahead, into the real architecture of intelligence.

This isn’t just a corporate acquisition. It’s the material foundation of AI — a move into land, energy, fibre, and silicon.

The Anatomy of the $40 Billion Deal

On 15 October 2025, a consortium led by BlackRock and joined by Nvidia, Microsoft, Abu Dhabi’s MGX, and Global Infrastructure Partners (GIP) announced the acquisition of Aligned Data Centres.

The reported value? $40 billion.

The scale? Aligned Data Centres controls more than 50 campus-style data centre facilities across the United States and Latin America. These facilities offer more than 2.5 gigawatts of existing or developable capacity, making it one of the largest privately held data centre platforms globally.

What does this say about the future? Simply that data — and the AI systems that rely on it — will need dedicated infrastructure at a pace and size never before seen.

Building the “AI Factory”

Jensen Huang, CEO of Nvidia, called these facilities “AI factories”. These are not traditional server farms. They’re power-intensive, heat-sensitive, custom-built environments for training large language models and executing real-time AI workloads.

These AI factories are essential to transforming raw data into computation. They are not optional. Without them, models stall, inference fails, and global AI ambitions slow down.

These facilities sit at the intersection of several critical industries: real estate, energy, computing, and cloud software. And the consortium backing this acquisition understands this convergence intimately.

The AI Infrastructure Partnership (AIP)

The acquisition wasn’t done by BlackRock and Nvidia alone. It’s part of a broader vehicle: the AI Infrastructure Partnership (AIP). This is a cross-sectoral alliance that combines the capital discipline of asset management with the technical urgency of AI development.

The AIP is designed to function as an operating partnership rather than a typical private equity fund. Its stated ambition is to mobilise up to $100 billion in investment—starting with $30 billion in equity—aimed at securing the global AI infrastructure pipeline.

This model combines deep pools of institutional capital with operational oversight, signalling that AI infrastructure is no longer just a technology issue. It’s a capital markets priority.

Why the Data Centre is the New Asset Class

What do investors see in a data centre? For one, stable long-term returns. For another, a critical role in one of the fastest-growing industries worldwide.

With demand for compute infrastructure expected to grow over 20% annually through 2030, according to McKinsey estimates, infrastructure funds are re-evaluating how they view energy, cooling, and location. Data centres are no longer seen as risk-heavy tech investments — they are utilities.

Aligned Data Centres’ platform is especially attractive because of its scalability and energy alignment. The company has developed a track record of securing power supply agreements and grid partnerships — key in a world where AI infrastructure is heavily dependent on electricity and sustainability commitments.

A Global Game, Not a Regional One

While the deal centres on assets in North and South America, the implications are global.

AI workloads are not restricted by geography. The datasets may be trained in one region and deployed in another. That makes infrastructure portability — or at least reliability — essential. Sovereign wealth funds, national regulators, and international cloud providers are watching this transaction closely.

It is quite probable for European countries, especially those looking to link energy goals to digital expansion, to bookmark this model. Likewise, the Asia-Pacific governments interested in grounding artificial intelligence ecosystems locally might want to consider the cross-industry approach behind this agreement.

Compute is the New Commodity

The growing commodification of compute—as essential as oil, water, or steel—reframes the Nvidia-BlackRock deal in more than financial terms.

By moving upstream into data centre ownership, Nvidia gains leverage in controlling not just chip supply but deployment environments. This ensures its GPUs are not just sold but used optimally — in infrastructure, it can help configure.

This gives Nvidia end-to-end insight into the AI stack: from silicon to software deployment. BlackRock, meanwhile, diversifies its exposure into a physical asset tied directly to the AI wave, offering its institutional clients a chance to participate without having to pick winning applications or algorithms.

What Does Microsoft Get Out of It?

Microsoft’s presence in the consortium may seem redundant — after all, it has its own global data centre network. But this partnership allows Microsoft to ensure continuity and redundancy in its infrastructure layer, especially as demand from Azure OpenAI workloads grows.

It also offers operational access to facilities not originally under its control. This can improve latency, workload distribution, and expansion timelines — key to competing against AWS and Google Cloud.

Will Others Follow?

The likely answer is yes. If this model works—tech firms investing directly into infrastructure, asset managers providing scale capital, and operators managing long-hold data centres—we’ll see replications.

Sovereign wealth funds like Singapore’s GIC or Canada’s CPP may back similar partnerships. Infrastructure developers in India, the UAE, and Africa may look to replicate the vertically integrated approach. AI infrastructure is now a globally investable category.

Why You Should Care

This deal changes the equation for any business building on AI. Whether you’re a founder, CIO, or investor, understanding who controls your compute supply is now critical.

In the same way we once mapped oil pipelines or semiconductor supply chains, today we need to trace where models are trained, where inference happens, and who owns that physical terrain.

Ask yourself:

• Is your AI supplier vertically integrated?
• Are your computing costs fixed or floating?
• Do you have visibility into the environmental and energy strategies of your infrastructure provider?

The Energy Reality

AI’s hunger for power is no secret. A single training run for a large language model can consume hundreds of megawatt hours of electricity. Data centres require cooling, redundancy, and regulatory compliance. That means they often need to be built close to power sources — whether hydro, nuclear, or solar.

Dark data centres emphasise surplus power availability more than most providers in history. They never embark on site selection without firm power commitments. They proactively engage local government entities and utilities in power planning.

Hence, the company itself, through the Nvidia Consortium perspective, and more widely from the BlackRock Consortium, finds itself strategically placed more in the context of energy trading than technology.

What’s Next?

If AI continues to accelerate — and there’s every reason to believe it will — then infrastructure will follow. We may see the emergence of AI-specific zones, preferential power rates for AI processing, and even infrastructure nationalism as countries seek to retain data and compute sovereignty.

The Nvidia-BlackRock deal doesn’t just anticipate this future. It lays the groundwork for it.

For global decision-makers, the question isn’t whether to watch this trend. It’s how to participate in it.