Host Aditya Agarwal enthusiastically welcomes Mike Schroepfer ("Schrep") to SPC for his second visit, noting they now have "bigger digs." Aditya describes Schrep as a dear friend, former boss, and mentor who has shown him the way through his journey of trying to do interesting things in life.

Aditya shares an anecdote from their recent Founder Fellowship kickoff where Ruchi (presumably another colleague) was asked to name three people who profoundly affected her life over the past decade, and she named Schrep as one of them. Aditya jokes that he himself didn't make Ruchi's list, emphasizing that Ruchi is "a tough cookie."

"Schrep is a dear friend, also my former boss, and a mentor to me. I think as I go through my journey of life trying to do interesting things, he's someone I constantly look back to as someone who has kind of shown me the way in many ways."

Timestamp: [0:07-1:13]

Aditya Agarwal provides an overview of Mike Schroepfer's impressive career in Silicon Valley. Schrep graduated from Stanford in 1997 and had several ventures before joining Facebook in 2008, where he remained until just a couple of years ago. He's still involved with Facebook in some capacity.

Aditya mentions they want to first reminisce about Facebook before diving into Schrep's current work with Gigascale and climate tech. He prompts Schrep to share what he's most proud of from his time at Facebook.

Schrep humbly responds by first acknowledging Aditya's earlier compliments, stating that the relationship has been mutually beneficial: "You describe it as much more of a one-way relationship than it is. I think the real joy of being in tech and being in this industry is I get to learn from everyone I work for, work with, and that definitely includes you and Ruchi."

"The real joy of being in tech and being in this industry is I get to learn from everyone I work for, work with, and that definitely includes you and Ruchi."

Timestamp: [1:13-2:31]

When asked by Aditya what he's most proud of from his time at Facebook, Schrep identifies two main areas: specific aspects of the engineering culture, and Facebook's consistent ability to deliver developer platforms used worldwide.

He highlights several examples of Facebook's influential developer tools: React became the de facto standard for web development, React Native for mobile, and in AI development, PyTorch and Jax are used for major research and development. These accomplishments span three different technology stacks.

Schrep emphasizes that it wasn't just about creating individual tools, but building a culture that could consistently develop tools so well-loved that they were adopted globally: "It's sort of less the one thing we did and more that—and I think the two are related—we built a culture that was able to build a set of tools that were so well-loved that kind of got adopted by the whole world."

"We built a culture that was able to build a set of tools that were so well-loved that kind of got adopted by the whole world."

Timestamp: [2:31-3:21]

Aditya notes that Facebook has often led with product and user impact rather than engineering, contrasting it with Google's engineering-first DNA. He mentions that Facebook's engineering achievements like its first open source project, Thrift, have often been kept in the background. He asks how Schrep made engineering excellence something celebrated within the company.

Schrep explains that it starts with humility at the top—acknowledging that leadership doesn't have all the answers. As VP of Engineering or CTO, his personal opinions about technologies mattered less than "the wisdom of the crowds"—what developers chose when given options because they want tools that are easier and more fun to use.

He describes Facebook's engineering culture as the right combination of top-down and bottom-up approaches. The top-down element focuses on problem identification and prioritization: "What problems do we need to work on, which are the highest priority, and how would you sort amongst them?" Once priorities are established, there's tremendous bottom-up energy directed at solving those problems.

"It actually starts from having a little bit of humility sort of at the top of not knowing the answers to all the questions... my opinion is a lot less important than sort of the wisdom of the crowds—which is, hey, when developers are given a choice, what do they choose? Because they want the one that is like easier to use and more fun."

Timestamp: [3:21-5:32]

Schrep elaborates to Aditya on why he loves open source, describing it as "the introduction of market forces within a company." When Facebook open-sourced their tools, it created external validation—if the broader community didn't adopt them, it signaled they needed improvement or replacement.

This market pressure ensured their tools had to be genuinely great. If another open source project was overtaking Facebook's solution, they would consider switching to it. This approach combines bottom-up innovation with market forces and a touch of humility at the leadership level.

However, Schrep cautions against taking the bottom-up approach too far into "absentee management." He maintained strong opinions on priorities and was consistently clear about them: "You should be able to predict with very high precision...if you ask me 'Hey Schrep, what's up?' you know what's coming out of my mouth...because I'm going to say the same damn thing over and over again."

These priorities would shift deliberately over time—from reliability one year to performance the next, then mobile, then AI. The key was being "on message" about the goals while giving teams freedom to determine how to achieve them.

"Open source is the introduction of market forces within a company...if we open source something and nobody else likes it, but we think it's awesome, I think the world's right and we're wrong."

Timestamp: [5:32-6:48]

Aditya recognizes the nuanced leadership approach Schrep describes, noting there's always a tension between confident leadership and avoiding dictatorial management. He appreciates how Schrep chooses priorities and identifies problems without necessarily dictating solutions.

This observation leads to a discussion about a recent blog post by Moxie Marlinspike that examined how engineering breakthroughs often come when people see beyond abstraction layers. The post drew an analogy to engineering organizations, contrasting the standard "divide and conquer" scaling approach with more integrated approaches.

Aditya explains that while most organizations scale by siloing work and partitioning problems—expecting that combining well-executed parts will result in a functioning product—the most interesting breakthroughs often come when teams transcend these boundaries. This is one reason why startups can innovate more easily than larger organizations.

"There's always this balance in leadership where you want to lead confidently but you don't want to be dictating everything. You want to take in feedback, but navigating that tension can be tricky sometimes. I liked what you said there—you're choosing priorities, you're choosing the problems to look at and the problems to push away for another day. You're not necessarily dictating the solutions to all those problems."

Timestamp: [6:48-7:35]

Schrep confirms Aditya's analysis by discussing Facebook's deliberate choice to maintain a monolithic architecture for their codebase rather than adopting microservices, which were becoming popular elsewhere. At Facebook, engineers could check out the entire codebase, refactor any part of it, and check it all back in. This approach was often criticized by outsiders who advocated breaking the system into microservices.

Schrep explains that microservices create friction points between modules, making it difficult to conceptualize and implement changes across many services. By contrast, Facebook's monolithic approach enabled engineers to quickly refactor entire parts of the codebase.

He then extends this principle to hardware development and climate tech, noting that relying too heavily on third-party suppliers creates similar friction points. When a company only builds a small part of a product and depends on thousands of suppliers, it's essentially created a "microservice" architecture in the physical world that slows innovation.

"Microservices were a big hot thing for a while and a lot of people—and we were explicitly at Facebook for a very long time a monolith. Like, you could check out the whole code base, you could refactor the whole thing if you wanted and check the whole damn thing back in and start it up and run it... What you do [with microservices] is you build these friction points between each of these modules and now if I want to change across these 2,000 microservices, man, I just can't even conceptualize it."

Timestamp: [7:35-9:51]

Schrep extends the discussion with Aditya about integrated versus modular development to hardware and climate tech. He explains how companies that build most components in-house can innovate much faster than those relying on extensive supply chains.

With a heavily outsourced approach, even minor modifications—like needing a slightly larger bolt or switching to magnesium materials—require lengthy communications with suppliers and manufacturing cycles. By contrast, companies that control the entire manufacturing process can quickly implement changes, test them, and iterate.

Schrep praises SpaceX as a pioneer in this vertically integrated approach, noting that they built almost everything in-house because they didn't have an established supply chain. He observes that with modern technologies like 3D printing and rapid prototyping, more companies can feasibly adopt this integrated approach, at least at small scales.

"Companies rely too much on third party suppliers. When you have a product where you only build a small part of it and I have 2,000 suppliers, I basically built a microservice. Because now if I'm like 'Hey, I need a bolt that's a little bit bigger' or 'Can you make it out of magnesium?' I got to talk to them, they got to do it, they manufacture. If I build the entire thing in house, I can go talk to [someone] like 'Hey, can you spin another version of this?' Got it tomorrow, and then boom, we've got a new version done."

Timestamp: [9:22-9:57]

• Engineering excellence at Facebook was built on a balanced approach of top-down priority setting and bottom-up solution development
• Open source functions as a market force within companies, providing external validation for internal tools
• Leadership requires the right balance of confident direction-setting without dictating solutions
• The monolithic architecture at Facebook allowed engineers to make sweeping improvements quickly without friction between modules
• The same principle applies to hardware: vertically integrated companies can innovate faster than those dependent on extensive supply chains
• Humility at the leadership level creates space for innovation while clear priorities provide direction
• Learning is mutual in effective mentorship relationships—even leaders learn continuously from those they work with
• Facebook's developer tools (React, PyTorch, etc.) became industry standards because they prioritized ease of use and developer experience
• Consistent messaging about priorities helps everyone understand the direction while giving freedom on implementation
• The "wisdom of crowds" is often more valuable than executives' opinions when selecting technologies and approaches

Timestamp: [0:07-9:57]

People:

• Aditya Agarwal - Host, SPC
• Ruchi - Colleague mentioned as having named Schrep as one of three people who profoundly affected her life
• Moxie Marlinspike - Author of blog post discussing how engineering breakthroughs often come from seeing past abstraction layers

Companies:

• Facebook - Where Schrep served from 2008 until recently
• Stanford - Where Schrep graduated from in 1997
• SpaceX - Mentioned as a pioneer in vertical integration for hardware development
• Google - Referenced as having engineering-first DNA, contrasting with Facebook's product-first approach

Technologies:

• React - Facebook-developed web development framework that became industry standard
• React Native - Facebook-developed mobile development framework
• PyTorch - Facebook-developed AI research and development framework
• Jax - AI development framework from Facebook
• Thrift - Facebook's first open source project
• Monolith vs. Microservices - Architectural approaches compared with Facebook favoring monolithic approaches

Organizations:

• Gigascale Capital - Climate tech fund co-founded by Schrep
• SPC - The organization hosting the conversation (Founder Fellowship mentioned)

Timestamp: [0:07-9:57]

Schrep continues his discussion on vertical integration in hardware development, emphasizing that companies who control their entire production process can iterate at a vastly different pace: "When you're rapidly iterating what the product is...your cycle time is not like 10% faster, it's like two orders of magnitude faster, and that's the difference between success and failure."

