Crusoe Sheds Bitcoin Unit to Build Exxon-Style AI Supermajor
Sequoia Capital
Crusoe modeled its vertically integrated infrastructure on oil supermajors to insulate compute margins against volatile power cycles and turn nominal competitors into pipeline customers.
AI-assisted summary of Crusoe CEO Chase Lochmiller: Turning Stranded Energy Into Intelligence — original source. Check the original for full context.
Source published:
Key ideas
- Crusoe closed a funding round at a $31 billion valuation after divesting its profitable Bitcoin mining operation to NYDIG to remove crypto-mandate friction among institutional investors and achieve full operational focus on AI infrastructure. (30:00)
- The company operates as a vertically integrated 'AI supermajor' modeled on ExxonMobil, spanning upstream power generation, in-house switchgear manufacturing, high-density AI data center construction, and a downstream managed AI cloud platform. (16:30)
- Crusoe employs over 2,000 people across radically diverse disciplines, intentionally integrating skilled trade workers like welders and electricians directly with AI researchers optimizing KV cache and inference workloads. (42:39)
- To address community pushback against data center power and water demands, CEO Chase Lochmiller frames the industry's friction as a marketing deficit that requires transparent local engagement rather than political rhetoric. (44:48)
Detailed notes
## Background Published by Sequoia Capital on October 29, 2026, on their podcast feed, this episode features host Brian Halligan in conversation with Crusoe co-founder and CEO Chase Lochmiller. According to the show description, the discussion explores how Crusoe operates as an operationally complex AI infrastructure powerhouse valued at $31 billion, modeling itself on integrated energy giants like Exxon while managing gigawatt-scale data centers, power generation, and cloud compute after divesting its Bitcoin mining arm.
## TL;DR Chase Lochmiller details Crusoe's journey from using flared oilfield gas for Bitcoin mining to building a deeply vertically integrated AI infrastructure supermajor that spans power generation, electrical equipment manufacturing, and managed cloud platforms. He discusses executive leadership mechanics at scale, including managing cross-functional alignment across 2,000 employees, treating continuous self-improvement as an operational necessity, and navigating the painful realities of outgrown personnel and executive turnover. Lochmiller also addresses the public relations challenges facing massive data centers, why startups die from internal mistakes rather than competition, and how he applied rigorous risk-mitigation frameworks to survive summiting Mount Everest.
## Atmosphere & style The conversation carries an intimate, masterclass tone that feels like sitting in on a candid late-stage founder peer roundtable. Lochmiller speaks with measured, analytical precision and disarming humility, openly admitting to early awkwardness in public speaking and the psychological strain of making tough executive decisions. Halligan acts as an experienced, empathetic foil, drawing out tactical playbooks through shared operational scars, creating an intellectual yet grounded atmosphere that balances boardroom realities with high-stakes mountaineering grit.
## Bulletin - `30:00` Crusoe closed a funding round at a $31 billion valuation after divesting its profitable Bitcoin mining operation to NYDIG to remove crypto-mandate friction among institutional investors and achieve full operational focus on AI infrastructure. - `16:30` The company operates as a vertically integrated 'AI supermajor' modeled on ExxonMobil, spanning upstream power generation, in-house switchgear manufacturing, high-density AI data center construction, and a downstream managed AI cloud platform. - `42:39` Crusoe employs over 2,000 people across radically diverse disciplines, intentionally integrating skilled trade workers like welders and electricians directly with AI researchers optimizing KV cache and inference workloads. - `44:48` To address community pushback against data center power and water demands, CEO Chase Lochmiller frames the industry's friction as a marketing deficit that requires transparent local engagement rather than political rhetoric.
## Key takeaways - `02:12` Founders hold dizzyingly complex hyper-scaled companies in their heads through 'boiling frog syndrome'—having lived through every iteration from day one rather than parachuting in cold. - `04:55` Crusoe avoids traditional colocation by vertically integrating the entire stack: generating electrons, manufacturing electrical switchgear, building data centers, and running an AI cloud. - `10:00` Scaling leadership requires repetitive alignment; by the time an executive is sick of hearing themselves repeat the mission and DRIs, the broader company is only just beginning to internalize it. - `11:00` Urgency at scale is sustained by balancing visionary upside with existential survival pressure, preventing large organizations from succumbing to sluggish corporate inertia. - `14:55` Competitor paranoia is counterproductive because most startups die from internal suicide—such as flawed execution and fractured culture—rather than homicide from market rivals. - `16:30` Crusoe models its business architecture on integrated energy supermajors like ExxonMobil, creating structural margin hedges between upstream energy, midstream infrastructure, and downstream cloud compute. - `20:00` Crusoe's flexible energy strategy brings compute directly to underutilized power sources, evaluating flared gas, geothermal in Iceland, offshore nuclear with Blue Energy, and SMRs with Oklo. - `23:26` Bitcoin mining was never Crusoe's end-state; it served as a low-friction stepping stone to monetize stranded gas while validating the modular hardware needed for high-density AI compute. - `30:00` Divesting the cash-generating Bitcoin business to NYDIG was necessary to clear fund mandate restrictions that prevented crypto-averse institutions from investing in Crusoe's AI cloud. - `32:36` Executive communication and public speaking are learned skills developed through intentional reps and coaching, particularly for technical founders emerging from secretive industries like quantitative trading. - `37:48` CEOs almost universally fire underperforming executives too slowly, often struggling with the reality that exponential company growth routinely outpaces linear human development. - `42:39` Crusoe leverages cross-disciplinary collective genius by allowing frontline trade insights, such as electrician wiring schemes, to directly inform building management systems and data center architectures. - `44:48` Public backlash against AI data centers over water and power usage is primarily an industry communications problem that requires authentic, transparent engagement with local officials. - `50:00` Effective executives must deliberately mine for bad news across all organizational tiers, rewarding employees who challenge leadership assumptions and raise operational alarm bells. - `50:54` During his Mount Everest summit campaign, Lochmiller redesigned the expedition's acclimatization schedule on neighboring peaks to reduce dangerous passes through the stochastic risk of the Khumbu Icefall by half.
