Two deals. One day. One acquirer.
On March 13, Mogin Law LLP’s coverage of AI deal activity reported that BlackRock has agreed to lead the acquisition of Aligned Data Centers in a transaction reportedly valued at approximately $40 billion. Nvidia and Microsoft are reported to be participating in the acquiring consortium. Aligned Data Centers reportedly operates more than 50 campuses across the Americas with over five gigawatts of capacity. The same day, Fladgate’s March 2026 AI roundup and the same Mogin Law LLP source reported that BlackRock’s Global Infrastructure Partners and EQT have agreed to acquire AES Corp, a major U.S. power utility, in a transaction reportedly valued at approximately $33.4 billion including debt. The combined reported figure across both transactions is approximately $73 billion, a number that, if confirmed, would represent one of the largest single-day infrastructure capital deployments in recent history.
Source note: Both transactions are sourced to T3 law firm commentary blogs. Article text was not fully readable during verification, and independent cross-reference searches were unavailable in this cycle. All figures and deal terms are reported, not confirmed from primary financial sources. TJS recommends treating deal terms as reported pending T1/T2 confirmation.
The full-stack thesis
AI infrastructure is not a single layer. It’s a stack. At the top sits the model layer, the foundation models and the inference compute that serves them. Below that sits the data center layer: the physical campuses, the cooling systems, the networking, the power distribution infrastructure that keeps the GPUs running. And beneath that, the energy layer: the electricity generation capacity that feeds the data centers in the first place.
The prevailing narrative in AI infrastructure investment has focused on the top of that stack. GPU allocation, cloud capacity, inference optimization. The reported BlackRock deals are a statement about the bottom two layers. A data center network acquisition and a power utility acquisition announced on the same day, by the same institution, in the same week, aren’t independent decisions. They’re a coordinated position on the full stack, a thesis that says the durable value in AI infrastructure isn’t in any single layer, it’s in owning the relationship between the layers.
Why the energy layer matters now
The electricity demand story behind AI is well established in market reporting. Data centers are power-intensive at baseline. AI inference workloads are substantially more so. The largest hyperscale data center buildouts currently underway are constrained not by capital or by land, but by grid connection timelines and electricity supply commitments.
Analysts and reports have linked the AES Corp acquisition to this dynamic, though that link is an inference from market context, not a confirmed rationale from the transaction parties. The inference is sound: a firm acquiring a major data center network would logically benefit from owning reliable electricity generation, particularly as grid connection queues lengthen and power purchase agreement costs rise. Whether or not that’s the explicit reasoning behind the AES deal, the structural logic holds.
What the Nvidia and Microsoft participation signals
The reported involvement of Nvidia and Microsoft in the Aligned Data Centers acquisition is the detail that sharpens the thesis. These are not passive financial investors. Nvidia’s core business is the GPU infrastructure that runs inside data centers. Microsoft’s Azure cloud depends on data center capacity as a core operational input. If either entity is in fact participating in this acquisition, and the “reportedly” qualifier matters here, as that participation couldn’t be confirmed from primary sources, the structure of the deal is not just asset management capital seeking yield. It’s vertical integration.
A data center operator in which the GPU manufacturer and a major cloud provider hold equity interests is a different competitive entity than an independent data center operator. It raises questions about capacity allocation, pricing, and access for competitors that are worth tracking as the transaction progresses toward its reported H1 2026 target close.
Implications for practitioners
Three groups should be watching this pattern closely.
Enterprise AI buyers evaluating platform stability need to understand that the compute infrastructure underpinning the AI services they rely on is in active transition. Ownership of data center capacity is concentrating. The implications for pricing, availability, and access terms over a multi-year horizon aren’t knowable today, but the direction of travel is visible.
AI infrastructure investors have a new comps dataset, if the deals confirm. Two reported transactions above $30 billion in a single day establish a pricing floor for institutional interest in AI infrastructure assets. Asset managers evaluating infrastructure holdings now have clearer evidence of what the largest pools of capital are willing to pay.
Policy and regulatory observers should note that these transactions, if confirmed, would represent a meaningful increase in concentration of AI infrastructure ownership. A small number of institutional investors and technology companies owning both the compute layer and the energy layer of AI infrastructure creates systemic dependency questions that regulators in multiple jurisdictions are likely to examine, regardless of whether the current regulatory frameworks are designed to address them.
What to watch
The H1 2026 target close for Aligned Data Centers and the late 2026 / early 2027 target for AES Corp are the first tracking milestones. Confirmation from T1/T2 financial sources will be the immediate priority for this coverage. Beyond that: watch for any regulatory review filings in jurisdictions where data center or utility acquisitions trigger antitrust scrutiny, and watch for any disclosed information about how capacity will be allocated post-close. The thesis is clear. The execution details will determine how it plays out.