It's Not About Energy, It's About Power: Matt Eggers (Prelude) on China, the Grid, and the AI Boom

Matt Eggers is a managing director at Prelude Ventures, a climate fund with more than two billion dollars under management. Prelude backs the companies building the energy and infrastructure behind the AI era, from new power for data centers to first-of-a-kind manufacturing plants. Portfolio names include LuxWall, Terabase, Augmenta, and Graphite.

Before he invested, Matt operated. He worked at Sunrun, Tesla, and Bloom Energy, and he grew up on a farm in central Iowa. That background comes through when he talks. He thinks like a finance guy, but he also talks like someone who has gone out in a blizzard to thaw a frozen cattle waterer because the alternative was a dead herd.

The hard part of reindustrialization is not the software. It is atoms, power, permits, and time. Matt gets into why AI's real bottleneck is power and not energy, why China is out-building the US on solar and EVs, how you finance a plant no bank will lend against yet, and what he looks for in a founder before committing for a decade.

The conversation

Jay: How has being a climate VC changed in the last year?

Matt: It has changed. But we have seen this before. This is not the worst downturn I have seen in climate tech. That was clearly 2008 through 2012. These are massive industries now. EVs are a giant industry. Solar is a giant industry. Wind is a giant industry. They do not just go away because somebody in the White House does not like them. There are bright sides and dark sides. Trump took away a lot of the IRA, though we still have more government support than before the IRA. And there is a pall cast by the zeitgeist, less support for solar, and, oddly, less permitting support in certain places. Take the Grain Belt Express, a huge transmission line, fully permitted, carrying very cheap wind from Kansas to the Midwest. The Trump administration, out of the blue, killed it.

Jay: How does that change what you think is investable?

Matt: If a company is getting incentives, or special permitting, or tax credits, there is now added risk, because people do not feel they can rely on the word of the government. Some things are better. The government is pouring money into certain aspects of power, and into critical minerals. But businesses like predictability, and predictability is down. That was the toughest thing about the tariffs. He started with a number, then a pause, then it came back, then it is on copper, now it is not on copper. Think about an entrepreneur trying to buy feedstock, cost it out, and price it for customers. It is so hard to plan.

Jay: Where have you seen AI demand intersect with energy and climate?

Matt: The power sector is booming in a way it has not in my lifetime. AI is a massive driver. The amount of power being consumed and asked for by AI is staggering compared to power growth over the last 30 or 40 or 50 years.

Jay: Is that a problem startups are equipped to solve, or are the grid and utilities so archaic that a longtime VC would say avoid?

Matt: It is the best of times and the worst of times. Geothermal is doing really well right now. Prelude is an investor in one called Quaise. That is a great place to be, a 24/7 source of clean power. One thing people who are a little close to power but not real close misunderstand: right now it is not about energy. It is about power.

Jay: Explain the distinction.

Matt: Energy is measured in a kilowatt hour. It is power over time. Power is the instantaneous need for energy. We have a lot of energy and not enough power, and the shortfall hits at certain times of day. A data center is both good and bad for the grid because it runs 24/7. At night it actually makes energy cheaper for everybody, because you are running more through a system you already paid for. It only hurts during the day. Maybe 20, 30, 40 hours a year are the really hard ones, and that is when prices go crazy high. We have a portfolio company, Voltus, in demand response. Its business is booming, because big businesses are willing to scale back their power use at certain times and make a lot of money doing it.

Jay: Nuclear has become the darling. I am still a believer in solar and wind. Am I a hopeless optimist?

Matt: In the first half of 2025, China built 285 gigawatts of solar. Just this year, in half a year. They are on pace for more than 500 gigawatts in 2025. Trump likes to call these incentives the green new scam, and I keep scratching my head. Is the chairman of China in the pocket of the tree huggers? I do not think so. Solar is the cheapest and fastest, emphasis on fastest, because the hyperscalers are in a massive race to get power on the grid. A nuclear power plant is about a gigawatt. So in half a year, China added the equivalent of 285 nuclear plants. Batteries can extend that capacity four hours, eight hours. They cannot extend it 24 hours yet, but there are other ways to provide peaking power. Solar is still the fastest and cheapest way to add energy.

Jay: The detractors say the sun does not shine at night, and data centers need a reliable source.

Matt: Even if AI demand triples, it will still be a small proportion of the grid. And the rest of the grid is not 24/7 either. We have decades of watching power move like this. Here is the other thing China probably has in mind. When you give people a lot of a cheap, valuable resource, they figure out how to use it. When you build 500-plus nuclear plants worth of daytime energy a year, that is a lot of cheap power. It will advance batteries fast. When we have all that power, we will figure out how to use it.

Jay: Where do investors overreact on China, and where are we underestimating them?