## Detailed walkthrough ### `00:00` Introduction and the Throughline of Hard Problems
Brian introduces Chase Lochmiller, CEO and co-founder of Crusoe, highlighting how Crusoe has grown from scratch into an operationally intensive powerhouse building the data centers that underpin the current age of intelligence. Brian notes that Chase’s trajectory looks fascinating from the outside: he was originally on track to become an academic professor, pivoted to working as a quantitative hedge fund trader, summited Mount Everest, stepped into the energy sector as an entrepreneur, and now stands at the forefront of gigawatt-scale data center infrastructure. When Brian asks if there is a single throughline connecting these disparate pursuits, Chase explains that his entire career has been driven by an innate gravitation toward tackling difficult, intimidating problems.
As Chase reflects on the challenge of founding and scaling a company, he emphasizes that the CEO job itself is one of the hardest puzzles an entrepreneur can take on:
> “As a person, I’ve always been attracted to trying to solve hard problems. I think that’s been the throughline of my career: take on big challenges, solve hard problems that are exciting and inspiring to me.”
`02:12` The Boiling Frog: Holding a Scaling Company in Your Head
Brian points out that outsiders often misunderstand the realities of stepping into the CEO seat and building a fast-growing enterprise. Chase observes that people frequently look at an established, hyper-scaled company and mistakenly assume the path to getting there was linear. Right now, Crusoe is experiencing an intense, high-velocity environment where enormous projects move simultaneously across multiple continents and disciplines. Dropping a newcomer directly into that environment can feel dizzying and overwhelming. However, a founder CEO benefits from what Chase terms the "boiling frog syndrome": because the founder lived through every historical iteration—from the days when it was just Chase and his co-founder in a room, to their very first hires, to the early milestone of ten people when things already felt crazy—they possess the full historical context required to hold the operation in their mind.
To keep pace with the business as it outgrows each previous phase, Chase explains that a founder must treat self-improvement as an ongoing operational imperative rather than an afterthought. At Crusoe, this philosophy is embedded directly into the company culture:
> “And, you know, this is partially reflected, like, you know, at Crusoe, we have a core company value that’s like live on the infinite growth loop.”
> “Like nobody is a finished product, right? Nothing we’ve done here is a finished product. Everything is a work in progress. Every person is a work in progress. And every single day we try to improve on who we were yesterday.”
Chase stresses that when you harness the compounding returns of daily, iterative personal growth over months and years, the resulting expansion in leadership capacity is astounding.
`04:55` Vertical Integration: From Electrical Manufacturing to AI Research
Brian shares an anecdote from earlier that day, when he hosted the "Kids Table" and "Adults Table"—Sequoia’s peer mastermind roundtables for founder CEOs managing teams ranging from 20 to 2,000 employees. One founder running a rapidly accelerating company confessed: *"My biz is becoming dizzyingly complex beneath me. How the hell do I keep up?"* Chase admits that if someone took him off the street today and parachuted him into Crusoe cold, he would experience that exact same paralyzing anxiety. The only reason he feels calm and grounded is that he has been present for every inch of the journey and has systematically assembled world-class domain specialists across every major pillar of the business.
Crusoe’s operational scope is uniquely demanding because it does not stop at software or traditional colocation. Instead, the company operates as a deeply vertically integrated stack. On the physical energy side, Crusoe generates power, builds power plants, and collaborates directly with independent power producers and utilities. In manufacturing, it operates facilities that produce critical electrical switchgear and infrastructure hardware. In data center operations, it designs and deploys everything from small-scale edge sites to massive multi-hundred-megawatt AI factories. Finally, at the top of the stack, Crusoe runs a managed AI cloud platform staffed by software engineers and AI researchers who optimize model serving, inference throughput, and KV cache management.
```mermaid flowchart TD accTitle: Crusoe Vertically Integrated AI Infrastructure accDescr: Four-tier stack spanning energy generation, physical manufacturing, data centers, and AI cloud platform.
subgraph Power["1. Upstream Energy"] A["Power Generation & Utilities"] B["Plant Construction & Electrons"] end
subgraph Industrial["2. Hardware & Facilities"] C["Manufacturing Electrical Switchgear"] D["Building AI Factories (Small to Gigawatt Scale)"] end
subgraph Software["3. Cloud & Intelligence"] E["Managed AI Cloud Platform"] F["AI Research & KV Cache Serving Optimization"] end
Power --> Industrial Industrial --> Software ```
Chase highlights how having genuine specialists in each pillar allows him to zoom in and out without getting bottlenecked:
> “What’s made me very scalable as a leader is that we have these critical areas of domain expertise where I can move around and I can go very deep with any individual, but I trust them a lot to be making the right decisions as we’re sort of scaling things.”
- **Energy and Electrons:** Developing power plants, tapping stranded energy, and coordinating with regional utilities and IPPs. - **Physical Manufacturing:** In-house production of custom electrical equipment to eliminate supply chain chokepoints. - **Data Center Real Estate and Operations:** Constructing and running high-density AI factories spanning varied scales. - **Managed AI Cloud:** Delivering cloud software, developer APIs, and deep algorithmic optimization for inference and model hosting.
`07:15` Trust, DRIs, and Navigating Cross-Functional Friction
Brian notes that when he surveyed the founders in his CEO group about how many executives they truly trust to act as a Directly Responsible Individual (DRI) across the entire company, most leaders answered that they only trusted three or four individuals at that level. He asks Chase whether this deep trust came naturally or required deliberate unlearning of micromanagement habits. Chase clarifies that his style is neither hands-off abdication nor passive delegation. Rather, his framework centers on establishing deep alignment around the overarching strategic plan and understanding how decisions in one division reverberate through the others.
Once alignment is locked, Chase grants leaders autonomy and holds them strictly accountable to their commitments:
> “Look, it’s not that I’m like, ‘Hands off, go for it, go do whatever you want.’ That’s not really my approach to management. It’s really: I want to understand what we’re doing in this vertical, we have alignment on a direction we’re headed, I get it and 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 from a plan perspective.”
