Peter Smith
The AI revolution isn't just about software—it's about power, real estate, and hardware. A buyer group led by Nvidia, Microsoft, BlackRock, and xAI is moving to acquire Aligned Data Centers for around $40B, setting the stage to build and lease next-generation AI facilities at massive scale. If it goes through, adding ~20 GW of capacity annually would fundamentally reshape the global data-center landscape, from GPUs and switches to transformers and cooling systems.
Inside the $40B Aligned Deal
As Shay Boloor pointed out, the consortium's bid for Aligned gives them instant access to large, power-ready sites and a development pipeline across North America and Latin America. Here's the strategic play:
- Nvidia locks in capacity for its Hopper and Blackwell GPUs, NVLink, and networking gear
- Microsoft gets a partner-built route to scale AI workloads beyond Azure's own infrastructure, solving deployment bottlenecks
- BlackRock funnels long-term capital into hard assets with strong AI-driven fundamentals
- xAI secures compute at scale for rapid model training and inference
It's a self-reinforcing loop: hyperscaler demand attracts capital, which funds more capacity, which absorbs more Nvidia chips—keeping the AI buildout humming.
Why This Matters
The deal signals two big shifts. First, AI infrastructure has gone mainstream—data centers aren't just a niche real estate play anymore; they're critical national compute assets. Second, a $40B transaction as step one of a $100B program means we're looking at a multi-year build cycle, not a one-off acquisition. For investors, that's a clear demand signal for everything that powers an AI campus: chips, optics, networking, switchgear, generators, cooling systems, and a lot of electricity.
Why Now?
Several forces are converging. Training cutting-edge models requires thousands of GPUs clustered together with ultra-low-latency connections, driving campus-scale builds rather than incremental expansions. Many regions lack the grid capacity to support this, so owning shovel-ready land with transmission access becomes a major advantage. Long-term hyperscaler leases support project finance economics—ideal for infrastructure funds hunting yield with growth. And buying an operator with existing sites, teams, and permits in place speeds up delivery versus starting from scratch.
Who Benefits
Nvidia remains the performance anchor—more campuses mean more high-end systems shipped. Microsoft can offload or complement Azure builds, accelerating capacity for Copilot and OpenAI workloads. BlackRock gains exposure to a secular growth asset class with inflation-linked rents. Then there's the wider ecosystem: networking vendors (optics, switches), power equipment makers (transformers, switchgear, backup generation), thermal solutions (liquid cooling, heat rejection), and construction firms all stand to see multi-year order backlogs.
Risks to Watch
Execution won't be simple. Interconnection queues, substation lead times, and local approvals can delay projects. Supply chains for long-lead items like large transformers, chillers, and high-density racks remain tight. Higher interest rates challenge returns if lease pricing doesn't keep up. And rapid GPU and networking upgrades require flexible designs to avoid stranded assets.
Peter Smith