Matt: We are underestimating them in almost every way. I was shocked when they released DeepSeek. I was shocked when I saw 285 gigawatts. They are flooding the world with cheap, very well-built EVs. The BYD, I was in Europe over the summer, that is a great car. It is better than a Tesla. I hate to put that on record, but it is a really well-built car. Maybe space is a sector where we are a bit ahead. But I would not bet against them on anything right now. As an investor, though, it is not a risk we really think about. I would not encourage a seed-stage founder to worry too much about China. Make a product people need in a market big enough to support your company, and you will do great.

Jay: What technologies are best situated to close the gap between power needed and power created?

Matt: Probably Form Energy. Form makes large, stationary, grid-connected batteries that act like a peaker plant. They shift 100 hours of power at a time. These compete with gas peaker plants, so they might only run hundreds to a thousand hours a year. In this power crunch they are incredibly valuable. It is easier to smooth power out than to build new stuff, especially in this country where it is still hard to build. It takes five years to get a new combined-cycle gas turbine. There is a five-year waitlist right now. Copper is really interesting because it addresses aging parts of the grid, in old northeastern cities where the gas and electrical infrastructure in hundred-year-old buildings is in bad shape. Someone smells gas, ConEd shows up and shuts off the gas to the building. Now someone is sitting in a five-million-dollar condo with a gas range and no gas. The building's options are to replace the entire gas infrastructure, expensive and invasive, or install induction stoves, which normally means replacing the entire electrical infrastructure. Copper comes in with a battery and smooths out that high power demand from the stove. You charge it overnight and it covers the peak.

Jay: How did you think about the risk of convincing 70-year-olds on the Upper East Side to switch from gas to induction?

Matt: Have you tried an induction stove? They are great products. And you go to that 70-year-old, you turn off their gas because it is leaking, and you tell them it is 25,000 dollars per condo to fix the gas. They say, okay, I will take the induction. It is meeting them at the point of need. Same with LuxWall, which makes extraordinarily efficient windows. They are doing an install in a big multifamily building around 86th and Broadway. Someone visited a unit that had it, was blown away by how quiet it was, and now demands it in their own unit. That is how it spreads.

Jay: Can there be software-only VCs in climate, or is that a setup to fail?

Matt: There definitely can be. Probably 20 to 25 percent of our companies are software. To a lot of people it is shocking how many hardware companies there are, but software can do a lot of good in climate too.

Jay: You have been vocal about splitting equity financing and capital project financing. Explain that difference.

Matt: Debt is much cheaper on a return basis than equity, but debt holders cannot afford to lose their money. So companies need to tackle the hard problems first, reduce risk so the valuation keeps rising, and plan the right steps from pilot to first plant to major commercial plant. Typically you cannot get project finance on a first-of-a-kind plant. That is where the real challenge is.

Jay: Why has there not been a first-of-a-kind wave of funding when the need is so clear?

Matt: I think there is, it just depends on the sector. Some places where it has been painfully obvious there is not are cement, and liquid fuels in the US. The products you compete against are so perfected and so cheap that it is difficult even on the first plant. But there are segments where investors are happy to take that risk. Form Energy, investors have spent nearing a billion building that factory pre-revenue, because they believe in the market.

Matt: Mining is a new one right now. Prelude invested in one eight or so years ago called MineSense. They took an XRF sensor that used to sit in a lab and figured out how to put it on a massive copper mining shovel. In a second or two it tells the operator how much copper is in that scoop. Before, they did not really know. You used to process a lot of waste and dump a lot of good copper. With this you do not. People suddenly discovered copper is one of the largest markets on earth, and with data centers, EVs, and electrification, it all runs on copper. Demand is expected to double.

Jay: In concrete, how easy is it to get someone to swap a mix they have used for 40 years?

Matt: In the world of atoms, physics and time and engineering risk really matter. The value has to accrue to the customer, and it has to come in a relatively quick time frame. I worry about a cement improvement that makes it last 10 or 20 years longer. Great for civilization, horrible for a customer who does not care. The sound benefit of LuxWall is one I was excited about at seed, because people walk in and like it immediately. The value is in the heating bill, but the time to value comes from the sound.

Jay: If a founder brings you an infrastructure-heavy, capital-intensive model, what do you stress-test?

Matt: Margins is a big one. It surprises people that a software company can raise 400 or 500 million just to scale sales and marketing. It is 85 percent gross margin and it recurs. Not much in hard tech is 85 percent, but they can be good, 50 to 60 percent, and they can have a high operating margin because they do not need much sales and marketing. Ultimately public investors pay for margin and growth. Total market size feeds growth. If you are already 40 percent of a market, you will not grow fast.

Jay: How do you feel about getting climate companies funded when AI is pulling so much capital?