Brian presses further on the organizational mechanics of cross-functional delivery, asking how Chase prevents the inevitable friction that occurs when initiatives cut across four radically different business units—such as power development, hardware manufacturing, real estate construction, and AI software engineering. Chase acknowledges that cross-functional coordination is an inherent struggle for any organization:
- **Program Management:** Dedicated program managers coordinate handoffs and resource allocation between distinct functional teams. - **Weekly Executive Staff Meetings:** Chase convenes his core executive leadership team once a week for an all-hands business review. Inter-team dependencies, resource bottlenecks, and structural friction points are intentionally surfaced and resolved in real time. - **Vertically Integrated Reality:** Because Crusoe cannot deliver its core product without tight cohesion between electrons, steel, and code, cross-functional alignment is an everyday survival skill rather than a theoretical management exercise.
`10:00` Urgency at Scale: Elon's Tension and Repetitive Alignment
Transitioning between the distinct layers of a vertically integrated company requires relentless collaboration across the entire stack. Reflecting on his transition from small teams to managing a sprawling organization, Chase Lochmiller notes that in early environments, stating a goal once was usually sufficient for everyone to align naturally. In a larger enterprise, however, alignment demands extreme, almost uncomfortable repetition. Lochmiller emphasizes that internalizing this dynamic was an essential personal adjustment:
> “As soon as I'm sort of getting sick of hearing myself say something, then we're just scratching the surface of when when I'm just hitting the point where, you know, I'm saying it enough times.”
He describes the communication layer—reminding people what the company is doing, why it matters, and who owns which responsibility—as a painful yet non-negotiable tax of scaling. Startups inherently enjoy the agility to pivot and move faster than incumbents. To prevent Crusoe from succumbing to the sluggish inertia typical of large corporations, leadership intentionally engineers an intense culture of pace and urgency, rejecting any complacency or sense of having "made it."
Lochmiller points to Elon Musk as the undisputed benchmark for managing urgency at scale. Musk excels at holding two diametrically opposed psychological realities at the same time: inspiring his workforce with a grand, world-altering vision of financial and technological impact, while concurrently stoking existential accountability by warning that the enterprise will collapse tomorrow without maximum human effort. Lochmiller acknowledges that while he does not calibrate his own management style to quite the same apocalyptic extreme, maintaining that constructive tension remains paramount:
> “Elon has this superpower of, you know, simultaneously enabling people to believe that if they're successful, they're going to change the world and there will be this incredible financial outcome and this incredible outcome that's driving massive impact, while also holding people accountable through the high pressure of like, we're going to go out of business tomorrow.”
```mermaid flowchart LR accTitle: The dual-tension leadership model accDescr: Visualizing how visionary upside and existential urgency combine to drive sustained operational pace without complacency. subgraph Drivers["Leadership Pressures"] Vision["World-Changing Impact<br/>(Inspiration & Upside)"] Urgency["Existential Pressure<br/>(Accountability & Survival)"] end subgraph Outcome["Organizational Execution"] Alignment["Eliminate Entitlement<br/>& Coasting"] Speed["Sustained Urgency<br/>at Scale"] end Vision --> Alignment Urgency --> Alignment Alignment --> Speed ```
`13:30` Learning to Lead: Books, Coaches, and Scar Tissue
Crusoe is the largest company Lochmiller has ever worked for, recently crossing the threshold of 2,000 employees. Because he had never managed an enterprise of that magnitude, his transition required deliberate study and humility. Prior to founding Crusoe, his reading habits ranged broadly across science fiction, technical manuals, novels, and economics. Once the company launched, he pivoted completely, binging on every business book recommended to him, working closely with seasoned mentors, and hiring an executive coach to systematically evolve into a true people leader.
This growth process demanded that Lochmiller embrace mistakes and accumulate firsthand "scar tissue" on the job. A critical shift in his mindset came from letting go of the constant founder paranoia that stems from watching competitors announce large funding rounds or launch overlapping products. Lochmiller points out that early on, watching market announcements on platforms like X could easily trigger distraction, until an advisor gave him a piece of enduring startup wisdom:
> “Most startups die from suicide, not from homicide.”
He explains that outside competitors almost never inflict the fatal blow on an ambitious enterprise. Instead, startups fail internally by stumbling on execution, fracturing company culture, mismanaging personnel, or drifting away from their highest-priority business opportunities. If an addressable market is large and compelling, peer companies pursuing the same frontier should be viewed as validation of the opportunity rather than an existential threat.
`16:30` The AI Super Major: Why Crusoe Models Itself on Exxon
When asked where he sits on the founder paranoia sine wave regarding rival infrastructure plays, Lochmiller clarifies that he feels very little competitive anxiety. He views the physical compute layer supporting artificial intelligence—or superintelligence (SI)—not as a single-winner market, but as a vast foundation that mirrors the oil and gas industry value chain:
| Segment | Oil & Gas Function | AI / Compute Infrastructure Analog | |:--- |:--- |:--- | | **Upstream** | Drilling wells and extracting raw oil and gas from the earth | Sourcing and developing raw stranded energy, power plants, and generation assets | | **Midstream** | Transporting hydrocarbons via pipelines to plants and refineries | Data center infrastructure, networking, electrical equipment, and physical capacity | | **Downstream** | Refining hydrocarbons into jet fuel, gasoline, and usable plastics | Running the AI cloud stack, compute clusters, and delivering usable intelligence workloads |
Just as ExxonMobil and Chevron operate as integrated supermajors across all three tiers, Crusoe is built to be an "AI supermajor." This integrated architecture provides an organic, structural hedge against cyclical commodity swings. For example, when crude oil prices decline, Exxon's upstream margins compress at the wellhead, but downstream refining margins expand because input feedstocks become significantly cheaper.
Furthermore, this structural posture enables Crusoe to partner constructively with entities that might appear to be direct competitors on the surface. In traditional energy, an upstream producer may compete directly with Exxon in drilling, yet willingly pay to run its extracted volume through an Exxon-owned regional midstream pipeline. Lochmiller adopts that identical strategy in AI infrastructure, noting:
> “We're willing to partner with businesses that we may compete with in other cases. [There are] more ways for us to help each other than for us to work against each other.”