He pushes back against common misconceptions about Meta's hardware divisions, noting that despite sarcastic comments suggesting web companies don't understand rapid iteration, the exact opposite is true: "The best hardware companies are actually the fastest at iterating, and they're asking the question at every moment in time—what's the quickest, dumbest, cheapest way I can build something?"

Schrep shares an example from one of Gigascale's portfolio companies that's building a high-pressure, high-temperature turbine. Their first prototype was assembled from Home Depot parts: "They're like, 'Let me get a blower from Home Depot and a bunch of tin foil and some other stuff and light some crap on fire in my backyard and learn some lessons.' That is the team that's going to make a lot of rapid progress."

"The best hardware companies are actually the fastest at iterating, and they're asking the question at every moment in time—what's the quickest, dumbest, cheapest way I can build something?"

Timestamp: [10:04-11:01]

Aditya affirms Schrep's point about rapid iteration, sharing that throughout his career he's encouraged teams to measure progress in cycles: "If I need 30 to 50 cycles to get something that's good in v1 and people will start to use, if my cycle time is a month, that's a world of difference from a cycle time of three days."

This prompts Schrep to reflect on Facebook's engineering culture, particularly how new engineers were onboarded through a "hazing ceremony" of checking out the entire codebase and being tasked with fixing a random bug anywhere in the stack. This approach fostered a culture of total ownership and agency, which led to tremendous speed improvements in ways that weren't fully appreciated at the time.

He recalls the evolution from weekly website updates to continuous deployment—a goal many thought was a terrible idea. But by prioritizing speed above all else, they built infrastructure and tooling designed specifically to make deployment safer while maintaining rapid iteration. This included techniques like canary deployments, where changes would first be pushed to a small number of machines to avoid breaking the entire site at once.

"When you set these goals, speed is the most important thing. Time to make a change is the most important thing. So when we do something and something goes wrong, the question isn't 'Oh no, should we do something different?' It's 'How do we fix that problem? How do we resolve it more quickly?'"

Timestamp: [11:01-13:07]

Aditya asks Schrep what other unorthodox approaches beyond vertical integration he's identified in the climate tech sector that enable more innovation and speed.

Schrep responds enthusiastically, explaining that as he's explored the field, he's become increasingly optimistic due to several "compounding factors" or "tailwinds" that make climate tech particularly promising now. He draws a parallel to Meta's AI strategy, where they benefited from compute and data tailwinds that made progress easier each year.

In climate tech, Schrep identifies several key tailwinds:

1. Rapidly decreasing costs: Renewable energy sources like solar, batteries, and wind are becoming dramatically cheaper every year, even without additional innovation. He notes that lithium-ion batteries have become 85% cheaper in the last decade, 97% cheaper since their introduction in 1991, and about 50% cheaper in China in just the last year.

2. Improving economics: This continuous cost reduction means that even if a clean technology is currently more expensive than its fossil fuel counterpart, waiting just a couple of years could make it cost-competitive or even cost-advantaged without requiring any additional breakthroughs.

"In the climate world, you have a bunch of tailwinds in your favor... lithium-ion batteries have gotten 85% cheaper in the last 10 years, 97% cheaper since they were introduced in the market in '91, they've got about 50% cheaper in China in the last year alone. So if I just like... 'I want to build this thing and it's going to be battery-powered, but it's like twice as expensive as the gas-powered thing'—wait two years. Without doing anything, you're cost competitive."

Timestamp: [13:07-14:55]

Addressing Aditya's question about speed in climate tech innovation, Schrep highlights how modern prototyping tools have dramatically accelerated development cycles:

3D Printing Revolution: While 3D printing didn't transform the consumer market as expected, it has become a "total game-changer" in industrial applications. Rather than waiting six weeks for parts to arrive or paying for expensive air freight, companies can now fabricate components in their labs the same day. Schrep notes that Meta built its own 3D printing lab specifically for this rapid prototyping capability.

Computational Modeling: Schrep shares an example of a Gigascale portfolio company developing an offshore wave power system—a 200-meter tall, 10-meter wide tube that generates clean energy by bobbing in ocean waves. Before building such a massive structure, they use computational fluid dynamics to simulate its performance. Their physical prototype performed within 3% of the simulation predictions, giving them confidence to make design improvements virtually before building the next version.

This combination of advanced modeling and rapid physical prototyping has collapsed development timelines: "When we do a redesign and do the rev 2, we can do it all in software... 'Alright, we made these changes to make these optimizations and it's a 20% efficiency improvement plus or minus 3% error, we're good to go.' Cycle time is really, really fast."

"3D printing was this thing that everyone... thought everyone would have a 3D printer in their home, and it kind of didn't happen on the consumer side. On the industrial side, it's a total game-changer... because you can rapidly prototype materials and rapidly prototype things. Instead of waiting six weeks for my part to come back in or to air freight it from somewhere, I have it in my lab today."

Timestamp: [14:55-16:17]

Schrep describes a third key factor accelerating climate tech innovation: a cultural shift toward first principles thinking. He acknowledges the influence of companies like SpaceX and Tesla in promoting this approach, where engineers question established practices rather than defaulting to catalog solutions.

"There's a whole set of people who are trained to like, when I solve a problem, I literally get the catalog out and see which parts I can buy to solve that problem... And then you've got this new generation of people that are like, 'Wait, what the hell is this thing made of? Why is this $10,000?'"

He shares an example from another Gigascale portfolio company that's disrupting the chemical industry with a cleaner process for producing ethylene, "the most made organic molecule on Earth" found in products like yogurt cups. When faced with expensive membrane components, the company's engineers broke down the costs to fundamentals—weighing materials and calculating what they should reasonably cost based on raw materials.

This first-principles approach led them to manufacture their own membranes at 1/100th the price, dramatically improving both cost and cycle time. Schrep explains that this mindset starts with basic questions: "Anything in the physical world was—how much does it weigh and what is it made of? And then with that and a little bit of a notebook, I can tell you about how much it should cost. And if it's like 10x that, somebody's ripping you off."

"There's a new generation of people that are just willing to first-principle something out... You start with anything in the physical world: how much does it weigh and what is it made of? And then with that and a little bit of a notebook, I can tell you about how much it should cost. And if it's like 10x that, somebody's ripping you off."

Timestamp: [16:17-18:01]

Aditya observes that Schrep's approach to climate tech strongly resembles the engineering ethos he cultivated at Facebook: "This reminds me a lot of the Facebook ethos, which is that whatever the problem is, just work the problem, go into the source, go deeper—there's nothing that a good engineer can't solve. That was kind of the way I used to think about my job."

He highlights a key question that drives this mindset: "The best question in the world is 'why?' Like, 'Hey, we can't do this.' 'Why?' And until you get to 'because it violates the second law of thermodynamics,' I don't like the answer."

Aditya then observes how this engineering-oriented, quantitative problem-solving approach seems to form the foundation of Gigascale's investment thesis. He characterizes it as bringing a toolbox of solutions while remaining agnostic to the specific problems—just looking for opportunities to create impact. He then asks Schrep where this approach has served him well in climate tech and where it has fallen short.

Schrep acknowledges that while he's shared many successes, there have definitely been challenges and mistakes along the way. The primary limitation he's encountered is failing to fully account for market dynamics and customer behavior: "You can get a little overly fixated on the technology... This thing is 5x cheaper, 10x cheaper, but I didn't at all understand the market dynamics—like, who are the people paying for this and what do they care about?"

He notes the frustrating reality that many market participants don't behave rationally: "Unfortunately, there are a lot of non-rational actors. There are a lot of people that when you show up with a cheaper product, won't buy it for a whole variety of reasons. So you've got to get a lot smarter on the business side."

"The best question in the world is 'why?' Like, 'Hey, we can't do this.' 'Why?' And until you get to 'because it violates the second law of thermodynamics,' I don't like the answer."

Timestamp: [18:01-19:58]

Schrep shares a compelling example of computational modeling's impact through a Gigascale portfolio company developing offshore wave power technology. Their innovation is "an amazing technology" that functions as "just a little buoy that floats and bobs up and down in the water" to generate continuous clean energy.

The company has designed a 200-meter tall, 10-meter wide tube that harnesses wave energy in a way that could potentially become "the cheapest form of energy production we'll have on the planet." However, building and testing full-scale prototypes of such massive structures would be prohibitively expensive and time-consuming.

