Joel Litman calls it Dark Energy, and he says this new energy technology will create a $10 trillion boom. SpaceX is deploying it right now in Tennessee and Mississippi. The technology is already live, with satellite imagery confirming the installations.
The first question anyone should ask: is $10 trillion even a real number?
Let’s put it in context.
The global energy market is roughly $8 trillion a year. The entire oil and gas value chain — from exploration to refining to the gas pump — sits somewhere between $10 and $15 trillion depending on the year and the price of crude. Apple, Amazon, and Microsoft combined trade for about $9 trillion on a good day.
So when Litman says $10 trillion, he is talking about replacing oil and a shift big enough to create new industry giants from scratch.
That sounds ridiculous until you look at what is actually happening.
What Is Dark Energy SpaceX?
Dark Energy SpaceX refers to the natural gas turbine power generation that SpaceX and xAI are deploying behind the meter for AI data centers. Joel Litman coined the term to describe the energy foundation being built right now in Tennessee and Mississippi, where aeroderivative gas turbines mounted on flatbed trailers bypass the multi-year grid interconnection queue. The technology is already live, with satellite imagery confirming turbine installations at the Colossus data center cluster.
The SpaceX Turbine Data
SpaceX’s IPO filing — page 224, in the fine print — discloses more than $2.8 billion in natural gas turbine purchases. The company committed to spending $2 billion on mobile gas turbines and related equipment in a single April purchase agreement, plus another $925 million through 2029. This is signed paperwork filed with the SEC.
xAI is operating roughly 46 natural gas turbines at its Colossus data center cluster in Memphis and Southaven, Mississippi. The state of Mississippi approved a permit for 41 turbines generating 1.2 gigawatts. That is more than double what Amazon and Microsoft’s biggest individual power projects deliver. The state waived taxes on the sites. The DOJ intervened to block a lawsuit against xAI’s turbines, arguing the power is a national security priority.
This is Dark Energy in the flesh.
Litman calls it Dark Energy, and despite the science-fiction name, the reality is concrete. These are aeroderivative gas turbines — think jet engines bolted to generators, mounted on flatbed trailers, running on natural gas. They hit 99.7% uptime, come online in minutes rather than years, and bypass the five-to-ten-year utility connection queue entirely by operating off-grid.
The technology is natural gas, deployed with military-grade urgency and hyperscaler budgets.
The key claim in Litman’s thesis is not about the turbines themselves. The global gas turbine market is $20-30 billion today. Growing at 5-7% a year, maybe $25 billion by 2030. That alone does not get you to $10 trillion.
Who Is Joel Litman?
Joel Litman is a macroeconomic analyst and the founder of Altimetry, a research firm that uses Uniform Accounting to identify mispriced stocks. He publishes the Altimetry hidden alpha newsletter and has built his career on the thesis that institutional-grade financial data, when properly adjusted, reveals opportunities the market misses. The Dark Energy SpaceX thesis extends his methodology to the energy infrastructure layer of the AI buildout.
The $10 Trillion Claim
The $10 trillion claim is about what the turbines enable.
The AI Power Demand
Here is the real math. AI data centers consumed roughly 415 terawatt-hours of electricity in 2024. The IEA expects that to nearly double to 945 TWh by 2030. Gartner says global data center power demand will hit 132 gigawatts in 2026, up 27% in a single year. AI-optimized servers will account for 31% of that power — up from 20% last year. By 2027, AI servers will consume more electricity than every other server on earth combined.
IDC forecasts AI infrastructure spending at $487 billion in 2026, crossing $1 trillion by 2029.
The Grid Bottleneck
The grid cannot handle it. Utilities in Northern Virginia, the Dallas-Fort Worth metroplex, and Dublin are already rationing power or calling moratoriums on new data center connections until 2030 or beyond. The queue for a new grid interconnection runs five to ten years in many regions. AI labs cannot wait that long for grid interconnection, so they are building their own power — the setup Litman identified early. The hyperscalers — SpaceX, Microsoft, Google, Amazon, Oracle, Meta — are all pursuing behind-the-meter natural gas generation because it is the only option that works right now. Small modular reactors are years away and fusion is decades away. Renewables cannot provide the 24/7 baseload these facilities need.
Sam Altman called Boom Supersonic about turbines for AI — the Boom Supersonic turbine call covers that angle in detail. Jensen Huang, Mark Zuckerberg, and Jeff Bezos are all backing the same trend. The Wall Street Journal described it as “reshaping American power.”
Is the $10 Trillion Boom Real?
The dark energy spacex thesis rests on real, documented infrastructure: SpaceX filed $2.8 billion in turbine purchases with the SEC, Mississippi approved 41 turbines generating 1.2 gigawatts, and AI data center power demand is forecast to nearly double by 2030. The $10 trillion figure is a prediction about the total economic value created when AI’s power bottleneck is removed, not a gas turbine market cap prediction. Whether that value materializes at that scale is the part the market has not answered yet.
The Oil Parallel
So what does a $10 trillion boom actually look like?
It looks like what happened to oil in the 20th century. Oil was a niche product in 1900. By 1950, it powered the global economy. The companies that controlled its production, refining, and distribution became the most valuable enterprises in the world. The whole value chain — upstream, midstream, downstream — was worth tens of trillions.
Dark Energy follows the same pattern. The turbines themselves are the production layer. The natural gas supply is the feedstock. The grid interconnection and transmission infrastructure is the midstream. And the AI compute that runs on top is the final product.
If Dark Energy unlocks the next generation of AI capability — the models that would otherwise stall from lack of electricity — the productivity gains alone could run into the trillions. The infrastructure build-out to support it could rival the interstate highway system or the electrification of rural America.
Litman is saying the energy foundation of the AI revolution is being built right now, in remote fields in Tennessee and Mississippi, with equipment sitting on trailers, while most of Wall Street has yet to notice.
Dark Energy spacex is a keyword gaining traction as investors catch on to what is sitting on those flatbed trailers in the South.
The turbines are bought. The contracts are signed. The power is flowing. The question is how big the boom gets.
And $10 trillion may prove conservative.