- Lochmiller views the infrastructure required for artificial intelligence as the single largest economic opportunity of our lifetime. - Crusoe intentionally rejects winner-take-all assumptions, expecting multiple massive enterprise winners to coexist across the compute and energy landscapes. - The interview pivots as the host interjects to bring up a pitch he received at his ocean-focused fund, Propeller.
`20:00` The Energy-First Approach: From Flared Gas to Geothermal and Nuclear
Responding to an analogy about extreme vertical integration—comparing deep-sea rare-earth mining to building an end-to-end modern Exxon—the discussion turns to Crusoe’s evolving power strategy and investments in frontier energy, including offshore nuclear developer Blue Energy. While Crusoe gained prominence by capturing stranded oilfield flare gas, projects like their massive Abilene facility increasingly run on natural gas and grid resources. Chase Lochmiller explains that Crusoe’s mandate is not dogmatically tied to a single fuel source, but rather focused on activating megawatts wherever compute can unlock underutilized power. To catalyze new clean energy technologies, Crusoe is testing everything from small modular reactors (SMRs) with Oklo at the Idaho National Laboratory to ocean-based nuclear deployments that can repurpose existing steam turbine infrastructure as a bridge between gas generation and steady nuclear baseload.
That flexibility extends internationally and across the entire renewable spectrum. In Iceland and Norway, Crusoe takes advantage of geological anomalies where cheap, abundant geothermal energy and hydropower sit far from dense population centers, enjoying natural cold-climate cooling. Instead of spending billions attempting to transmit electricity across long distances to where people live, Crusoe’s philosophy is to bring the compute demand directly to the power generation. Lochmiller stresses that meeting the scale of modern intelligence will also require long-duration grid storage, highlighting battery ventures like Redwood Materials, which repurposes spent electric vehicle batteries, and companies like Form Energy and Antora Energy, which develop cost-effective 100-hour storage. As Lochmiller puts it, “The demand from intelligence for energy is voracious, and it’s really going to accelerate human progress as it has more access to energy.”
```mermaid flowchart TD accTitle: Crusoe Energy Portfolio Strategy accDescr: Diagram showing diverse energy sources routed through Crusoe infrastructure to power AI compute.
subgraph EnergySources["Diverse Power Generation"] A["Flared & Natural Gas"] B["Geothermal & Hydro<br/>(Iceland, Norway)"] C["Advanced Nuclear & SMRs<br/>(Blue Energy, Oklo)"] D["Storage & Batteries<br/>(Redwood, Form Energy, Antora)"] end
subgraph Infrastructure["Crusoe Compute Infrastructure"] E["Modular Generators & Data Centers"] F["Behind-the-Meter AI Factories"] end
subgraph Output["End Application"] G["Ubiquitous AI Compute & Intelligence"] end
EnergySources --> Infrastructure Infrastructure --> Output ```
- **Frontier Nuclear Partnerships:** Crusoe is working with Blue Energy on floating offshore nuclear power and has partnered with Oklo to develop the first SMR-powered AI factory at the Idaho National Laboratory. - **Steam Turbine Bridge:** Modern gas generation and nuclear plants share compatible steam turbine technology, offering a mechanical pathway to transition fossil-based generation into clean nuclear baseload over time. - **Geographic Arbitrage:** Locations like Iceland and Norway offer abundant, low-cost hydropower and geothermal power paired with inexpensive ambient cooling, making them ideal targets for modular compute deployment. - **Long-Duration Storage:** Lochmiller sees massive potential in 100-hour battery solutions from companies like Form Energy and Antora Energy to smooth out renewable intermittency.
---
`23:26` The "AI-Pilled" Plan All Along: Turning Flared Gas into Compute
Crusoe’s origin as a Bitcoin miner using stranded oilfield gas looked like an extraordinarily lucrative, specialized business—making their eventual decision to divest Bitcoin assets and build an AI cloud seem like a radical pivot. Lochmiller rejects that narrative, maintaining that early investors and veteran employees knew that compute for artificial intelligence had been the endgame from day one. He describes becoming deeply “AI-pilled” years earlier after reading influential essays by Tim Urban on *Wait But Why* regarding the path to artificial superintelligence, which convinced him that multilayer neural networks would scale nonlinearly as long as they were fed sufficient compute and data. That belief was cemented at Stanford by his professor Andrew Ng, who famously analogized AI to electricity: a generalized utility that would transform every domain rather than a single vertical application.
Lochmiller reasoned that if data and compute were the primary bottlenecks to intelligence, then compute would inevitably become an energy problem at scale. In oilfields such as the Bakken in North Dakota, energy companies drill for oil and vent or burn off the associated natural gas because building pipeline infrastructure is cost-prohibitive. Crusoe stepped in with small modular generators and data centers, turning an environmental liability into immediate cash flow. Bitcoin mining was merely the initial, lowest-friction monetization engine to validate the modular concept: “It was the easiest way to take energy and turn it into money.” When Lochmiller began telling people he wanted to build an AI cloud alongside mining, early observers scoffed, questioning why he would divert resources from an operation that was printing money.
The early efforts to launch the AI cloud were messy. Lochmiller’s team stumbled through iterations of building GPU clusters, cold-calling academic researchers at universities and asking if they could run workloads on GPUs powered by flare gas in the middle of a North Dakota field. Everything changed on November 30, 2022, with the public release of ChatGPT, which ignited explosive enterprise demand for the exact storage, networking, and high-density compute infrastructure Crusoe had spent years prototyping. While Lochmiller still believes Bitcoin and AI complement one another on the grid—because Bitcoin provides a flat, highly responsive baseline load that can ramp down instantly during peak demand while AI training and inference exhibit higher variance and seasonal spikes—the market shift made clear that intelligence, not hashing, was the ultimate destination for Crusoe's power.