Instead, they leverage computational fluid dynamics to simulate their system's performance with remarkable accuracy. When they finally deployed a physical prototype in the Pacific Ocean off the coast of Washington, the real-world performance matched their simulations within plus or minus 3%.

This precision in computational modeling allows them to rapidly iterate designs entirely in software: "When we do a redesign and do the rev 2, we can do it all in software—'Alright, we made these changes to make these optimizations and it's a 20% efficiency improvement plus or minus 3% error, we're good to go.'" The result is dramatically faster cycle times for innovations that would otherwise require months or years of physical testing.

"You have computational modeling... They can use computational fluid dynamics to simulate their system. Their system is a 200-meter tall tube that's 10 meters wide, so it takes a while to build it. So you kind of want to make sure it works before you build the 200-meter tall one... The final thing, when we dropped it in the Pacific Ocean off the coast of Washington, was plus or minus 3% to our simulations."

Timestamp: [15:24-16:11]

Schrep shares another example of first-principles innovation from a Gigascale portfolio company working to disrupt the chemical industry. They're focusing on ethylene—a compound most people haven't heard of despite it being "the most made organic molecule on Earth" and a component in everyday items like yogurt cups.

Traditional ethylene production uses a "dirty, nasty process" that Schrep doesn't detail. In contrast, the portfolio company has developed a cleaner electrochemical approach that directly converts electricity into ethylene, potentially transforming how this essential industrial chemical is manufactured.

However, their progress was initially hindered by a specialized membrane component that suppliers were charging exorbitant prices for. Rather than accepting these costs, the team disassembled the membrane to understand its composition, then began manufacturing their own version that proved to be "100x cheaper" and faster to produce.

This example highlights how the combination of first-principles thinking and vertical integration can simultaneously address both environmental impact and economic viability. By challenging industry assumptions and bringing critical components in-house, the company dramatically improved both their cost structure and cycle time for innovation.

"We have another company that's building—they're basically trying to disrupt chemicals, a big, dirty industry that nobody cares about. They make this chemical ethylene—probably no one's ever heard of ethylene, it's the most made organic molecule on Earth. If you've ever had a yogurt cup, you've got ethylene in it. The way we make it is this dirty, nasty process... These folks have figured out an electrochemical way, so you just take electricity in and you get ethylene out."

Timestamp: [17:07-17:38]

• Speed of iteration in hardware development can be two orders of magnitude faster with vertical integration, making the difference between success and failure
• The best hardware companies prioritize rapid, cheap prototyping over perfection—like building initial versions from Home Depot parts
• Measuring progress in development cycles rather than calendar time helps teams understand the impact of cycle time reduction
• Facebook's journey from weekly deployments to continuous integration required prioritizing speed while building safety mechanisms like canary deployments
• Climate tech benefits from multiple "tailwinds" including rapidly decreasing costs of renewable technologies—lithium batteries have become 97% cheaper since 1991
• 3D printing has transformed industrial prototyping by eliminating weeks-long waits for components
• Computational modeling allows climate tech companies to test designs virtually with high accuracy (within 3%)
• First-principles thinking helps engineers break free from expensive catalog parts by understanding fundamental material costs
• The engineering approach of "working the problem" and always asking "why" until reaching physical limitations drives innovation
• Technical superiority doesn't guarantee market success—non-rational market dynamics can prevent adoption of objectively better solutions

Timestamp: [10:04-19:58]

People:

• Aditya Agarwal - Host, SPC
• Mike Schroepfer ("Schrep") - Guest, co-founder of Gigascale Capital, former CTO of Facebook/Meta

Companies & Organizations:

• Gigascale Capital - Climate tech investment fund co-founded by Schrep
• Meta - Mentioned regarding their 3D printing lab and engineering culture
• Facebook - Discussed regarding their deployment practices and engineering culture
• SpaceX - Credited as a pioneer in vertical integration and first-principles thinking
• Tesla - Mentioned alongside SpaceX for first-principles approach to problem-solving
• Home Depot - Referenced as source of components for rapid prototyping

Technologies & Concepts:

• Lithium-ion batteries - Noted to have decreased in cost by 97% since 1991
• 3D printing - Discussed as transformative for industrial prototyping but less impactful in consumer markets
• Computational fluid dynamics - Used to simulate wave energy systems before building physical prototypes
• Canary deployments - Facebook's technique of rolling out changes to a small subset of machines first
• Continuous deployment - Evolution of Facebook's deployment strategy from weekly to immediate updates
• Ethylene - "Most made organic molecule on Earth," used in yogurt cups and other products
• Electrochemical process - Cleaner approach to ethylene production using electricity
• Wave power system - 200-meter tall tube that generates clean energy from ocean waves
• First-principles thinking - Approach of breaking down problems to fundamental physical and economic components

Concepts:

• Tailwinds - Forces that make progress easier over time without additional effort
• Cycle time - Speed of iteration in product development
• Vertical integration - Controlling the entire production process rather than relying on suppliers
• DORA metrics - Industry metrics for software deployment frequency and rollback capabilities

Timestamp: [10:04-19:58]

Schrep continues his discussion about the limitations of a purely technological approach to climate solutions, emphasizing the crucial importance of market dynamics and customer adoption: "If you don't have customers that want to buy this thing, nothing else matters because you need the funding to go after and do it."

When asked if he considers himself a capitalist in his climate work, Schrep responds emphatically: "Hell yeah, I'm a capitalist! We're going to use technology to drive down prices, make better products, but it's got to be a business. You've got customers that want to buy this thing, and if you don't, nothing else matters."

He reflects on his early investing mistakes, acknowledging that he initially got overly excited about fascinating technologies without adequately evaluating their market potential: "In the early days of my investing, I got super excited about nerding out on all these crazy cool technologies, and I certainly picked a couple of technologies that I still think are super cool, but it turns out the markets they were going after didn't have customers with enough willingness to buy. From a business perspective, it was a cool technology, bad business. Not making that mistake again."

"Hell yeah, I'm a capitalist! We're going to use technology to drive down prices, make better products, but it's got to be a business. You've got customers that want to buy this thing, and if you don't, nothing else matters because you need the funding to go after and do it."

Timestamp: [20:05-20:40]

Ryan Barrett joins the conversation, asking Schrep to walk through his "minus one journey"—the process of figuring out what to do next after leaving Meta/Facebook. He notes that Schrep was in the enviable position of being able to tackle almost anything with his background and connections, and asks how he navigated that period of infinite possibility to eventually focus on climate tech.

Schrep acknowledges the emotional challenge of having too many options: "For anyone who's been in a situation where you have a lot of choice, it's hard. This is something for those who don't have choice, it's hard to have empathy for." He references studies showing that having too many choices often leads to decision paralysis and second-guessing.

He explains that when starting something entrepreneurial, the key consideration is finding what genuinely excites you: "Your primary job is to inject energy into it. If I don't show up for work, things don't happen... If you have to inject energy, you have to have that energy come from somewhere, and it's got to be you, which means you got to be really damn excited about what you're working on."

"If you're doing anything entrepreneurial or anything where you're driving it, you realize that your primary job is to inject energy into it. If I don't show up for work, things don't happen... If you have to inject energy, you have to have that energy come from somewhere, and it's got to be you, which means you got to be really damn excited about what you're working on."

Timestamp: [20:40-23:04]

Schrep describes how the COVID-19 pandemic in 2020 became a turning point that eventually led him to focus on climate change. During lockdown in the Bay Area, he found himself with unexpected free time despite still working hard at Meta.

Before the pandemic, his life had been in "survival mode" for much of his career: "If people asked 'What are your hobbies, Schrep?' I was like, 'Uh, it's like trying to stay alive, trying to do my job, see my kids, sleep a little bit. I don't have time for hobbies.'" He recalled his previous entrepreneurial experience, joking that when he started his company in 2000, "We took Sunday afternoons off—that was our day off."

During COVID, he had more time to reflect because despite still working intensely, the pandemic eliminated many logistical demands: "I wasn't driving my kids anywhere, there's no activities, I wasn't driving into the office. Things on Zoom were actually more efficient." This gave him space to observe the global crisis unfolding, which crystallized an insight about human behavior:

"It reinforced a view I had, which is like humans are actually pretty good in an acute crisis... We rally, we're an awesome species on that. When it's a slow burn going to happen over the next 10 or 20 years and we got to start stocking up toilet paper to be prepared—we suck." This realization led him to consider what major slow-burning problems humanity faces, and "climate is the mother of all the slow-burn problems."

"Humans are actually pretty good in an acute crisis... We rally, we're an awesome species on that. When it's a slow burn going to happen over the next 10 or 20 years... we suck. And so I was just like, 'Well, okay, what are the real big slow-burn problems?' And climate is the mother of all the slow-burn problems."

Timestamp: [23:04-24:26]

As Schrep began studying climate change, he quickly realized the severity of the situation: "If you actually start to study it, it gets really scary really fast. You're in a situation where in 20-30 years, parts of the earth will be uninhabitable. Sea level rise could be like 3 feet, maybe 6 feet by the end of the century. This is just horrific—billion climate migrants, really scary, scary stuff."