Matt: You are going to talk to a Series A investor who earlier that day spoke to someone who went zero to 20 million in ARR, and you are great, you hit 5 million, but it took you three years. It feels like there need to be two entirely different kinds of investors. But AI needs power, it needs critical materials. It is surprising how many of our companies have a side thing the Department of Defense is asking them to do. A lot of things in hard tech have other uses.

Jay: Do you feel a Sputnik moment in climate is still coming?

Matt: I think it is already happening. There are knock-on effects of AI, like rising power prices. When power prices go up, other technologies look a lot more valuable. LuxWall windows, the ROI keeps climbing as power prices rise. You do have to play the game on the field. We also think about how to use AI inside a company that is not an AI company, to compress development time. One of my companies, Vaulted Deep, is a carbon-removal and waste-management company. It used to take them five or six months to decide and permit a site. Now it takes six weeks, because they made it programmatic with AI tools. If you are a founder today and you are not using these tools to go faster, you are getting left behind.

Matt: I have encouraged my founders that in interviews with everyone they hire, a question should be: how are you using AI to make yourself better at your job? If the person does not have a good answer, that is a big issue.

Jay: What is a recent non-consensus investment most folks passed on?

Matt: I made two carbon-removal investments last year, Vaulted and Graphite. Especially in the Trump era, that is a market people are not as excited about. So why did I make two? These companies store carbon with biogenic sources, so-called BiCRS, biomass-based storage. They take waste biomass, remove a waste problem, and generate carbon storage. Their costs look extremely cheap to me, without any massive technology leap. And they are really scalable. These have projects that cost 5, 10, 15 million. Bite-sized. Much easier to get that debt financing. You do not have to get everything to line up to build one massive project where you spend a billion dollars.

Jay: You have described your ideal founder as a calm maniac. What does that mean?

Matt: The maniacal piece is that people have to love what they are doing and be willing to go so hard for so long, through so many ups and downs. But they also have to be calm. A great people person. They have to roll with the punches and inspire people through all of it. How do I test for it? I really try to get to know these people as humans. I actually like it when a difficult conversation comes up during diligence, because I get to see how they act with me when things are hard.

Jay: Is there anything earlier in your career that resonates now?

Matt: I grew up on a farm in central Iowa. Part of my appreciation for the founder journey comes from moments like a raging blizzard when a cattle waterer freezes and the cattle will die without water. It does not matter that it is 35 below with 40-mile-an-hour wind. You go fix the waterer. And it does not matter that you do not know how to fix it right now. You figure it out. That is a lot like the founder's journey. My father used to say, if you are going to do something halfway, do not even bother starting.

Jay: Ten years from now, what consensus view in climate investing will be proven wrong?

Matt: One is carbon removal. I think the world will have to do a lot of it, and we will come around. A lot of things will force it, rising seas, rising temperatures, and the massive insurance crisis, which I think is only beginning in the US. People need to be able to insure their homes or they cannot live there. Our economy falls apart if real estate is uninsurable. That is the thing that scares me the most.

Jay: What climate theme is deeply under-discussed today that will look obvious in hindsight?

Matt: EVs. They are fundamentally better cars, cheaper to maintain, cheaper to drive. And autonomy is going to surprise us on the upside. My oldest child is 10 years old. He is never going to have to drive. Once you know how to make batteries, it is cheaper and simpler to build an EV than a gas car. I think by the end of this decade, the US automakers will do some combination of merging down and federal bailouts. We are dramatically underestimating how this will affect society, the combination of autonomy and cheap EVs. You do not need auto body shops, gas stations. You do not need parking lots because the cars are always in motion.

Jay: How does your operating background help you support founders?

Matt: I have led sales, marketing, product, and operations across companies. Part of it is empathy, and part is that we can get into some level of detail on any of those topics. I frequently interview people for my founders because they find the feedback useful. Having walked in people's shoes, you get more of a sense of what is hard and where the tricks are.

Jay: Matt, that is a wonderful place to land. Thank you for investing in these founders, and thank you for joining me.

Matt: It is truly a pleasure and a privilege to work on these problems. It is a special thing.

Pull quotes

1. "One thing that people who are a little close to power but not real close misunderstand: right now it's not about energy. It's about power."

2. "The BYD is a great car. It's better than a Tesla. I hate to put that on record, but it is a really well-built car."

3. "My oldest child is 10 years old. He's never going to have to drive."

4. "We're dramatically underestimating how this is going to affect society: the combination of autonomy and cheap EVs."

5. "People need to be able to insure their homes or they can't live there. Our economy falls apart if real estate is uninsurable."

Source

From CLIMB Episode 083 with Matt Eggers (Prelude Ventures). Transcript cleaned from the published episode. Watch the full episode: https://youtu.be/L6Hm9XgL920