- **Foundational Thesis:** Instead of building costly transmission lines to transport stranded energy to markets, Crusoe moves the compute market directly to the stranded energy source. - **Academic and Intellectual Influences:** Tim Urban's writing on artificial superintelligence on *Wait But Why* and Andrew Ng’s lectures at Stanford helped convince Lochmiller that compute-driven intelligence would serve as a universal meta-science. - **The Bakken Proving Ground:** Crusoe deployed its first modular generators and data center trailers directly onto oil well pads in North Dakota, using waste flare gas that oil producers were previously paying to dispose of. - **Load Characteristics:** Bitcoin mining provides an almost perfectly flat, throttleable electrical load ideal for demand-response balancing, whereas AI workloads introduce wider power swings and volatility driven by model training runs and inference traffic.
`30:00` Divesting the Bitcoin Mining Business for Focus
PUE (power usage effectiveness) naturally fluctuates throughout the day: chillers and air conditioning units must run aggressively during the heat of the afternoon, whereas nighttime allows for far more natural cooling, with cluster workloads causing additional ebbs and flows. Against that technical operational backdrop, the interviewer pushes Chase Lochmiller on a major strategic pivot: if Lochmiller envisioned Crusoe as an AI company from the start, why sell off the profitable Bitcoin mining business that was actively paying the bills rather than keeping both?
Lochmiller explains that the transaction was made easier because they found an ideal acquirer in NYDIG. He had built a strong friendship over the years with Ross Stevens and held deep respect for CEO Tejas Shah and their team. NYDIG represented the rare intersection of operators who understood natural gas assets at scale while remaining deeply embedded in the Bitcoin ecosystem, making them trustworthy stewards for the business. Ultimately, the choice came down to operational focus and capital allocation:
> “What drove me to divest that business was really—focus. And, you know, we are doing a lot at Crusoe and increasingly we were doing more and it felt like 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?”
Beyond internal bandwidth, investor appetite created an external forcing function. Pitching both lines made Crusoe’s narrative far too complicated for institutional allocators. Multiple investors told Lochmiller they loved the AI infrastructure story but were structurally prohibited by their own fund mandates from investing in anything tied to crypto. To establish Crusoe as the category-defining, vertically integrated AI infrastructure company, they had to eliminate the Bitcoin distraction entirely. Lochmiller emphasizes that he remains a massive personal supporter and holder of Bitcoin—viewing it as an essential, non-state-controlled, decentralized, digitally native open monetary ecosystem—yet underscores that going all-in on AI infrastructure was unconditionally the right strategic move for Crusoe.
```mermaid flowchart TD accTitle: Crusoe Divestment and Focus Strategy accDescr: Diagram illustrating the operational and investor pressures that led Crusoe to divest its Bitcoin mining operation to NYDIG to focus entirely on AI infrastructure.
subgraph Tension["Strategic Tensions"] direction TB A["Bitcoin Mining Business<br>(Profitable, gas assets)"] B["AI Infrastructure Business<br>(High growth, capital intensive)"] C["Investor Friction<br>(Crypto mandates blocked AI checks)"] end
subgraph Divestment["The Divestment Solution"] direction TB D["Partner: NYDIG<br>(Ross Stevens & Tejas Shah)"] E["All-In Strategic Focus:<br>Vertically Integrated AI Cloud"] end
A -->|"Competed for capital & focus"| B C -->|"Demanded clean narrative"| E A -->|"Transferred to natural gas/crypto stewards"| D B -->|"Retained full corporate attention"| E ```
`32:36` Learning Polish and Shane Metcalf's Coaching
Observing Lochmiller’s communication style on the call, the interviewer remarks that he comes across as unusually polished compared to most startup CEOs, asking whether that was deliberately developed through coaching. Lochmiller confirms that public speaking never came naturally to him. Coming from quantitative finance—a hyper-secretive industry where quants collaborate openly with internal teammates but speak to nobody externally for fear of leaking trading alpha—he spent years never presenting in public beyond occasional university guest lectures.
Transitioning into an outward-facing CEO required deliberate, grueling effort to plan out what to say and how to say it. Over time, constant practice transformed the skill: Lochmiller found that over-preparation eventually became unnecessary, and he now performs best speaking off the cuff because he has built the muscle to articulate clearly what is in his head. When asked whether every founder CEO should hire an executive coach, Lochmiller gives an unequivocal "Yes."
He credits his coach, Shane Metcalf, with transforming his leadership:
> “Public speaking was never something that came naturally to me. It was something I had to work really, really hard at because I just wasn't doing it.”
- Shane Metcalf worked intimately with Lochmiller, serving as both an executive sounding board and an emotional ballast. - In quantitative finance, secrecy is paramount to protect proprietary alpha; founders from that background often have zero prior reps communicating with press, investors, or large public audiences. - Continuous real-world reps eventually allowed Lochmiller to shift from scripted, anxious presentations to speaking fluidly and thinking aloud off the cuff.
`35:00` Felicis's Founder Budget and Navigating People Problems
Lochmiller highlights an investor practice from venture firm Felicis that he believes the broader venture community should emulate: providing portfolio founders with a dedicated budget explicitly reserved for personal development, executive coaching, off-sites, and leadership up-leveling. Felicis takes the explicit stance that early-stage investing is an underwriting of human potential—betting that the CEO is not a static individual, but someone capable of expanding their personal capacity as the enterprise scales.
Lochmiller compares a great executive coach to a therapist. Even during routine sessions where immediate tactical breakthroughs seem elusive, the ongoing cadence lays an indispensable foundation of mutual trust. When an existential crisis hits, the coach already possesses an intimate map of the founder's blind spots, emotional triggers, and instinctive coping habits, allowing them to intervene and steer the CEO through high-stakes moments.
Ultimately, Lochmiller argues, corporate leadership boils down almost entirely to interpersonal dynamics:
> “People say this and it's—you know, I may sound like I'm, you know, repeating the common folklore, but like, you know, so many of the problems with building companies are people problems, right? Being able to effectively navigate the people side of this is like it's that's like most of the challenges.”
