Crusoe CEO Chase Lochmiller: Turning Stranded Energy Into Intelligence

1. Key Themes

Vertical Integration as the Strategy, "Electron to Token"

Crusoe's bet is that owning every layer of the AI infrastructure stack (power, data centers, manufacturing, and cloud) is more valuable than owning any single layer. Lochmiller describes the scope: "We have a complex business cuz it's vertically integrated, it spans like a huge scope of different work streams. Everything from energy generation and bringing electrons online, building power plants, working with independent power producers and utilities, to building data centers... and then building and operating a managed AI cloud platform with managed AI services." 00:00:00 He also notes the stack extends into hardware: "we have factories where we're manufacturing critical electrical equipment for data centers, and then we have AI researchers that are working on the best, most effective ways to serve and host AI models and manage the KV cache." 00:05:23 The cross-pollination is a feature: "there have been critical innovations that people have gotten from... electricians that are... working on wiring things up to... the researchers... how we think about the whole building management system, how we think about sort of the operating system from the electron to the token, I think, is something that really sets us apart." 00:43:17

The Oil & Gas Analogy: Building an "AI Supermajor"

Lochmiller maps AI infrastructure onto the upstream/midstream/downstream structure of oil and gas. "I view it as we're trying to build an AI supermajor in the same way that Exxon and Chevron are supermajors in the oil and gas industry. And with that, you sort of have this embedded risk management where... Exxon very famously doesn't hedge their oil exposure because when oil prices come down, their upstream business at the wellhead... those margins actually get compressed, but their downstream business at the refinery, those margins actually expand because the raw input costs actually go down." 00:17:02 The implication is that a vertically integrated player has natural hedges across the stack, and that competitors in one layer can be customers or partners in another.

Energy-First, Energy-Agnostic Infrastructure

Crusoe's founding thesis was "bring the market to the energy" rather than the reverse. "The original thesis for the business was like, go find stranded energy that is otherwise being underutilized and bring a market to that stranded energy with compute, instead of trying to bring that stranded energy to a market where it can be bought and sold, bring the market to the energy." 00:25:13 Today the portfolio is deliberately diversified: "We've done stuff with behind-the-meter wind with behind-the-meter gas, with behind-the-meter solar, with batteries... We've gone to Iceland... Norway is another great market for us." 00:20:48 And the demand picture: "The demand from intelligence for energy is voracious." 00:20:48

Focus Through Divestiture: Selling the Bitcoin Business

The Bitcoin business was profitable, but Lochmiller sold it to NYDIG to focus. "What drove me to divest that business was really focus... we were having to balance this tension between do we allocate capital and resources and attention towards our Bitcoin business or against this AI infrastructure business? And we were also getting feedback from investors where... 'my firm, we're actually blocked from making any investments that are tied to crypto.'" 00:30:39 The decision removed a capital-allocation distraction and an investor-universe constraint at once.

Scaling the CEO: Domain Experts, Accountability, and Repetition

Lochmiller's scalability comes from domain-expert leaders paired with accountability rather than blind delegation. "I have been able to find domain experts in each of these key verticals from energy, data centers, manufacturing, and AI cloud computing... I trust them a lot to be making the right decisions." 00:05:23 But trust is conditional: "It's not that I'm like... hands off... I want to understand what we're doing in this vertical, we have alignment on a direction we're headed. I get how it influences the other pieces of the business. And then I hold my team accountable to go execute on what we agreed upon." 00:08:00 On communication at scale: "As soon as I'm sort of getting sick of hearing myself say something, then we're just scratching the surface." 00:10:19

Speed and Urgency at 2,000+ Employees

Lochmiller says the company is moving faster, not slower, as it grows: "It feels faster and crazier than it's ever been." 00:11:50 He cites Elon Musk's technique of holding two tensions at once, a belief you will change the world plus the pressure of existential urgency, as the model, while saying he leans toward the inspirational side rather than the "we're going to go out of business tomorrow" extreme. 00:11:50