This led him to a critical question: "Is there anything we can do about it, or are we just totally screwed?" To find out, Schrep adopted a "learner's mind" approach while still working full-time at Meta. He hired a consulting firm to educate him on climate basics, initially thinking he would direct his philanthropic efforts toward the problem.

"I still had my full-time job. I figured I would just keep doing my thing at Meta. 'Get me smart'—I just did the basic Climate 101, like what are sources of emissions, what can we do about them, how could you get rid of them?" He spent about a year asking foundational questions and began directing philanthropic resources toward promising solutions, initiatives that continue today.

Through this education process, Schrep started meeting climate entrepreneurs and gained a deeper understanding of the scale of the challenge: "This is like a 10 trillion dollar problem, it's 50 trillion—tens of trillions of dollars." This massive scale led to an important realization about what it would take to address climate change effectively.

"You start looking at that and go like, 'What problems has philanthropy solved that are on the order of 10 [trillion]?' Like none. Governments can't do it. Philanthropy and government can get the ball rolling, but at the end of the day, what you're doing is reforming six, half a dozen $2 trillion industries, and the only way that works is self-interest."

Timestamp: [24:26-25:57]

Schrep explains his investment philosophy for climate solutions, which focuses on moving from what he calls "the green premium"—a term he admits to disliking—to what he terms "the green discount": "Our job is to move stuff from the green premium to what we invest in, which is the green discount—it's like, I'm going to win because my thing's cheaper."

He describes how his conversations with climate entrepreneurs revealed compelling cost curves that resembled patterns he'd seen in successful technology transformations. These companies were showing trajectories where clean technology costs were declining rapidly while approaching the cost of fossil fuel alternatives: "You have a cost curve that looks like this, which is we're going on this crazy cost curve down, and then there's this line which is the current fossil fuel one. And you're like, 'Hey, as soon as we cross this line, we're cheaper.'"

This realization—that climate solutions could soon become the economically superior choice, not just the environmentally responsible one—became the focal point for his investments: "I've seen a lot of things that look like that. Let's go!"

Schrep concludes by returning to his advice for entrepreneurs facing too many options: "It's hard. Pick something you're really excited about. I don't care, doesn't have to be climate—it's my thing, you can do something else: biology, social issues. There's a lot of problems to solve in the world. Pick it because you like it and you're excited about it. I think it's the most important thing."

"Our job is to move stuff from the like green premium, which is a word I hate, to what we invest in, which is the green discount—it's like, I'm going to win because my thing's cheaper."

Timestamp: [25:57-27:07]

Aditya praises Schrep's origin story of coming to mission-driven work and transitions to a key question about business models in climate tech. He notes that for years, climate change has been an externality to markets—a cost society would slowly pay over the next century but wasn't priced into economic decisions.

He points out that while this is changing through regulatory measures in some places, for a long time the dominant business model in climate has been carbon offsets or credits, which have faced recent challenges. Aditya asks Schrep whether his "green discount" approach is meant to replace these mechanisms or whether both approaches have a place.

Schrep responds by identifying "two giant forces, two giant tailwinds" driving climate business models. First, he clarifies that philanthropy and government have crucial roles to play: "If you look at any interesting industry transition, they happened on the back of philanthropy and government. I think they have a massive role to play here, as does industry."

He then describes the first major market force: increasing regulatory pressure. "People get fixated on carbon markets, but there are a whole bunch of regulations in Europe, in the US, in California, from shipping to food to others, that start to put pressure on people to understand their emissions." This regulatory pressure isn't voluntary—it's becoming legally mandated.

Schrep offers an example from Meta's data centers, where their first major innovation focused on efficiency: "We didn't put air conditioners in it. We use evaporative cooling. It's a massive savings over a very long time period." He then provides a concrete example of regulatory impact: "In New York City, if you have a building over 35 stories, you have to pay taxes on your imputed carbon output. That's a law on the books that's being enforced."

These regulations create business opportunities for solutions that help building owners reduce their emissions and avoid costly tax bills.

"There is the sort of pricing in of [climate impacts]... There are a whole bunch of regulations in Europe, in the US, in California, from shipping to food to others, that start to put pressure on people to understand their emissions, reduce them in different ways. And that doesn't just mean carbon—that often means being more efficient in their supply chain, using better electricity."

Timestamp: [27:07-29:57]

• Climate solutions must be economically viable businesses with real customers willing to pay—technological superiority alone is insufficient
• Having too many options can be paralyzing; entrepreneurs should focus on problems they're genuinely excited about to sustain the energy needed to drive progress
• The COVID pandemic provided Schrep with unexpected time for reflection, revealing how poorly humans handle slow-burning crises compared to acute emergencies
• Climate change education revealed increasingly alarming projections: uninhabitable regions within 20-30 years, potential 3-6 feet of sea level rise by century's end
• The scale of climate change (tens of trillions of dollars across multiple industries) means philanthropy and government alone cannot solve it—market forces must drive adoption
• Schrep's investment strategy focuses on "green discount" rather than "green premium"—solutions that win because they're cheaper, not just because they're cleaner
• Cost curves that show clean technologies becoming cheaper than fossil fuel alternatives are key indicators for promising climate investments
• Two major forces drive climate business models: increasing regulatory pressure (mandatory emissions regulations) and pure economic advantage
• Regulations across Europe, the US, and California are creating concrete business opportunities for emissions-reducing solutions
• Efficiency remains the first rule in addressing climate impacts, as demonstrated by Meta's focus on energy-efficient data centers using evaporative cooling

Timestamp: [20:05-29:57]

People:

• Mike Schroepfer ("Schrep") - Guest, co-founder of Gigascale Capital, former CTO of Facebook/Meta
• Aditya Agarwal - Host, SPC
• Ryan Barrett - Host, SPC (joins conversation in this segment)

Companies & Organizations:

• Gigascale Capital - Climate tech investment fund co-founded by Schrep
• Meta/Facebook - Schrep's former company, mentioned regarding data center efficiency innovations
• Home Depot - Referenced in the previous segment for rapid prototyping components

Technologies & Concepts:

• Evaporative cooling - Energy-efficient cooling system used in Meta's data centers instead of air conditioning
• Green premium - Term Schrep dislikes, referring to clean technologies that cost more than conventional alternatives
• Green discount - Schrep's preferred investment focus, clean technologies that win by being cheaper than fossil fuel counterparts
• Carbon markets/offsets/credits - Climate business models Aditya mentions as having faced recent challenges
• Cost curves - Declining trajectories for clean technology costs that eventually cross below fossil fuel alternatives

Events:

• COVID-19 pandemic (2020) - Catalyst that gave Schrep time to reflect and eventually focus on climate change
• Climate change projections - Referenced predictions including uninhabitable regions within 20-30 years, 3-6 feet sea level rise by century's end, and billion climate migrants

Policies:

• New York City building emissions tax - Specific regulation mentioned that taxes buildings over 35 stories based on carbon output
• Regulations in Europe, US, California - Various regulatory frameworks putting pressure on emissions across different sectors

Concepts:

• Acute crisis vs. slow-burn problems - Contrast between human responses to immediate threats versus gradual challenges
• Decision paralysis - Psychological effect where too many choices makes decision-making harder
• Climate as an externality - Economic concept where environmental costs haven't been priced into market decisions
• Industry transition - Process by which major economic sectors transform, requiring both government/philanthropy and market forces

Timestamp: [20:05-29:57]

Continuing his discussion of climate business models, Schrep highlights the second major force alongside regulatory pressure: technological advancement. He explains how electrochemistry—using electricity rather than heat to make products—has seen an 80% cost reduction in the last decade, bringing clean alternatives increasingly close to the point where they're simply cheaper than conventional options.

Schrep then pivots to energy demand, noting that carbon markets aren't the biggest opportunity: "Carbon markets is not the biggest market. Energy is the market. It's a two trillion dollar market." He points to emerging AI data centers requiring gigawatts of power and the growing electric vehicle market as indicators of rapidly increasing electricity demand.

He shares a crucial insight that most people don't realize: "Even in the US, we don't use most of our energy for electricity right now. Most of our energy—including stuff we dig out of the ground and burn—is used in industrial heat, in transportation, in heating in homes and others." Electrifying these sectors will require dramatically scaling up clean electricity production.

"We're probably going to 4x, 5x, 10x our electricity production in the next 20 years. We've been going on like a 1% growth or flat for the last decade. And then you're like, 'We can't do that.' China did it—7x in the last two decades. That's how much more power they're generating, and it's predominantly clean power."

Timestamp: [30:02-31:37]

Schrep shares a specific example of a Gigascale portfolio company developing next-generation biomass turbine technology. The team includes former SpaceX engineers and GE veterans who previously worked on large gas turbines.

Their innovation involves a high-pressure combustion system that takes wood waste (such as forestry thinnings from California wildfire prevention efforts) and oxygen as inputs, producing clean electricity, water, and pure CO2 that can be sequestered or utilized. This provides reliable "firm power" that's available 24/7, unlike intermittent renewable sources like solar and wind.

The company has already secured a significant deal with Microsoft for "5 megawatts of clean 24/7 firm power and 25,000 tons of sequestration." Schrep notes that this first plant will be followed by a second facility five times larger, and the business model is "very profitable."