The central dilemma of the CEO seat is steering brilliant, opinionated, and highly ambitious leaders toward a single vector—getting them to commit to a path even when consensus is impossible. The hardest recurring executive decisions involve deciding whether to continue investing in an underperforming team member or replace them, and arbitrating deep philosophical disagreements between two exceptional employees.
`37:48` The Mosaic of Firing and Exponential Scaling
The conversation turns to the common Silicon Valley adage to "hire slow and fire fast." The interviewer admits to having routinely violated this advice by firing far too slowly, and asks Lochmiller whether he has shared that regret and what indicators finally trigger a termination. Lochmiller confirms that nearly all CEOs suffer from an inherent bias toward moving too slowly when separating from employees. In practice, by the time an executive begins seriously contemplating letting someone go, that initial moment of doubt was usually the objective point when the trigger should have been pulled.
When assessing when an executive or employee is no longer the right fit, Lochmiller explains that it rarely stems from a single glaring complaint or a peer declaring someone subpar. Instead, the CEO must assemble a disparate "mosaic of information"—synthesizing small data points collected across totally different functions, from deeply technical engineers to intuitive people leaders.
He grounds this operational friction in organizational physics, citing a classic insight from First Round Capital:
> “People scale linearly, companies scale exponentially. So you sort of have this dynamic...”
Because elite startups grow at an exponential rate while individual humans develop along a linear curve, high-growth companies inevitably outpace early contributors who were indispensable at prior stages. Identifying this delta is an ongoing, high-stakes judgment call that sits squarely on the CEO's shoulders.
`40:00` Outgrowing Roles and Executive Turnover
Chase Lochmiller reflects on the painful reality that ambitious company growth often outpaces individual capability. Even when an executive team is challenged to invest in themselves and learn continuously, someone who was exceptional at one stage may simply not be equipped to run their division two years later when the enterprise has scaled tenfold. Letting those leaders go is emotionally taxing because founders forge deep personal bonds working late nights and early mornings alongside them. Lochmiller notes that every firing decision is unique—"almost like every case is a snowflake"—and the burden rests entirely on the CEO to make the pragmatic judgment call that protects the broader business.
Brian Halligan contrasts two executive management styles from recent conversations: Ali Ghodsi at Databricks, who kept almost the exact same leadership team intact over a decade with only a single executive departing, versus Dev Ittycheria at MongoDB, who cycled through an average of roughly two executive changes per year. Lochmiller positions himself squarely between those two philosophies. He notes that Crusoe’s current COO and CFO both worked alongside Ittycheria at MongoDB for roughly ten years without turning over. Drawing on his background in competitive sports, Lochmiller explains that a locker-room bond cannot supersede the ultimate goal of winning; when a teammate struggles and becomes a weak link, everyone on the roster feels it and expects leadership to make a change.
> “And that's a hard reality to face as a leader because you end up in the trenches with these people.” — Chase Lochmiller
- Firing decisions must balance personal empathy for people giving everything to the mission against the structural needs of a company growing 10x. - Executive team stability varies radically among top software leaders, ranging from Ali Ghodsi’s near-zero turnover to Dev Ittycheria’s regular refreshes. - High-performing teams instinctively recognize when a peer is struggling, making decisive personnel changes necessary for team morale and execution.
`42:39` Bridging Welders and AI Researchers: Collective Genius
Halligan observes that while software companies like HubSpot hire employees with relatively similar profiles, Crusoe spans an unusually broad workforce that includes trade professionals like welders alongside cutting-edge AI researchers. Lochmiller explains that Crusoe’s culture is anchored in a core value called "tapping into the collective genius." Because building power plants, manufacturing electrical enclosures, and managing cloud clusters requires vastly different capabilities, no single person possesses all the required domain knowledge. Instead, the company deliberately fosters cross-disciplinary pollination where insights from the shop floor inform computational architecture.
Practical innovations regularly emerge from tradespeople that directly benefit the AI research team. Electricians working on physical wiring schemes, for example, have contributed critical insights that reshaped how Crusoe designs building management systems and optimizes data center airflow. Lochmiller emphasizes that framing the organization around mutual learning enables physical builders and computer scientists to see how their respective crafts fit together into an integrated, defensible platform.
```mermaid flowchart TD accTitle: Crusoe Collective Genius Cross-Disciplinary Integration accDescr: Diagram showing how physical trade disciplines and AI research feed into Crusoe's unified infrastructure platform.
subgraph Trades["Physical & Electrical Trades"] W["Welders & Assemblers"] E["Electricians & Field Techs"] end
subgraph Software["AI & Cloud Computing"] R["AI Researchers"] S["Cloud Infrastructure Engineers"] end
subgraph Core["Unified Infrastructure Platform"] B["Building Management Systems"] G["Gigawatt-Scale Data Centers"] end
W --> B E -->|"wiring insights"| B R -->|"compute requirements"| G S --> G B --> G ```
> “It's like no one person has all the domain expertise to deliver the full solution that we're trying to bring to bear.” — Chase Lochmiller
- Crusoe’s multi-industry workforce deliberately pairs blue-collar skilled trades (welders, pipefitters, electricians) with AI research scientists. - Front-line electrical and mechanical innovations feed directly into the software running the building management systems. - Success relies on ensuring neither cultural cohort feels subordinate to the other.
`44:48` The Data Center Backlash: A Marketing Problem
Halligan brings up the growing public backlash against AI infrastructure, pointing to protests at venture conferences and data center development becoming an active election issue over concerns about power grids and water consumption. Lochmiller responds that everyone at Crusoe genuinely wants to do good for the world, arguing that their infrastructure can be deployed in ways that benefit local water conservation, generate municipal tax revenue, and fund public schools. However, he acknowledges that the broader data center sector faces a severe communications deficit where misinformation fills the void.
When entering contentious local jurisdictions, Lochmiller bypasses defensive talking down and directly meets with county judges, civic officials, and community members to lay out transparent, ground-truth facts. He argues that voters and local officials are fundamentally receptive to authenticity over political rhetoric when developers show respect and present balanced economic realities. Despite this constructive engagement, Lochmiller admits that the heightened political climate has forced Crusoe to scale up executive security to safeguard against bad actors looking to make a public statement.