Coaching and Personal Growth as Infrastructure for the Company

Lochmiller credits executive coach Shane Metcalf and says all CEOs should hire a coach. 00:34:52 He highlights Felicis's practice of giving founders a personal development budget: "they have a budget for like, personal investment, essentially... executive coaching and offsites and just ways of upleveling yourself as a leader, because they're like, 'we're betting on you being able to get better, that you're not a static version of yourself.'" 00:34:59 This is reinforced as a company value: "Live on the infinite growth loop... every single day, we try to improve on who we were yesterday." 00:02:47

Radical Openness to Bad News and Intellectual Flexibility

"I care way less about the good news than I do about the bad news." 00:50:12 His method is "mining for it, just asking for it relentlessly," plus visibly changing his own mind: "When people see examples of you taking criticism and feedback, responding to it, changing your mind, I think it empowers people to be like, 'Oh, actually, my voice should... I should voice my concern.'" 00:50:12

Winning the Social License for Data Centers

Lochmiller argues the data center industry has a messaging problem, not a substance problem, and that local, authentic engagement works. "What the data center industry like their biggest problem is we have a marketing problem. We haven't been able to get the truth out there, and to me, the truth is on our side... when you sit down, you really explain to them what's happening... people are receptive to authenticity and facts." 00:45:41 He also acknowledges real security costs: "We've had to increase our security footprint just given there's crazy people out there." 00:45:41

Conscious Risk-Taking: The Everest Playbook as Management Philosophy

Lochmiller turned a personal Everest experience into a company value ("think like a mountaineer"). He de-risked his 2018 summit by acclimatizing on nearby peaks, cutting passes through the Khumbu Icefall from eight to four: "me and my climbing partner actually designed a different approach to climbing Everest... so that we would acclimatize without taking this stochastic risk in the Khumbu Icefall... conscious risk-taking and intense planning and preparation that really went into a very successful Everest summit in 2018." 00:51:30

2. Contrarian Perspectives

Competition Isn't the Threat. "Startups Die From Suicide, Not Homicide"

Most founders obsess over rivals' funding rounds. Lochmiller argues the opposite: "Most startups die from suicide, not from homicide... it's you fundamentally not executing on the thing that's in front of you and you not building the right culture, not attracting the right people." 00:15:21 And on market structure: "any real, meaningful business opportunity worth pursuing... it's not typically going to be a single winner-take-all business... if someone else is pursuing it, you should almost kind of view that as a compliment, almost a validation." 00:15:21 In a land-grab AI infrastructure market where many assume a winner-take-all race, he sees "far more ways for us to work together than to work against each other." 00:17:02

Sell a Profitable, Cash-Generating Business to Chase a Harder One

Most operators keep the cash cow. Lochmiller divested Bitcoin mining even though "it was a really good business" (per Halligan) and despite internal and external criticism. His claim is that it was never a pivot: "externally, a lot of people thought this was a pivot for us, but internally... this was sort of the plan all along, was this energy-first approach to computing." 00:24:27 Early critics said "Why are you spending any resources on [AI cloud]? You have this great thing. You're like making all this money in Bitcoin." 00:25:13

Bitcoin and AI Are Complementary Loads, Not Competitors

Even while divesting, he argues the two workloads fit together: "AI data centers do not consume energy in the same way that Bitcoin mining does. Bitcoin mining... you can almost give it like a fixed-draw amount of power... AI has a bit more variance in terms of its overall power utilization." 00:25:13 The non-obvious implication is that flexible, interruptible compute can be paired with variable AI loads and energy supply.

Don't Be Mission-Pure Everywhere: Use Every Energy Source

Rather than anointing one winning energy technology, Lochmiller says "we don't think it's a silver bullet... you need to use different energy solutions in different cases." 00:20:48 That cuts against both the nuclear-only and renewables-only camps. Gas today, SMRs, geothermal in Iceland, hydro in Norway, and batteries all fit into a single energy-agnostic platform.