He also mentions another portfolio company, Dioxide Materials (referred to as "Dioxycle"), which produces ethylene through electrochemical processes. Their first production line is also expected to be successful financially. Schrep explains that these companies are succeeding because they're "riding these cost curves down" where clean technologies become increasingly cost-competitive.

"They're a biomass next-generation biomass turbine, bunch of ex-SpaceX engineers including some folks from GE who worked on big gas turbines, and they're like, 'Hey, we have technology now that if we do a very high-pressure burn, you can put wood waste—think forestry thinning from California—oxygen in, you get clean power out, firm power, water, and pure CO2 that you can then sequester or do something else with.'"

Timestamp: [31:37-32:30]

Schrep summarizes his climate investment thesis, explaining how three converging forces create a compelling opportunity: "Market pressure plus technological innovation plus demand for electricity is causing this intersection that the more I dug into it, I was just like—even if you don't care about [climate]—I think this is a really good sector to be investing in."

He argues that the scale of the opportunity should be attractive even to investors focused purely on financial returns: "Wouldn't you want to be investing in a 12 trillion dollar re-engineering of the market? That sounds like a place where there's opportunities to make money."

Aditya responds positively, noting how encouraging it is to hear that climate solutions don't need to rely as heavily on problematic carbon offset and credit markets as previously thought.

Schrep offers to provide more specific examples beyond "big platitudes," emphasizing his preference for concrete case studies over abstract concepts. He clarifies that while he's enthusiastic about his portfolio companies, he's not intending this as a promotional pitch—he simply wants to illustrate diverse business models that work in climate tech.

"I do think that this Market pressure plus technological innovation plus demand for electricity is causing this intersection that the more I dug into it, I was just like—I mean, even if you don't care about [climate]—I think this is a really good sector to be investing in. Wouldn't you want to be investing in a 12 trillion dollar re-engineering of the market? That sounds like a place where there's opportunities to make money."

Timestamp: [32:30-33:18]

Schrep shares another climate tech business model that's succeeding through addressing everyday consumer pain points rather than explicitly marketing its environmental benefits. The company, Mill, was founded by the creator of Nest and produces a home food waste device that looks like a standard kitchen trash can.

The $1,000 appliance offers a compelling solution to the challenges of composting: "Here in California we've got green bins. Most of the United States doesn't have green bins. Even if you have one, most people don't use them. Why? It's like, 'I dump my eggs and my raw chicken somewhere in my kitchen and it freaking smells.'"

Mill's approach is simple but transformative—users put all food waste into the device, which dries it out and grinds it up, eliminating odors and reducing volume by 80%. After about a month of use, a typical family produces only a shoebox-sized container of dry, coffee ground-like material that can be placed in green bins, used in gardens, or shipped back to Mill with a prepaid USPS label. Mill then processes this material into chicken feed (with appropriate regulatory approvals).

Schrep notes the product has exceptional customer satisfaction: "People freaking love this thing. I've worked on a lot of consumer hardware products, and I can't tell you what the NPS is... but it's the highest NPS of anything I've seen." The environmental impact is substantial—Mill has already diverted 2 million pounds of food waste from landfills, preventing methane emissions (a potent greenhouse gas).

This exemplifies a climate tech business succeeding on pure consumer value: "It's massive climate impact on the back of people buying stuff because they're like, 'I'm tired of carrying messy, smelly trash out. I don't have to do that anymore.'"

"Mill's basic pitch is, 'Yeah, don't do any of that. You dump all this stuff into the Mill, it dries it out, grinds it up so it doesn't smell, it also gets 80% smaller. So you do this for like a month, and then after a month of a normal family's use, you have like a shoebox size of what looks like non-smelly coffee grounds.'"

Timestamp: [33:18-34:53]

Aditya asks a "spicy question" about whether the climate tech perspective suggests global conservation movements simply don't make sense: "Humans don't want to globally cooperate to do less of something." Instead, he frames the aspiration differently: "I want to use 10x more energy. How much more awesome would my life be if I could have access to low-cost 10x more energy? What are the things I could do?"

He notes that while per capita energy consumption in the U.S. has remained relatively flat, this statistic is misleading because it fails to account for the embedded energy in imported goods. He asks Schrep for the actual figure.

Schrep acknowledges he doesn't have the exact calculation but confirms Aditya's point about offshore production masking true energy consumption. He then emphasizes the fundamental connection between energy use and prosperity: "If you do a historical plot of energy use and GDP, it is—I mean, correlation isn't causation—but holy crap, it is the most correlated graph you can possibly make."

He frames energy access as addressing basic human needs: "The way to get people out of poverty is to get them energy. Can you read books at night? Can I cook things? Can I stay warm and cool when the weather is out there, or do I suffer and can't do anything?" From there, he extends to higher-level needs like AI-powered services that could benefit billions more people.

Schrep argues this abundance-focused approach is both possible and morally necessary: "We can do it. We've done it in the U.S., we've done it in China, we can do it throughout the world. That's the bogey we should be going after." The path forward involves deploying existing renewable technologies at scale while pursuing breakthrough technologies like fusion that could deliver "unlimited clean energy for everything."

"The way to get people out of poverty is to get them energy. Can you read books at night? Can I cook things? Can I stay warm and cool when the weather is out there, or do I suffer and can't do anything? These are really basic human needs, and that's the bottom of the Maslow triangle here."

Timestamp: [34:53-37:16]

Aditya builds on Schrep's energy abundance vision, expressing frustration with prevailing climate narratives that focus on limiting growth: "One thing that has frustrated me about a lot of the conversation about climate is this idea of like degrowth will fix it. We have too many people. Let's stop growing, let's not develop."

He argues that population reduction alone is an ineffective approach: "Even bringing down the population to 90% won't fix it." More importantly, he highlights the moral dimension of denying prosperity to developing nations: "Telling billions of people in India and China and elsewhere, 'No, you can't get raised out of poverty, you can't get to what we have'—this is the problem."

Aditya points out that people around the world aspire to basic comforts that wealthy nations take for granted: "We've got a whole lot of the world that looks at how we live and [thinks] 'I'd like to eat meat, I'd like to have air conditioning, I'd like to have a refrigerator.'"

Schrep emphatically agrees with Aditya's moral stance: "I think it is like immoral to say no. It is completely immoral." He reinforces that addressing climate change requires economic expansion rather than contraction: "We need to go spend tens of trillions of dollars. To your point, we need growth and development and all that."

"Telling billions of people in India and China and elsewhere, 'No, you can't get raised out of poverty, you can't get to what we have'—this is the problem. We've got a whole lot of the world that looks at how we live and [thinks] 'I'd like to eat meat, I'd like to have air conditioning, I'd like to have a refrigerator.' And I think it is immoral to say no. It is completely immoral."

Timestamp: [37:16-37:56]

Ryan shifts the conversation to request personal career advice, noting that he's spent the last three years in climate work and expects to continue in the field, but has struggled to find his optimal path.

He observes that Gigascale's portfolio focuses heavily on hard technologies rather than software solutions, mentioning Dioxide Materials, Mill, and Chement (working on cement and steel alternatives, which each account for 8-10% of global emissions). As a software engineering leader who doesn't plan to return to school for a climate science degree, Ryan has found it challenging to identify where he can make a meaningful impact.

He's focused on MRV (monitoring, reporting, verification) and carbon accounting because most companies lack accurate understanding of their emissions, especially in their supply chains: "Carbon accounting, like hugely important, there's great startups out there doing it. At the end of the day, it's compliance software. You've got to be a special person to wake up and get out of bed every morning and say 'I am psyched to go to work to build a reporting tool.'"

Ryan questions whether his assessment might be missing something important: "Maybe this is me just not being big-hearted enough, but how am I thinking wrong about this?"

This prompts Schrep to respond that he has "good news and bad news" about Ryan's situation, beginning with what he sees as the fundamental problem: "I think too many people start with the tool and..."

"One thing I noticed in your portfolio—Gigascale—it's a lot of hard tech and not necessarily software... I'm an engineer and engineering leader. I've learned a bit about climate. I'm not going back to school for undergrad to become a climate scientist, and I don't expect I need to. But when I've looked at what to do in climate, I've struggled... Carbon accounting is a dirty secret—most people have no idea what they're emitting, even your big companies, whether it's in their supply chain or direct."

