AI's physical layer dominates 48-hour deal flow

The biggest transactions of the past 48 hours all ran through the physical layer of AI—power, data centers, and compute—carrying infrastructure-sized price tags while fund mergers and RIA roll-ups sat this one out. KKR's $2.2 billion deal with the IMM Investment-Stonebridge consortium, SK Telecom, and SK Horizon buys into SK Telecom's data-center assets and prices hyperscale compute real estate as core infrastructure; around it, OpenAI announced a deal with Georgia Power, Nvidia and Amazon Web Services announced a partnership, and EdgeConneX announced a transaction. The compute and power contracts that underwrite the AI build-out are now trading with blue-chip liquidity.

Brookfield announced a deal with Precinct Capital, EQT Real Estate closed with LBA Realty, and Potentia Energy completed a transaction, real-asset investors reaching for the same conclusion; the energy ledger stayed just as busy, with ACME Solar announcing a $164 million deal, Quaise closing $180 million, and Sowitec, Sembcorp, and Eurowind each announcing their own. None of those names made the RIA headlines, but together they describe a capital rotation that dwarfs the wealth management channel's M&A activity.

Data centers and the power that feeds them were once the province of specialist funds making opportunistic bets. The two-day stretch—a $2.2 billion data-center acquisition, a new power agreement for OpenAI, a compute partnership between Nvidia and AWS, and a wave of energy transactions—suggests they have been reclassified as core holdings, the construction of the next decade's asset class rather than a side bet.

Market structure is following the capital. Texas Stock Exchange and Brookmont announced a deal, the public-market plumbing being rewired for the same flows; an exchange announces a partnership when it expects listings, and commitments like the ones in this window tell it where to look.

For private wealth managers, the allocation message is already visible in interval funds, private credit vehicles, and co-investment sleeves. The pace of the past two days marks the AI trade's shift from venture story to real-asset story, with term sheets now written at infrastructure scale.