Fire Faster, and Accept That Good People Get Outgrown

Using First Round's "Startup Laws of Physics," Lochmiller says "people scale linearly, companies scale exponentially... the person who was the right person for that role at one point in time may not be able to grow and scale to be the person in that role two years later when the company is 10 times bigger." 00:38:57 His view is that "there's a bias towards you just end up firing people too slowly" and that the right time to fire is when you first start thinking about it. 00:38:21

3. Companies Identified

Crusoe: Vertically integrated AI infrastructure company spanning energy, data centers, manufacturing, and managed AI cloud. Mentioned as the subject. Recently closed a $31 billion valuation round per Halligan, with 2,000+ employees. "Crusoe is building so many of the data centers that power the age of intelligence." (Halligan) 00:01:10; "we recently surpassed 2,000 employees." (Lochmiller) 00:13:38

NYDIG: Bitcoin-focused financial services firm affiliated with Stone Ridge, and the buyer of Crusoe's Bitcoin mining business. "We had a great partner that we felt like this business we built would land in a really good home, which was NYDIG... Stone Ridge has built out one of the largest natural gas businesses in the United States." 00:30:39

Blue Energy: Startup building nuclear power, including nuclear in the ocean; Halligan's Propeller fund is an investor. Lochmiller highlights its approach: "can we leverage the same steam turbine technology that is needed for gas generation... to be a sort of bridge between gas and baseload firm nuclear power." 00:20:48

Oklo: Advanced nuclear (SMR) developer. Crusoe's partnership aims for the first SMR-powered AI factory: "we've another partnership with Oklo that we announced recently where we're trying to bring to life the first SMR-powered AI factory at the Idaho National Lab." 00:20:48

Redwood Materials: Battery recycler and repurposer. Cited for second-life EV batteries: "what's happening with companies like Redwood Materials that are repurposing used electric vehicle batteries and used lithium-ion batteries." 00:20:48

Form Energy: Long-duration storage developer. Lochmiller mentions it among "long-duration storage solutions like Form Energy and Antora that we think will be cost-effective 100-hour batteries." 00:20:48

Antora: Long-duration thermal storage company, named alongside Form Energy. 00:20:50

Felicis: Venture investor in Crusoe, praised for its founder personal-development budget. "Felicis, they had this really cool thing they do... a budget for like, personal investment... executive coaching and offsites." 00:34:59

Exxon and Chevron: Oil supermajors used as the model for Crusoe's vertically integrated strategy. "We're trying to build an AI supermajor in the same way that Exxon and Chevron are supermajors." 00:17:02; "Exxon very famously doesn't hedge their oil exposure." 00:17:02

MongoDB: Public database company. Crusoe's COO and CFO came from there: "my current COO and CFO worked alongside Dave at MongoDB for about a decade." 00:41:46

Databricks: Data and AI platform. Halligan notes Ali Ghodsi "has basically had the same management team for 10 years." 00:41:09

HubSpot: Halligan's former company, cited for its culture and for years of coopetition with Salesforce. 00:43:17

Salesforce: Cited by Halligan as an example of productive coopetition over many years. 00:57:46

Sequoia: Venture firm where Halligan runs CEO peer groups ("The Kids' Table and The Adults' Table"). 00:04:44

Propeller: Halligan's ocean-focused fund, an investor in Blue Energy. Halligan also heard a pitch for vertically integrated ocean rare-earth mining modeled on the Exxon analogy. 00:19:50

First Round Capital: Venture firm whose "Startup Laws of Physics" blog Lochmiller cites on people scaling linearly while companies scale exponentially. 00:38:57

Tesla/SpaceX (via Elon Musk): Referenced as the exemplar of urgency culture. See Elon Musk below. 00:11:50

OpenAI (ChatGPT): Its November 30, 2022 launch was "the magic moment" that created the startup demand for Crusoe's compute infrastructure. 00:25:13

4. People Identified

Chase Lochmiller: Co-founder and CEO of Crusoe; former quant portfolio manager and Stanford grad student in AI; Everest summiteer in 2018. The subject of the interview. "As a person, I've always been attracted to trying to solve hard problems." 00:02:14

Brian Halligan: Host; HubSpot co-founder/former CEO, Sequoia partner, and founder of the Propeller ocean fund. He closes with the takeaway that "He is a very legit CEO scaling that business really well." 00:57:46

Elon Musk: Founder-CEO cited as the best model for sustaining urgency. "Elon has this superpower of simultaneously enabling people to believe that if they're successful, they're going to change the world... while also holding people accountable through the high pressure of 'We're going to go out of business tomorrow.'" (Lochmiller) 00:11:50; also "one of the greatest entrepreneurs... greatest visionaries that humanity's ever had." 00:13:38