Timestamp: [37:56-39:59]

• Electrochemistry (using electricity instead of heat for industrial processes) has seen 80% cost reduction in the past decade, making clean alternatives increasingly competitive
• Energy, not carbon markets, represents the largest opportunity ($2 trillion market) as AI data centers and EVs drive massive electricity demand growth
• Most energy consumption currently goes to industrial heat, transportation, and building heating—not electricity production—meaning electrification will require 4-10x more clean power
• Successful climate businesses are "riding cost curves down" to profitability, as demonstrated by next-generation biomass power securing Microsoft contracts
• Consumer-focused climate solutions like Mill succeed by solving everyday pain points (messy, smelly food waste) rather than leading with environmental benefits
• Energy consumption is fundamentally linked to prosperity, with energy use and GDP showing remarkably strong correlation throughout history
• The path forward requires clean energy abundance rather than conservation—enabling developing nations to access the same comforts enjoyed by wealthy countries
• Telling people in developing nations they can't have air conditioning, refrigeration, or meat consumption is described as "completely immoral"
• Climate tech investment represents a $12 trillion market opportunity to reengineer global energy systems
• Software engineers seeking climate careers may need to reconsider starting with their existing tools rather than focusing on the problems that need solving

Timestamp: [30:02-39:59]

People:

• Mike Schroepfer ("Schrep") - Guest, co-founder of Gigascale Capital, former CTO of Facebook/Meta
• Aditya Agarwal - Host, SPC
• Ryan Barrett - Host, SPC (asks about climate career options)

Companies & Organizations:

• Gigascale Capital - Climate tech investment fund co-founded by Schrep
• Microsoft - Mentioned as customer for biomass energy and carbon sequestration deal
• Mill - Company making home food waste processing devices
• Dioxide Materials (Dioxycle) - Company making ethylene through electrochemical processes
• Chement - Company working on cement and steel alternatives (referenced by Ryan)
• Nest - Previous company of Mill's founder
• SpaceX - Former employer of engineers at biomass turbine company
• GE - Former employer of engineers working on gas turbines who joined biomass company

Technologies & Concepts:

• Electrochemistry - Using electricity instead of heat (thermochemistry) for industrial processes
• Biomass turbine - Technology using wood waste and high-pressure combustion for clean power
• Carbon sequestration - Storing captured CO2 to prevent atmospheric release
• Firm power - 24/7 reliable electricity (contrasted with intermittent renewables)
• MRV (Monitoring, Reporting, Verification) - Tracking emissions and reduction efforts
• Carbon accounting - Measuring and reporting an organization's greenhouse gas emissions
• Degrowth - Economic strategy focused on reducing consumption and production
• Green bins - Municipal organic waste collection (common in California)
• NPS (Net Promoter Score) - Customer satisfaction metric mentioned regarding Mill

Statistics & Data:

• 80% cost reduction - Decrease in electrochemistry costs over the past decade
• $2 trillion market - Size of the global energy market
• 4x-10x increase - Projected growth in electricity demand over next 20 years
• 7x increase - China's growth in power generation over past two decades
• 5 megawatts - Size of biomass power plant deal with Microsoft
• 25,000 tons - Annual carbon sequestration in Microsoft deal
• 2 million pounds - Food waste diverted from landfills by Mill
• 80% reduction - Volume decrease when food waste is processed by Mill
• 8-10% each - Portion of global emissions from cement and steel production

Concepts:

• Maslow's hierarchy - Referenced regarding energy enabling basic human needs
• Energy-GDP correlation - Strong historical relationship between energy use and economic growth
• Embedded energy - Energy used in producing goods that are imported rather than produced domestically
• Green discount - Clean technologies that win by being cheaper than conventional alternatives

Timestamp: [30:02-39:59]

Responding to Ryan's question about software careers in climate tech, Schrep begins with what he sees as the fundamental issue: "I think too many people start with the tool and don't start with the problem. So you're like, 'I got a hammer, my hammer is software. What can I go hit?'"

He explains that while pure software solutions exist in climate tech, they're limited "because fundamentally we're talking about atoms, we're talking about electrons, we're talking about moving stuff, we're talking about growing food." Software can be a significant component, but Schrep advises Ryan to "open your aperture a little bit" beyond traditional software approaches.

To illustrate his point, Schrep shares the journey of Selina Tobaccowala, whom he describes as "a multi-time founder" and "classic Silicon Valley consumer product person" with experience at companies like Evite, a healthcare startup, and SurveyMonkey. Despite her background in digital products, she found her climate opportunity by focusing on home energy efficiency—a major problem where consumers could save money within a year but lacked knowledge about what actions to take.

Rather than trying to force a pure software solution, Tobaccowala's company developed a hybrid approach: a home kit with a thermal camera and other hardware components, paired with an app experience that guides users through their home assessment and recommends specific improvements like weather stripping, insulation, or appliance replacements. It then connects them with contractors who can implement these changes.

"Home energy efficiency is a huge thing. There's a lot of ways that you can help a consumer save money on their home that pays back within a year, so it financially makes sense for them to do this. They just don't know what it is."

Timestamp: [40:06-42:02]

Schrep shares a second example of problem-first thinking that led to a successful climate tech software company. He explains that Arch—"the only pure software company we've backed to date"—emerged from exploring a fundamental question about heat pumps: "Heat pumps are this magic technology... Why the hell isn't everyone in the world running around and installing them?"

He first explains how heat pumps work, moving 3-6 times more heating or cooling energy than the electricity they consume: "It's not turning electricity into heat or cooling, it's literally moving energy from one place to another. The transport cost of pumping the heat from outside to inside is less than the amount of heat I'm moving." This efficiency makes them an ideal climate solution, yet adoption has been slow.

To understand this market gap, the Arch founder took an unconventional approach to research: "I would call HVAC contractors and be like, 'Can I come with you to an installation?' The first three said no, like, 'What, are you some weirdo or something?' And then the fourth one's like, 'Yeah, come along.' I literally rode along in the truck for days to just go and be like, 'What is this person's business like?'"

This immersive research revealed that HVAC contractors were constrained by inefficient processes. The company built B2B software enabling contractors to "build and bid four systems rather than three a day," significantly improving their productivity and profitability while accelerating heat pump adoption.

Rather than beginning with a software solution looking for a problem, the founder "started off the back of a deep understanding of the problem. It wasn't like 'I want to do something in software, let me try heating and cooling.' It's like, 'Let me understand this problem.' And then, 'Oh, we can build some software for this.'"

"I just did a bunch of ride-alongs... I literally rode along in the truck for days to just go and be like, 'What is this person's business like?' And what are they doing? And just did this again and again and again and started to figure out—huh, it's a bunch of things they're doing that if we built some software for, literally they could build and bid four systems rather than three a day, and that would make their business and make us a bunch of money too."

Timestamp: [42:02-44:34]

Aditya interjects with a lighthearted comment about Ryan not needing to go back to school since he's getting his education at SPC, saying "Street smarts!"

Ryan acknowledges the value of Schrep's advice, admitting to a common cognitive trap: "It's great advice. I've been looking for my keys a bit under the streetlight because that's the software I have, and I should expand it out a bit." This refers to the "streetlight effect"—the tendency to search where it's easiest to look rather than where the solution might actually be found.

The metaphor resonates perfectly with Schrep's guidance to start with the problem rather than the tool, suggesting that Ryan has been limiting his climate impact by focusing only on pure software solutions when hybrid approaches or deeper problem understanding might lead to more meaningful opportunities.

"I've been looking for my keys a bit under the streetlight because that's the software I have, and I should expand it out a bit."

Timestamp: [44:34-45:00]

The conversation transitions to audience questions, with an attendee named Sam asking about the capital-intensive nature of climate tech: "Climate plays can take a long time and be really capital intensive. So as you've gotten into investing, how do you manage the risk/reward of that? Because it's very different from the software investing here where it's like you get $400K and you improve traction and then you get your seed."

Sam extends the question to founders as well: "Even starting a climate tech company—how do founders think about it, and how do you think about it?"

Schrep acknowledges the challenge but begins by putting it in historical context: "I've been in Silicon Valley since the '90s, and it's hilarious because Silicon Valley—it's in our name—it started in hardware. We were chips, and then we were hard drives, and then we moved into software." He points out that one of the world's most valuable companies, Nvidia, is a hardware company, demonstrating that "you can build amazingly big hardware businesses."

This historical perspective serves as a reminder that while software has dominated recent startup culture, the Valley has deep roots in capital-intensive hardware businesses, suggesting that the climate tech approach isn't as radical a departure as it might seem to today's founders.

"It's hilarious because Silicon Valley—it's in our name—it started in hardware. Like, we were chips, and then we were hard drives, and like, then we moved into software... If you look at one of the most valuable companies in the world right now, it's a hardware company. It's Nvidia. So you can build amazingly big hardware businesses, and we've had a history of that."

Timestamp: [45:00-45:44]

Schrep shares insights from his hardware experience at Meta, where he learned the crucial differences between various types of hardware development. When he joined Meta (then Facebook), it was purely a web company. Their first hardware venture was building data centers, with the initial one coming online in 2011. Within a decade, they had "10 million square feet under operation in 12 regions around the world."

This enterprise hardware experience gave him confidence, but he faced a steep learning curve when Meta entered consumer hardware with Oculus VR headsets: "When you ship a box to someone and then they return it because they don't like it, it's like a whole different ball game than building buildings that you own yourself." This required him to master both enterprise and consumer hardware approaches.

From these experiences, Schrep identifies several key principles for hardware startups:

1. Planning is more important: "Pivots are really hard, and there are some just fixed time frames when you're building hardware... If I need to do a major design spin, you're probably talking six to eight weeks, and it's really hard to compress." The consequence is that an additional design iteration can add months to development, making the "measure twice, cut once" approach essential.

2. Leverage modeling and metrics: Unlike software, hardware performance can often be modeled beforehand—"You can kind of do some basic napkin math on like, 'Okay, if I get everything right, how good is the system going to perform before I built a single thing.'"

3. Fundraising becomes more central: Hardware founders must be strategic about raising capital and demonstrating progress. Rather than showing up with "three parts on the ground that kind of work," successful founders present clear roadmaps: "Here's the endgame we're getting to... In order to get there, I have to do six things right and in this order: 1, 2, 3, 4, 5, 6. And then I'm going to come back to you on the following dates."