> “Their biggest problem is we have a marketing problem. We haven't been able to get the truth out there.” — Chase Lochmiller
- Data center siting has evolved into a prominent political and electoral flashpoint surrounding utility grids and natural resource consumption. - Crusoe emphasizes tangible community benefits: expanding local tax bases, supporting public schools, and investing in water-conscious cooling designs. - Heightened industry protests and polarized rhetoric have necessitated increased personal and physical security protocols for leadership.
`47:45` Intellectual Flexibility, Thick Skin, and Speaking Truth to Power
Facing packed town halls with angry, shouting residents offers zero playbook training, but Lochmiller handles hostility by leaning on empathy and intellectual flexibility. Rather than becoming defensive, he listens to the underlying anxieties, acknowledging that even strongly held convictions must adapt when confronted with new local facts. By maintaining a growth mindset, leadership can separate performative public hysteria from valid community grievances that Crusoe should incorporate to become a better environmental and civic steward.
Halligan asks whether having thick skin is essential for a chief executive, and Lochmiller affirms that a CEO’s primary daily function is serving as the shock absorber for bad news. The toughest challenge as Crusoe scales is the growing distance between the executive suite and frontline workers. Lochmiller constantly works to flatten that gap so employees throughout the organization feel empowered to challenge leadership, voice ground-level operational problems, and speak truth to power without fear of reprisal.
> “A hundred percent. Yeah. I mean, just your job is to like be the receiver of bad news very frequently and then be able to think with a clear mind and make decisions in light of that bad news.” — Chase Lochmiller
- De-escalating hostile municipal meetings requires emotional composure and the humility to separate raw anger from actionable constructive feedback. - Strong opinions loosely held: Lochmiller stresses willingness to revise operational plans whenever presented with legitimate new facts. - A critical executive discipline at scale is keeping feedback loops open so frontline operators can voice bad news and challenge leadership assumptions directly.
`50:00` Mining for Bad News and Encouraging Dissent
Brian asks how a leader can cultivate an environment where employees feel safe pushing back, noting that while employee number 20 will openly challenge a founder, employee number 2,000 tends to place the CEO on an intimidating pedestal. Chase explains that hearing genuine criticism requires actively “mining for it” by soliciting negative feedback relentlessly across the organization. When employees observe leadership visibly accepting critique, engaging with it constructively, and publicly changing their minds in response, it breaks the default assumption that executives only want to be flattered with positive updates.
Chase emphasizes that while good news is always welcome, it does very little to help steer the company through operational bottlenecks. He stresses that an executive’s primary utility lies in running directly toward problems rather than celebrating wins. As Chase puts it: “Generally, like the good news is great. Like I want to hear it. But like I care way less about the good news than I do about the bad news.” By focusing attention on systemic flaws, broken buildings, and areas where the business is performing poorly, leadership signals to the entire company that raising alarm bells is rewarded rather than punished.
`50:54` Mount Everest: Designing Around the Stochastic Risk of the Khumbu Icefall
Turning to Chase’s personal background, Brian brings up his summit of Mount Everest and asks what preparation and standing on the peak actually feel like. Chase explains that Crusoe’s company culture directly mirrors this mindset through its formal company value to “think like a mountaineer”—a code built around meticulous preparation, planning, and enduring prolonged hardship. Chase had initially attempted to summit Everest from the Nepal side (the South Col route) in 2014, but that season ended in tragedy when a massive avalanche in the Khumbu Icefall killed 16 Sherpas. Although his wife hoped the disaster would close that chapter of his life for good, the mountain haunted him, leading to an agreement that he would attempt it once more before they started having children.
In planning his return, Chase dissected the mountain into distinct risk profiles. The area above Camp 4 (at or above 8,000 meters) is known as the Death Zone; while deadly due to severe oxygen deprivation, it represents a known physiological challenge that a climber can consciously train for and manage. In contrast, the section between Base Camp and Camp 1—the Khumbu Icefall—is an actively moving glacier that functions like a collapsing ice jungle gym. In that zone, climbers face purely stochastic, uncontrollable hazards: a hanging serac can collapse or an avalanche can sweep through without warning. As Chase explains: “There are ways I can consciously train and prepare for the high elevation and I can manage my risk up there more effectively and I know consciously I'm taking that risk. The risk I felt less comfortable with was this stochastic risk.”
On a typical commercial expedition, climbers make four full round trips (eight passes) through the Khumbu Icefall simply to acclimatize on the higher camps before their summit push. To eliminate that exposure, Chase and his climbing partner redesigned their entire campaign. Instead of running laps through the icefall, they spent weeks climbing and sleeping atop other nearby Himalayan peaks to build high-altitude red blood cell counts safely. When they finally set foot on Everest, they needed only a single acclimatization rotation through the icefall before launching their final push, cutting their exposure to the Khumbu Icefall in half from eight passes down to four.
```mermaid flowchart TD accTitle: Everest Acclimatization Strategy Comparison accDescr: Chart comparing the traditional 8-pass commercial acclimatization through the Khumbu Icefall with Chase's redesigned 4-pass strategy using neighboring peaks.
subgraph Traditional["Traditional Commercial Expedition"] T1["Base Camp"] -->|"Passes 1-2"| T2["Camp 1 & Return"] T2 -->|"Passes 3-4"| T3["Camp 2 & Return"] T3 -->|"Passes 5-6"| T4["Camp 3 & Return"] T4 -->|"Passes 7-8"| T5["Summit Push"] note1["Total: 8 Passes Through Khumbu Icefall"] end
subgraph Redesigned["Chase's Redesigned Expedition (2018)"] R1["Climb Neighboring Peaks"] -->|"Acclimatize at high summits"| R2["Everest Base Camp"] R2 -->|"Passes 1-2"| R3["Single High Rotation"] R3 -->|"Passes 3-4"| R4["Summit Push"] note2["Total: 4 Passes (50% Risk Reduction)"] end ```
- In 2014, a massive serac collapse triggered an avalanche in the Khumbu Icefall that killed 16 Sherpas, aborting Chase’s first attempt. - Standard commercial climbs require four rotations up and down Everest (Base Camp to Camp 1, then Camp 2, then Camp 3, and back down each time), forcing eight total transits through the icefall. - Physical conditioning for the 2018 campaign combined high-mileage trail running with dedicated high-altitude acclimatization trips.