Shane Metcalf: Executive coach to Lochmiller; credited with helping him grow into a leader of large organizations and with public speaking. "Shane Metcalf, he was a phenomenal coach for me. I learned so much from him." 00:33:10

Ross Stevens: Founder of NYDIG/Stone Ridge and a friend of Lochmiller's. "Ross Stevens has become a good friend, you know, over the years." 00:30:39

Tejas Shah: CEO of NYDIG, who took over the Bitcoin business. "Just deep amount of respect for that team." 00:30:39

Tim Urban: Author of the Wait But Why blog; his piece on the path to superintelligence influenced Lochmiller's early AI conviction. "One of the more influential things on me was actually Tim Urban had this blog called Wait But Why." 00:24:27

Andrew Ng: Stanford professor whose line "AI is like electricity" shaped Lochmiller's thesis. "He has this famous saying where he says, 'AI is like electricity.'" 00:25:13

Jake (Blue Energy): Blue Energy's founder, whom Halligan and Lochmiller both know. "I know you know Jake from Blue Energy building nuclear in the ocean." (Halligan) 00:19:50

Ali Ghodsi: Databricks CEO, cited by Halligan for a stable ten-year exec team. "He's basically had the same management team for 10 years." 00:41:09

Dev Ittycheria: MongoDB CEO and Halligan's friend; turned over "two per year on the exec team." Lochmiller's COO and CFO worked with him. 00:41:46

5. Operating Insights

Communicate Until You're Sick of It, Then Keep Going

Lochmiller's rule of thumb for message alignment at scale: "I found that as soon as I'm sort of getting sick of hearing myself say something, then we're just scratching the surface... it's sometimes painful the amount of times you have to remind people of what we're doing and why we're doing it, and who's doing what." 00:10:19

Use a Weekly Cross-Org Staff Meeting as the Surface for Friction

For cross-functional pain in a vertically integrated company, the mechanism is simple: "I run a staff meeting once a week that has a lot of the key executives and my direct reports... we cover things across the whole business, and when there are pain points of how teams are working together, they surface in those meetings." He pairs it with a dedicated program management function. 00:09:03

Trust = Aligned Direction + Accountability, Not Hands-Off

Rather than blind delegation, set the plan jointly, understand how it affects other pieces of the business, then hold leaders to it. 00:08:00 Pair this with domain experts in each vertical so you can "move around and go very deep with any individual." 00:00:00

Seek Bad News Actively and Reward Dissent Visibly

"Mining for it, just asking for it relentlessly." Show people examples of you changing your mind, and "run into the burning buildings." 00:50:12 Halligan's complement: train the org to deliver bad news in sizable settings, because people default to good-news reporting upward. 00:57:46

Read the "Mosaic" Before Making Personnel Calls

Rather than relying on a single signal for performance decisions, Lochmiller treats the CEO role as synthesizing "this mosaic of information... from very different parts of the organization... someone super technical... someone that's more of a people person." 00:38:57

6. Overlooked Insights

The Real Moat May Be Operating Loads, Not Just Securing Power

Lochmiller drops a quiet technical point: "AI has a bit more variance in terms of its overall power utilization... you also have seasonality with cooling, you tend to have higher PUE during the day when you're having to run your air condition units and your chillers more aggressively cuz it's hotter than at night." 00:25:13 The throwaway detail suggests that load profile, flexibility, and cooling geography (Iceland and Norway's "very low-cost... cheap, efficient cooling" 00:20:50) are central economic levers, and that operators who can manage variable, demand-responsive loads, a skill honed in Bitcoin mining, may have an edge with utilities and grids. Combined with his comment that Bitcoin and AI "go together very well," it hints that the divested business was a training ground for grid-flexible compute operations.

Founder-Backing Investors Are Now Underwriting CEO Growth Explicitly

Lochmiller mentions in passing that Felicis gives portfolio CEOs a personal-investment budget because "we're betting on you being able to get better, that you're not a static version of yourself." 00:34:59 He says "I haven't seen anybody else do this." The significance is that at least one venture firm treats founder development as a line item on par with capital, and for a company whose complexity scaled from two people to 2,000+ across energy, manufacturing, and AI cloud, it is a cheap but non-obvious source of leverage for investors and a differentiator for founders choosing investors.