"Planning is more important. Pivots are really hard, and there are some just fixed time frames when you're building hardware stuff. If I need to do a major design spin, you're probably talking six to eight weeks, and it's really hard to compress. So if I do three design spins versus four, you know, that's an extra two months."

Timestamp: [45:44-48:02]

Schrep highlights unique funding advantages available to climate tech startups that can offset some of the capital intensity challenges:

First, he emphasizes the substantial non-dilutive financing available: "There is a lot of non-dilutive financing. There's a lot of government loans, government grants, there's a lot of ways." He shares impressive examples from his portfolio: "One of our companies got a $150 million grant to help build their factory in West Virginia. That's massive. And they also got a $150 million grant to build a massive battery in Maine."

These grants don't require giving up equity, preserving ownership for founders and early investors while providing the substantial capital needed for hardware development.

Schrep then addresses dilution concerns directly, suggesting that founders should be pragmatic rather than overly protective of valuation: "Because you're raising a lot of money, dilution is an issue. To be honest, it's going to sound self-serving as an investor, but some people try too hard to get precious about it, really try to jack up valuations, and then you get into a point where you can't raise the money you need and the company's dead."

Instead, he recommends accepting "slow bump-ups as we hit these milestones," recognizing that the path to success may involve more funding rounds than a typical software startup.

Finally, he emphasizes the substantial advantage hardware companies gain once they achieve scale: "The good news at the end of all of this is—once you get escape velocity as a hardware company, good luck catching it." He points to Nvidia as an example, where their market lead allows them to "take all the cash, plow it into the next R&D," creating a virtuous cycle that's difficult for competitors to overcome.

"Once you get escape velocity as a hardware company, good luck catching it. Because if you're out on the market like, 'I've built this thing, I'm at a 500,000 unit scale,' and some new set of entrepreneurs come, 'We're going to build that, but we're two years behind...' Not going to work. Look at everyone chasing Nvidia and how much money they're making. We've got this massive lead, they take all the cash, they plow it into the next R&D. It is a really exciting exit if you can get there."

Timestamp: [48:02-49:21]

Aditya notes a tension between Schrep's earlier emphasis on rapid iteration and his current advice about careful planning for hardware ventures: "Those funding notes seem very real, and the planning part seems very real too, and also a little disappointing when we've talked about how iterating is so useful and really 'Let's just go iterate fast and learn and orient as we go.' But for hardware, it's just the necessity."

This observation highlights an important paradox in climate tech: while rapid iteration is ideal for innovation, the physical and financial constraints of hardware development often require more methodical approaches. The faster learning cycles that drive software development must be balanced against the longer timelines and higher costs of each hardware iteration.

The comment serves as an important bridge between the earlier conversation about Facebook's culture of speed and the current discussion of climate tech's capital intensity, acknowledging the need to adapt principles rather than directly transfer them across contexts.

"Those funding notes seem very real, and the planning part seems very real too, and also a little disappointing when we've talked about how iterating is so useful and really 'Let's just go iterate fast and learn and orient as we go.' But for hardware, it's just the necessity."

Timestamp: [49:21-49:40]

The final audience question asks Schrep about his personal transition from being a builder throughout his career to becoming an investor: "You've been a builder almost your entire career. Can you elaborate on your experience pivoting into investing?"

Schrep responds with candid vulnerability: "It's honestly been a little bit of a struggle, if I'm just being honest. If it's like being real about me as a person and a human, it's really fun to build stuff and to have a team, and to be like, 'I went into the office today and I went home, and the following things got done.'"

This brief but revealing answer provides insight into the emotional adjustment required when moving from operational roles to investment positions. Despite his success and impact as an investor, Schrep acknowledges the personal satisfaction that comes from direct building and the tangible daily progress that operational roles provide.

The conversation appears to be continuing beyond this segment, as Schrep is just beginning his answer when the transcript ends.

"It's honestly been a little bit of a struggle, if I'm just being honest. If it's like being real about me as a person and a human, it's really fun to build stuff and to have a team, and to be like, 'I went into the office today and I went home, and the following things got done.'"

Timestamp: [49:40-50:00]

• Effective climate tech innovation starts by understanding problems deeply rather than beginning with existing tools or expertise
• Software engineers can find meaningful climate roles by expanding beyond pure software to hybrid solutions that address physical-world challenges
• Heat pumps exemplify promising climate technologies with efficiency advantages (moving 3-6x more energy than they consume) that face adoption barriers solvable through software
• Immersive research like "ride-alongs" with industry professionals can reveal overlooked business opportunities that accelerate climate solutions
• Silicon Valley began with hardware (chips, hard drives) before software became dominant, showing capital-intensive businesses can succeed at massive scale
• Hardware development requires more careful planning than software, as design iterations typically take 6-8 weeks and are difficult to compress
• Climate tech benefits from substantial non-dilutive funding sources like government grants ($150M examples cited) that aren't available to most software startups
• Hardware companies that achieve scale gain significant competitive advantages that are difficult for followers to overcome
• Founders should be practical about valuation rather than "precious," accepting more dilution to ensure adequate funding for capital-intensive climate ventures
• The transition from building to investing can be emotionally challenging, as operators often miss the tangible daily progress and team dynamics of operational roles

Timestamp: [40:06-50:00]

People:

• Mike Schroepfer ("Schrep") - Guest, co-founder of Gigascale Capital, former CTO of Facebook/Meta
• Aditya Agarwal - Host, SPC
• Ryan Barrett - Host, SPC (asks about climate career options)
• Selina Tobaccowala - Multi-time founder mentioned as example of product person entering climate tech
• Sam - Audience member who asked about capital intensity in climate tech

Companies & Organizations:

• Gigascale Capital - Climate tech investment fund co-founded by Schrep
• Meta/Facebook - Referenced regarding Schrep's hardware experience with data centers and VR
• Arch - The "only pure software company" in Gigascale's portfolio, focusing on HVAC contractor software
• Evite - Company founded by Selina Tobaccowala
• SurveyMonkey - Company where Selina Tobaccowala previously worked
• Nvidia - Cited as example of successful hardware company with strong market position
• Oculus - Consumer hardware company acquired by Facebook that taught Schrep about consumer hardware challenges
• SPC - Organization hosting the conversation

Technologies & Concepts:

• Heat pumps - Technology that moves 3-6x more heating/cooling energy than electricity consumed
• Home energy efficiency kit - Product with thermal camera and app for home assessment
• Enterprise vs. consumer hardware - Distinction between building data centers versus consumer products
• Design spin - Hardware iteration cycle that typically takes 6-8 weeks
• Non-dilutive financing - Government grants and loans that don't require giving up equity
• Escape velocity - Point where hardware companies achieve scale advantages difficult to overcome
• Streetlight effect - Cognitive bias of looking where it's easiest rather than where solution might be

Statistics & Data:

• 6-8 weeks - Typical timeframe for a hardware design iteration
• 10 million square feet - Scale of Meta's data center operations after 10 years
• 12 regions - Global spread of Meta's data centers
• $150 million grant - Size of government grant for factory in West Virginia
• $150 million grant - Size of government grant for battery project in Maine
• 3-6x efficiency - Energy multiplication factor achieved by heat pumps

Concepts:

• Problem-first approach - Starting with deep understanding of problems rather than available tools
• Ride-along research - Immersive method of understanding industry practices by accompanying practitioners
• Capital intensity - Challenge of hardware businesses requiring more funding than software startups
• Dilution management - Balancing ownership preservation against capital needs
• Builder vs. investor mindset - Different satisfaction drivers in operational versus investment roles

Timestamp: [40:06-50:00]

Continuing his reflection on transitioning from builder to investor, Schrep explains the emotional satisfaction he derived from operational roles: "This week or this month, look at this thing I made in this world. It shows up, we're on stage talking about you're using my—that freaking feels great. That has fueled me for 20-plus years."

By contrast, investing requires accepting a more indirect role: "When you're an investor, you're kind of one step removed. You're like, 'I backed this company, they built this big thing, I helped them with a decision, I helped them get money.'" This requires "eating more humble pie" and maintaining "more humility about your role in the system."

Schrep criticizes investors who overstate their contributions: "Some investors run around like, 'I did X.' You didn't do that. You helped them, you gave them money. You're important, but let's not overshadow the entrepreneurs."

Beyond the emotional adjustment, Schrep identifies two additional challenges. First, the work style differs significantly: "As an operator, you're kind of in the office every day working on a small number of things. As an investor, I've got to be working on 30 different things at once."

The most difficult adjustment, however, involves a fundamental shift in success metrics: "As an operator, you get rewards for being right. If my success rate is 5% when I'm like, 'I'm going to do X, Y, and Z,' I am not going to last long as an operator. So you spend a lot of time trying to be 90% correct."

Investing completely inverts this mindset: "Your job isn't to correctly miss out on the opportunities that are going to fail. It's like, 'Is this the one 10x company that's going to happen or not?' You've got to identify these unicorns because it's all power law distributions."

