Three power deals moved on November 17. One is a giant utility take-private at the finish line. One is a battery developer changing owners. One is a pre-revenue power company going public through a SPAC. They sit at very different points on the risk curve, and all three are chasing the same thing: electricity for a grid that data centers keep stretching.
AES received a final order from the Federal Energy Regulatory Commission approving its sale to the consortium led by Global Infrastructure Partners (part of BlackRock) and the EQT Infrastructure VI fund, with CalPERS and Qatar Investment Authority as co-underwriters. FERC was the last suspensory approval. Stockholders, CFIUS, the Public Utilities Commission of Ohio and foreign regulators had already signed off. Closing is expected on December 2, 2026, after which AES leaves the NYSE. Shareholders get $15.00 a share in cash in a deal valued at about $33.4 billion including debt. AES Indiana and AES Ohio stay under their existing regulators. The asset the buyers wanted is clear enough. AES has signed more than 11.8 GW of clean energy agreements with large technology companies, and it needed private capital to keep building without cutting its dividend or issuing a pile of new stock.
The same day, EQT Transition Infrastructure agreed to buy Flatiron Energy from Hull Street Energy. Flatiron, founded in 2021, develops and runs utility-scale battery storage in New England, New York and the Mid-Atlantic, in dense areas where capacity is tight. Its pipeline is 24 GWh, almost ten times what it was when Hull Street invested in 2022. Of that, 6 GWh has long-term contracts and 2 GWh is under construction, including a project the company calls the largest urban storage installation in the world. The release puts storage in those markets at 2 GWh in 2026 and 108 GWh by 2035. No price was given. Closing is expected this quarter. What EQT is paying for is site control. Grid-connected land in New York City or northern New Jersey is scarce, and nobody can manufacture more of it.
AMPERA is the speculative end. It agreed to merge with Berto Acquisition Corp. (Nasdaq: TACO) at a pre-money equity value of about $900 million, and would trade as AMPR. The near-term product is a 7.5 MW natural-gas system built around a supercritical CO2 power cycle, with commercial production targeted for the fourth quarter of 2027. The long-term plan is a 15 MW thorium-fueled subcritical reactor. Committed money so far is just over $40 million, including a PIPE of roughly $30 million. Berto’s trust holds about $317 million, but that only matters if shareholders don’t redeem. The S-4 hasn’t been filed, the deal is meant to close in the first quarter of 2027 and Berto’s deadline is May 1, 2027.
So the capital stack for AI power now runs from a $33 billion utility with contracted customers, through a storage developer with scarce sites, to a company that hasn’t shipped a unit. Watch AMPERA’s redemption rate. It will show how far down that stack public investors are willing to go.