The first useful thing to know about Chase Lochmiller is that he once expected to spend his life expanding humanity's knowledge of the universe. The second is that he got impatient.
At MIT, where he studied mathematics and physics, Lochmiller pictured a future in theoretical physics. He did research there and at Los Alamos National Laboratory. On his first day at the lab, a principal investigator offered the durable joke of fusion science: practical fusion was 30 years away when the investigator began, and it remained 30 years away three decades later. Lochmiller had energy to spend and little appetite for waiting on the cosmic calendar.
He moved into quantitative finance, a world in which milliseconds had consequences. At GETCO and Jump Trading, he worked on algorithmic strategies. He later earned a Stanford master's in computer science, specializing in artificial intelligence, and joined the crypto investment firm Polychain Capital. Physics had taught him to respect first principles; trading taught him to respect price, timing and risk. Crusoe would eventually require all four.
Base camp was in Denver
Lochmiller grew up in Denver, competing in math contests and playing soccer. The 1994 World Cup helped turn the sport into a fascination. At MIT he became a center midfielder, a position that rewards a taste for connecting moving pieces. Nitin Perumbeti, now Crusoe's chief technology officer, played sweeper on the same team. Lochmiller jokes that he was the team's CEO. The line is convenient, but the older relationships are more telling: people from several chapters of his life later joined the company.
His co-founder, Cully Cavness, goes back further. The two attended Kent Denver School, then separated into complementary careers. Lochmiller learned compute and markets. Cavness studied geology, traveled through global energy systems and worked in oil and gas, where he encountered the strange economics of flaring. An oil well can produce natural gas as a byproduct, but if no pipeline or nearby customer can take it, the gas becomes a nuisance. Operators burn it at the site. A valuable fuel turns into flame because geography has vetoed the transaction.
The pair also shared an enthusiasm for mountaineering. Lochmiller had been drawn to high peaks since middle school, when a teacher's husband, the blind climber Erik Weihenmayer, was preparing for Everest. After college, Lochmiller pursued expeditions across the world. He was on Everest in 2014, the season an avalanche in the Khumbu Icefall killed 16 Sherpas and ended the climbing season. He returned in 2018 and reached the summit. By then he had climbed five of the Seven Summits.
“Getting up is optional, getting down is mandatory.”Chase Lochmiller, on mountaineering and risk
Crusoe's founders were on a mountain when they were sketching the company's origins. Their shared hobby supplied one of its values: “Think like a mountaineer.” Lochmiller's explanation is practical rather than decorative. Prepare hard. Carry what is necessary. Anticipate failed gear, sick partners and changing weather. Have a plan B and a plan C. Move quickly when speed reduces exposure, but remember that a summit without a safe descent is a botched expedition.
The stranded molecule
Founded in 2018, Crusoe began with an inversion. Instead of building a data center in a conventional market and dragging electricity toward it, the company would take modular computing equipment to the energy. A generator could turn stranded gas into power beside an oil well. The computers could turn that power into a product. Early on, that product was bitcoin, a workload that could travel to a remote site and tolerate an unusual operating environment.
Move the computing load to the energy source, instead of waiting for the energy to reach a conventional data center.
The proposition joined two industries that rarely shared a sentence. Cavness brought the oilfield experience and relationships. Lochmiller brought computing, AI and blockchain knowledge. The equipment was unglamorous: generators, switchgear, transformers, shipping-container-like modules. The elegance was in the coordination. Waste existed because energy and demand occupied different places. Software could make demand portable.
Even in a 2020 interview about bitcoin mining, Lochmiller was already discussing machine learning and protein folding as future consumers of inexpensive computing. His long-term question was broader than cryptocurrency: what might society do if computation cost less? Bitcoin was the first customer that fit the peculiar supply. It was not the final destination.
The pivot that wasn't
When generative AI demand accelerated after the 2022 release of ChatGPT, the shortage changed scale. The industry needed clusters of advanced GPUs, high-performance networks, liquid cooling and vast amounts of reliable electricity. Crusoe moved up the stack into cloud services and outward into hyperscale data centers. In 2025 it agreed to sell its bitcoin mining operation, making explicit what the business had become.
Calling this a pivot misses the stubborn continuity. The original business asked where cheap, underused energy could meet flexible compute. The AI business asks the same question with far more zeros. In Abilene, Texas, Crusoe found a grid region rich in wind and solar generation where power prices had at times gone negative because supply outran local demand. The company began building the first phase of the Stargate data center project there.
By September 2026, Crusoe said it had more than six gigawatts of gross contracted capacity, with more than one gigawatt delivered and operating. It announced the initial closing of an anticipated $3.9 billion Series F at a $30.9 billion post-money valuation. The company described more than $140 billion in total contracted value across its platform. Those are company figures, but their order of magnitude makes the underlying point: AI infrastructure is now a capital and energy business wearing a software industry's name tag.
From electrons to tokens
Lochmiller likes the term “AI factory.” It restores the machinery to a product that arrives on a screen looking weightless. Crusoe now works across land and power development, data center design, construction, cooling, networking, GPU clusters and managed software services. Its vertical-integration thesis is that these layers must be planned together when customers want hundreds of megawatts quickly.
The approach also widens the hunt for energy. Crusoe has worked with wind and solar resources, second-life electric-vehicle batteries, gas generation and proposed nuclear projects. Lochmiller argues that enormous, dependable AI demand can help finance energy technologies that struggled to find a sufficiently committed buyer. The environmental arithmetic is not automatic. Power sources differ, and fast construction can collide with climate goals. His wager is that control of more of the stack makes those trade-offs visible enough to engineer.
There is a family irony here. Lochmiller's father built a career as a commercial and multifamily real-estate owner and operator. The son spent years making his own path through physics, algorithms and crypto. Now he develops enormous buildings, secures tenants and worries about power, land and long-duration contracts. After all that software, he jokes, he became a landlord too.
Recognition followed the scale. Lochmiller joined the Aspen Institute's 2025 class of Henry Crown Fellows and appeared on the 2025 TIME100 AI list. Yet his public manner remains that of an engineer-operator: animated by systems, fond of physical analogies and willing to discuss downside. Asked about an AI investment bubble, he did not wave the risk away. He talked about leverage, counterparties and underwriting bad scenarios. Plan B and Plan C had followed him down the mountain.
The descent matters
Lochmiller's ambition is hardly modest. He wants Crusoe to help “orchestrate intelligence for the economy,” providing the energy, buildings and computing systems on which useful AI runs. In a 2026 conversation, he entertained a path to a trillion-dollar company. The claim sounds less peculiar when set beside the two markets he is trying to join: global energy and global computing.
The more revealing aspiration is abundance. He speaks of making both energy and intelligence more available, and of AI expanding scientific discovery and economic productivity. It is the old theoretical physicist's desire to enlarge human knowledge, returned in industrial clothing. The timetable is faster now. The laboratory covers acres.
There are still two summits on Lochmiller's Seven Summits list. He says they have been deprioritized. Crusoe supplies enough altitude for the moment. The company's route from a gas flare to gigawatt campuses is a lesson in looking sideways at scarcity. Sometimes the useful resource is already present; it is simply marooned from demand.
AI is often narrated as a parade of models, benchmarks and clever code. Lochmiller's career makes a less ethereal argument. Intelligence needs turbines, substations, pipes, batteries, cooling loops, contracts and people in hard hats. The future may arrive as tokens. First, somebody has to wire the place.