"As an operator, you get rewards for being right... For investing, it's the exact inverse. Your job isn't to correctly miss out on the opportunities that are going to fail. It's like, 'Is this the one 10x company that's going to happen or not?' And so you've got to identify these unicorns because it's all power law distributions."

Timestamp: [50:02-52:14]

Aditya asks whether Schrep believes certain professional backgrounds (product people versus engineers) might be better suited for the transition to investing, given the challenges he described.

Schrep responds with striking candor about the difficulty of predicting investor success: "Having now tried to hire a bunch of other investors to help me, it's a really hard profession to predict performance based on prior work, to be really honest. I don't believe anyone who makes any statements that the following experiences or attributes correlate with good investing performance."

He describes investing as "making important decisions on limited information," comparing it to having a certain "taste" in identifying opportunities: "Everyone looked at the same data, and you're like, 'Yeah, I think they're going to make it, so we're going to go for it.'" This human judgment element remains crucial even in the age of AI: "If ChatGPT can say 'yes' or 'no,' then we don't need humans. At the end of the day, you have humans who are like, 'I think this team's really good, and I know a lot of people are betting against them, but I'm going to go for it.'"

Despite his earlier skepticism about correlations, Schrep does draw one parallel: "The closest analog is actually product people. If you've ever worked with someone who's a really good product person, they just have taste... They'll bring you a product, you look at it, you're like, 'That's great!' [You ask] 'What did you do?' And they won't be able to describe a thing. They're just like, 'I don't know, I just thought maybe if we did it this way, it'd be better.'"

This contrasts with people who build "perfectly acceptable products that bring me no joy." Similarly in investing, the goal is finding "outlier people" who can say, "I know everyone else thinks this is dumb, but I think this might be it."

"If you've ever worked with someone who's a really good product person, they just have taste... They'll bring you a product, you look at it, you're like, 'That's great!' [You ask] 'What did you do?' And they won't be able to describe a thing. They're just like, 'I don't know, I just thought maybe if we did it this way, it'd be better.'"

Timestamp: [52:14-54:00]

Aditya reinforces Schrep's point about product intuition by sharing his own experience: "One of the quintessential product people I've worked with like that was Marissa Mayer, coming from Stanford and Google. Working with her... she couldn't always explain why something was a good product decision or bad, but she always knew and always had the conviction, and she was pretty much always right."

This anecdote validates Schrep's comparison between exceptional product people and successful investors—both rely on an intuitive sense that can't always be articulated but consistently leads to correct decisions. It highlights how certain individuals possess an almost ineffable talent for recognizing potential before it becomes obvious to others.

The exchange illustrates how in both product development and investing, analytical frameworks and methodical approaches are valuable but sometimes insufficient. The ability to make inspired leaps beyond what the data explicitly shows—and to maintain conviction in those insights even when they can't be fully explained—often distinguishes exceptional performers.

"One of the quintessential product people I've worked with like that was Marissa Mayer, coming from Stanford and Google. Working with her... she couldn't always explain why something was a good product decision or bad, but she always knew and always had the conviction, and she was pretty much always right."

Timestamp: [54:00-54:19]

An audience member asks Schrep about Meta's long-term commitment to AR/VR technology, noting they'd seen glasses on his desk at Facebook about seven years ago: "What made you continue working on that even though there wasn't like a direct signal or proof early that it was going to work?"

Schrep immediately credits Mark Zuckerberg for this perseverance: "I really have to give the credit to Mark Zuckerberg on this." He reveals that he still spends one day a week advising Meta because "it is still one of the best places in the world to incubate very risky long-term technology that has massive impact, whether that be AI or AR/VR or wearables."

He explains that every major technological transition faces substantial skepticism before breakthrough: "My experience in every new product category, including—we built an AI research lab in 2013, and I can't tell you how many meetings I was in where people [asked], 'Do we really need to have all these AI researchers? Is this stuff really going to matter?' You know, and then 10x that on wearables and VR."

The consistent pattern he's observed is that "every new technology that I've seen is like, up until it's obviously awesome to you personally, everyone's like, 'I don't know.'" This creates a critical need for leadership with unwavering commitment: "In that world, you need to have someone who has belief and vision and does so consistently, because you can't fund it one year and then cut the whole team... That doesn't work."

Schrep expresses profound admiration for Zuckerberg's steadfastness in the face of widespread doubt: "What Mark has done, which is just absolutely effing incredible, is when everyone—the markets, employees—are like, 'Why are you doing this?' Mark is just like, 'We're going to keep going.' And then when he's right at the end, everyone's like, 'Oh yeah, he's a genius, and I thought it was good all along.'"

He concludes by acknowledging the rarity of such conviction: "That is a very unique skill. Not a lot of people have the—honestly—fortitude to stand up to a whole bunch of people being like, 'This is dumb. You're wasting money.' So that's a skill and a trait I admire in Mark."

"Every new technology that I've seen is like, up until it's obviously awesome to you personally, everyone's like, 'I don't know.' And so in that world, you need to have someone who has belief and vision and does so consistently, because you can't fund it one year and then cut the whole team... What Mark has done, which is just absolutely effing incredible, is when everyone—the markets, employees—are like, 'Why are you doing this?' Mark is just like, 'We're going to keep going.'"

Timestamp: [54:19-56:05]

As the session reaches its time limit, Aditya acknowledges there are many more questions from the audience that they can't address. He expresses gratitude to Schrep for participating and suggests the possibility of a follow-up session: "Thank you again, Schrep, for coming here. This is super fun. Talk about any early war stories... we can do a part two at some point."

Schrep agrees, and Aditya extends the thanks to the audience "for coming and kind of getting a bunch of wisdom dropped from Schrep." The session concludes with mutual appreciation from all participants.

This ending reflects the enthusiasm generated by the conversation and indicates strong interest in hearing more from Schrep about his experiences in both his Meta/Facebook years and his current climate tech work. The reference to "early war stories" suggests there remain many valuable insights from Schrep's career that could benefit the audience in a future session.

"Thank you again, Schrep, for coming here. This is super fun. Talk about any early war stories... we can do a part two at some point. We would love to have you back anytime. But thank you all also for coming and kind of getting a bunch of wisdom dropped from Schrep."

Timestamp: [56:05-56:37]

• Transitioning from operator to investor requires accepting a more indirect role and less immediate gratification from visible accomplishments
• Successful operators aim for high accuracy (90% correct decisions), while successful investors must identify rare unicorns among many opportunities that won't succeed
• Prior professional experience is a poor predictor of investor performance, making it difficult to identify who will excel in venture capital
• The "taste" that distinguishes exceptional product people (like Marissa Mayer) resembles the intuition needed by successful investors—an ability to recognize potential that others miss
• Both great product people and investors often can't fully articulate why they believe in certain opportunities, yet maintain strong conviction that proves correct
• Every transformative technology (AI, AR/VR, wearables) faces significant skepticism until it becomes "obviously awesome" to end users
• Long-term technology development requires leadership with unwavering vision and the fortitude to withstand widespread doubt
• Mark Zuckerberg's willingness to persist with AR/VR despite criticism exemplifies the rare leadership quality needed for breakthrough innovation
• Meta continues to be "one of the best places in the world to incubate very risky long-term technology" according to Schrep, who still advises them weekly
• Consistent support is essential for developing transformative technologies—"you can't fund it one year and then cut the whole team"

Timestamp: [50:02-56:37]

People:

• Mike Schroepfer ("Schrep") - Guest, co-founder of Gigascale Capital, former CTO of Facebook/Meta
• Aditya Agarwal - Host, SPC
• Mark Zuckerberg - CEO of Meta, credited by Schrep for his long-term vision on AR/VR
• Marissa Mayer - Former Google executive with strong product intuition, cited by Aditya

Companies & Organizations:

• Meta/Facebook - Where Schrep served as CTO and still advises one day per week
• Gigascale Capital - Climate tech investment fund co-founded by Schrep
• Google - Former employer of Marissa Mayer
• Stanford - University mentioned in connection with Marissa Mayer
• SPC - Organization hosting the conversation

Technologies & Concepts:

• AR/VR - Augmented reality and virtual reality technologies Meta has pursued long-term
• AI research lab - Established at Facebook in 2013 despite skepticism
• Wearables - Technology category Meta has invested in (including glasses seen on Schrep's desk)
• ChatGPT - Referenced when discussing human judgment versus algorithmic decision-making in investing

Concepts:

• Operator vs. investor mindset - Different approaches to decision-making and success measurement
• Power law distributions - Investment returns where a tiny percentage of companies generate most returns
• Product taste - Intuitive ability to make good product decisions without explicit reasoning
• Long-term vision - Commitment to transformative technologies despite lack of immediate validation
• Success rate expectations - 90% for operators versus seeking rare unicorns as investors
• Investment conviction - Believing in opportunities despite widespread skepticism

Statistics & Data:

• 20+ years - Length of Schrep's career as an operator before investing
• 5% vs. 90% - Contrast between unacceptable and expected success rates for operators
• 30 different things - Scale of parallel work an investor must manage compared to an operator's focus
• 35,000 people - Size of the organization Schrep led at Meta by 2022
• 2013 - Year Facebook established its AI research lab
• 7 years ago - When audience member noticed AR glasses on Schrep's desk at Facebook
• One day a week - Time Schrep still spends advising Meta

Timestamp: [50:02-56:37]