`55:25` Pushing Through the Balcony to the Summit
Chase recounts the physical and psychological toll of the final summit push in 2018. Leaving Camp 4 at 11:00 PM in total darkness, he pushed upward into the sub-zero night without having eaten substantial food in nearly 48 hours. By the time they reached the Balcony at roughly 8,600 meters—just 250 vertical meters shy of the 8,848-meter summit—Chase was severely dehydrated, suffering from agonizing leg cramps, and on the verge of turning around. His climbing partner rallied him to keep moving forward, noting that despite the exhaustion and cramping, he was not exhibiting acute mountain sickness or putting his life at immediate medical risk.
When the sun broke over the horizon, the shifting light transformed his mental state completely. Seeing the sunrise from the upper slopes of the Himalayas flooded him with adrenaline and elevated his spirits for the final, grueling climb across the knife-edge ridges to the summit. Standing on the highest point on Earth on a crystal-clear, sunny morning, Chase was overwhelmed not by personal hubris, but by a profound sense of humility and cosmic scale. As Chase remembers: “And it makes you feel small, right? And I think there's something very interesting about like, you know, having that sense of like, I'm small and I'm inspired by the world around me and just like the magnificence of existence.” They spent a full hour taking photographs and taking in the panoramic views before beginning the descent.
- The push began at 11:00 PM out of Camp 4 in freezing darkness. - The Balcony sits at approximately 8,600 meters; the summit of Everest stands at 8,848 meters. - The team enjoyed rare, sunny, and calm summit conditions, remaining at the peak for approximately one hour.
`57:31` Brian's Key Takeaways on Scaling and CEO Qualities
After thanking Chase for the conversation, Brian reflects on his primary takeaways from watching Chase evolve into the CEO of a multi-billion-dollar enterprise over an eight-year journey. First, Brian points out that Chase possesses a deep innate hunger for difficult, intellectually daunting challenges. Coming into the role without prior CEO experience, he turned himself into a deliberate “learning machine”—reading extensively, hiring an executive coach who functioned as an objective sounding board and therapist, and methodically training himself to overcome natural deficits in public speaking until he became articulate and smooth in front of investors and employees.
Second, Brian highlights Chase's healthy orientation toward competition. While many Silicon Valley founders dismiss competitive threats, Brian recalls his own intense, multi-year market battles against Salesforce, arguing that founders should lean into competition and even establish direct lines of communication by regularly inviting competitors out for coffee. Finally, Brian praises Chase's active insistence on hearing bad news and his readiness to reverse course when presented with new facts. As Brian observes: “He has very strong opinions, as you've heard. But if the facts change, he's willing to change his mind. And being stubborn is a really tough quality in a CEO. It makes you quite hard to follow.”
- Chase relied heavily on executive coaching to gain context, listen through complex operational problems, and navigate the psychological burdens of the role. - Public speaking was an acquired, practiced discipline rather than an innate talent. - Effective CEOs intentionally train their growing organizations to deliver unfiltered negative reports rather than comforting pleasantries. - Brian closes the episode by teasing upcoming founder conversations on *Longest Strange Trip*.
## Notable quotes > `37:48` People scale linearly, companies scale exponentially. — Chase Lochmiller
> `20:00` The demand from intelligence for energy is voracious, and it’s really going to accelerate human progress as it has more access to energy. — Chase Lochmiller
## Mentioned Crusoe, Chase Lochmiller, Brian Halligan, Elon Musk, ExxonMobil, Chevron, NYDIG, Ross Stevens, Tejas Shah, Shane Metcalf, Felicis, First Round Capital, Ali Ghodsi, Databricks, Dev Ittycheria, MongoDB, HubSpot, Blue Energy, Oklo, Redwood Materials, Form Energy, Antora Energy, Tim Urban, Wait But Why, Andrew Ng, Stanford University, ChatGPT, Salesforce, Sequoia Capital
## Video description Chase Lochmiller runs one of the most operationally complex companies in AI: Crusoe builds power plants, manufactures electrical equipment, operates gigawatt-scale data centers, and runs a managed AI cloud on top, all under one roof. He just closed a round at a $31 billion valuation, and he got there by selling off the Bitcoin mining business that was paying the bills.
Chase explains why Crusoe is modeled on Exxon, not a cloud provider, and why he partners with companies that look like competitors. He makes the case that a company that outgrows the person who got it there is the rule, not the exception, and that CEOs almost always fire too slowly. He also talks candidly about hiring a coach, working hard to become a polished speaker when it didn’t come naturally, and why the data center industry’s biggest problem is a marketing problem.
I meet with a group of Sequoia founder CEOs between 20 and 2,000 employees, and one of them told me his business is becoming dizzyingly complex beneath him. Chase has the clearest answer I’ve heard for how to hold that in your head.
And yes, we talk about summiting Everest.
0:00 – Introduction
2:12 – The boiling frog: how a founder holds a 2,000-person company in his head
8:16 – DRIs, the weekly staff meeting, and saying it until you're sick of it
11:19 – Urgency at scale: Elon's "change the world or die tomorrow" tension
14:55 – Startups die of suicide, not homicide
19:29 – Energy first: flared gas, Iceland hydro, and nuclear
23:26 – Bitcoin to AI was the plan all along
32:36 – Polish is learned: coaches, public speaking, and Felicis's self-improvement budget
37:48 – Firing too slowly, and when the company outgrows the person
42:39 – Welders and AI researchers, the data center backlash, and mining for bad news
50:54 – Everest: designing around the Khumbu Icefall's stochastic risk
57:31 – Brian's takeaways
[Source](https://pscrb.fm/rss/p/traffic.megaphone.fm/CPUAI3487587675.mp3)