


CEO @ Corley Energy | Building Behind-the-Meter Power for AI Data Centers | Host, Energizing AI Podcast


Texas just handed SpaceX $SPCX the air permit for its $24B AI data center at Starbase: 35 gas turbines, about 748MW, on 444 acres. The filing names every turbine: - 18 Solar SMT-130s from $CAT’s Solar Turbines, the same model running xAI’s Colossus in Memphis - 9 Mitsubishi FT8s - 8 $GEV TM2500s, GE’s trailer-mounted turbines SpaceX says all of them are “used and/or refurbished,” and it bought none directly from the manufacturer. New turbines are sold out for years, so the AI build is now buying used ones. They’re also temporary. The filing says they’ll be replaced by 8 GE 7E heavy-frame turbines, about 708MW, also used. GE Vernova gets paid on both rounds through parts and service. The detail that matters most is the emissions cap. The permit limits NOx, the main smog pollutant, to 248.86 tons a year. The federal line that triggers a full major-source review is 250 tons. SpaceX came in 1.14 tons under. That cap also limits how hard the plant can run. A local outlet that read the filing estimates full output would emit about 458 tons a year. Under the cap, the plant can average only around half its 748MW. The clock matters too. The TCEQ, the state agency that signs off on anything that burns fuel, says it has “resumed” permitting AI facilities after Governor Abbott paused them in September. Democratic Governor-elect Gina Hinojosa takes office around January 19 and promises a data center freeze. An issued permit is hard to undo, since the TCEQ’s three commissioners are Abbott appointees on staggered six-year terms. Every AI power project waiting at the TCEQ now has about nine weeks. ERCOT, the Texas grid, hasn’t approved a new large load to connect since August.

@bmbrigham your team hosted me and some others a little while back at your office. You’ve got a great team. Excited for where you guys are headed.


"data centers [are projected to] drive residential electricity rates in Entergy Mississippi’s service territory 16% lower in 2030 than they would have been without data center growth." mississippitoday.org/2026/11…

Islanded power checks all of these boxes. Since we don’t connect to the grid, we pay for our own power. We don’t use municipal water. We don’t build near communities. We’re happy to pay taxes. Islanded power is the gold standard both Abbott and Hinojosa want. Open TCEQ
The governor-elect of Texas, Gina Hinojosa, has explicitly said that all currently permitted datacenters in Texas will be allowed to finish construction. This essentially covers all of the 2027 and likely well into 2028. She is going to pass laws that ensure that all *newly* permitted datacenters going forward for delivery in 2028/2029 will pay taxes, commit to no increase in power prices for consumers, don’t impact water availability for people, be constructed with union labor and give “Texans a seat at the table.” Sounds reasonable. She doesn’t want to lose a lot of good blue collar jobs and property taxes for small towns in rural Texas. Datacenter commentary starts at 6:30 in this interview. Also attaching her official campaign position from her website in the next post. Sidenote: After watching a bunch of her speeches and interviews, I do think she is a political talent and probably has a future in national politics. youtu.be/D-T6H3b8ZeE?is=mEWu…


When I was in the Marines, I was working on multi million dollar comm systems that were inferior to my $500 iPhone (circa 2012). It never made sense to me why we weren’t using the latest and greatest tech to enable our warfighters. Love to see Anduril changing that 🇺🇸
BlueChip. Built in America. America made the first electronic computer to win a war. Now we depend on our greatest adversaries to build the computers that run our economy and defend our nation. Today, we bring it home.

Said no one ever.

Save money. Build a 60 year old car.
Save money. Buy an 8 year old car.

As someone with a portfolio of energized land assets here is what I am seeing: 1. Neoclouds are disorganized and poorly capitalized. Some are definitely better than others however. 2. Hyperscalers are intractable, bureaucratic organizations that take 3 months to do an NDA and want 18mo escrow periods. We built our portfolio by competing against these terms because of our speed and simplicity. The more future assets orgs like ours win, the bigger the problem for the hyperscalers downstream (see 3). 3. There are delays in the existing pipeline that will catch people off guard in 2027. Hyperacalers will (obviously) react by focusing on their own internal compute needs over any third party. If the third party is a model company, they will tell them to go through their model router where they will also take a non-trivial vig. Them telling Ant they have no compute for them (rumored in Ant’s S-1) if true is a huge tell that this is already happening. All of this is to say that it does look like there will be a lot of partially stranded compute and compute sitting in inventory in 2027 than is currently forecasted. This is bought and paid for but effectively underearning (below total cost of capital) or collecting dust. We see this as true for everyone - model companies, hyperscalers, neoclouds except ONE (you can guess which one). This dislocation will benefit two kinds of orgs: the non-bureaucratic big scalers who can build industrial scale quickly and the nimble, well capitalized players feeding them. 2027 will be a turbulent year because it will clear out a bunch of wrong/bad bets. It will also be a consolidation year for a few.
I think we’re going to see a lot of stranded compute in partially complete data centers

Joules of energy × proof of work = Bitcoin
Elon predicts money will soon become obsolete. The new money will be called JouleWork. My friend @BrianRoemmele explains JouleWork as a thermodynamic unit of account tied directly to physical reality—energy consumed and useful work output (e.g., Joules of energy × efficiency/normalization factors for AI/robot labor). The future currency will be rooted in physics.

Could not be more excited to announce Spark Liquid: @CrusoeAI’s newest modular datacenter product manufactured in our newest factory. - builds on the proven Spark platform, which is the only modular datacenter platform to achieve NVIDIA Exemplar reliability and performance standards - deployable in weeks not years - compact and low-profile - compatible with any energy source - massively scalable - zero operational water use - designed for next gen @nvidia chips and DSX-compatible
New: Crusoe Spark™ Liquid, our liquid-cooled modular AI factory built for @NVIDIA Vera Rubin NVL72 rack-scale systems. 5x the compute density of air-cooled Crusoe Spark, and part of the NVIDIA DSX POD design partner ecosystem. Plus: Managed RL and Managed MLflow in private preview ⤵ crusoe.ai/resources/blog/cru…


BREAKING: NVIDIA introduces forward-deployable, containerized modular AI data centers. The company announced DSX POD, a factory-manufactured modular data center design intended to reduce deployment times from six months to as little as 14 days at prepared sites. The systems feature pre-integrated computing, networking, cabling, and liquid cooling. Key partners include @armada_ai and @anduriltech, among others, with initial deliveries expected in early 2027.

Underappreciated how much the shale revolution unlocked the AI revolution.