Why VCs Are Betting on Energy Startups in 2026

Lucas: So a16z-backed Base Power just announced they're offering cheaper electricity to the grid - specifically to the parts of the power grid that need it most. And I think this is one of those venture stories that a lot of people in software-focused VC are missing right now. Luna: I saw that. Base Power - they're building these small, distributed natural-gas generators that can ramp up quickly when the grid is strained.

It's almost like a virtual power plant model. Lucas: Exactly. And the reason this matters for venture is that for years, the conventional wisdom was that energy infrastructure was too capital-intensive, too regulated, too slow for VC returns. But the AI data-center boom is changing that math.

These data centers need massive, reliable power - and the grid isn't keeping up. Luna: Right. And we're seeing that play out in the market. Look at Palantir - down almost 16 percent in the last five days.

That's not just an AI hype correction. A lot of that sell-off is tied to concerns about data center energy costs and the infrastructure bottlenecks that could slow down AI deployments. Lucas: Yeah, Palantir's drop is brutal, and it's not alone - the whole AI cohort is getting hammered this week. But interestingly, ARKG, the genomics ETF, is up 7.

7 percent over the same period. So capital is rotating out of pure AI plays into areas like biotech and energy where the infrastructure story is more tangible. Luna: That rotation is exactly why VCs are looking at Base Power. It's not just a climate play; it's a capacity play.

The grid literally cannot meet demand from data centers and electrification at the same time without new generation sources that can be deployed fast. Lucas: And Base Power's pitch is smart - they're not trying to replace the grid, they're offering peak-shaving services to utilities. They get paid to be on standby, and then when prices spike, they dispatch power. It's a much lower-risk business model than trying to build a full-scale utility.

Lucas: But here's the thing - this is still hard tech. It requires manufacturing generators, navigating permitting, dealing with fuel supply chains. It's not a two-developer-and-a-Slack-bot startup. Luna: So the question is: can VCs actually make the 10x returns they expect from something like this?

Energy hardware has historically been a tough asset class for venture. Lucas: Right, and that's the debate. I think the answer is that the return profile might be lower - maybe 3x to 5x over a longer period - but the check sizes are bigger, and the market is enormous. If you're a16z and you're writing $50 million checks into AI, putting $20 million into an energy startup that could serve those same data center customers actually creates a portfolio hedge.

Luna: That's actually a really good point. It's not just about the energy startup alone - it's about how it fits into the broader portfolio. If you're backing the AI companies that need power, it makes sense to also back the companies that provide that power. Lucas: Speaking of things that fit together - if these conversations are useful for what you're building or running, we keep this show ad-free because of listener support.

If today's episode was worth a coffee to you, that's actually the link: buy me a coffee dot com slash fexingo. Luna: Yeah, it genuinely helps us keep digging into these corners of venture that don't always make the front page. Lucas: Exactly. So back to Base Power - one thing I find really interesting is that they're not even the only one.

There are a bunch of startups now building what are essentially grid-scale batteries, modular nuclear reactors, even compressed-air energy storage. All of them are getting VC attention. Luna: But there's a tension here. The VC timeline is usually 10 years.

Building a nuclear reactor can take 15 to 20. How do you square that? Lucas: You don't - for nuclear. That's why we're seeing more money going into natural-gas peaker plants and battery storage, which can be deployed in 2 to 3 years.

Base Power is a great example: they're using existing gas turbine technology, just packaging it differently. That's a venture-appropriate timeline. Luna: And the economics work because of the volatility in power markets. When a heat wave hits and prices spike to $1,000 per megawatt-hour, these plants can make a year's worth of revenue in a week.

Lucas: Exactly. So the VC thesis here is really about optionality. You're betting that the grid will remain stressed, that regulation will slowly ease, and that you can build a defensible position in a local market before the utilities wake up. Lucas: Now, I want to connect this to something we talked about a few episodes ago - the AI infrastructure play.

Remember when we discussed how VCs were pivoting from AI models to AI infrastructure? This is a natural extension. Luna: Right, because AI infrastructure isn't just chips and data centers - it's power, it's cooling, it's the physical stuff that makes the digital stuff work. Lucas: Exactly.

And that's why I think we're going to see more VC dollars flow into energy over the next 12 to 18 months. The question is whether the returns will follow. Luna: What's the biggest risk for these startups? Is it regulatory, or technological?

Lucas: I'd say it's actually market structure. Utilities are regulated monopolies in most places. They can make it very hard for new entrants to connect to the grid, or they can undercut on price if they want to. So the startup needs a really clear go to market that avoids direct competition with the incumbent.

Lucas: Base Power's model of being a supplier to the utility rather than a competitor is clever for that reason. They're selling reliability, not electricity per se. Luna: That's a good distinction. And it's similar to what we saw in the early days of solar - the companies that succeeded were the ones that partnered with utilities, not the ones that tried to replace them.

Lucas: Yeah, exactly. So if you're a founder looking at this space, I think the lesson is: find the pain point the utility can't solve internally, and build a solution that fits into their existing infrastructure. Luna: And for VCs, the lesson might be that energy is back on the menu - but you have to be patient and think about portfolio fit, not just standalone returns. Lucas: Right.

And that brings us back to the broader market context. With the AI names selling off, VCs have capital that needs to go somewhere. Energy infrastructure is one of those places where the thesis is solid even if the timeline is longer. Luna: Do you think we'll see more traditional energy VCs - the ones who did cleantech 1.

0 - come back into the fold? Lucas: I think we already are. A lot of the partners who got burned in the 2010s cleantech bust are now looking at this new wave with different technology and better business models. Base Power is a good example - it's not a science experiment, it's proven tech applied to a new market.

Luna: So this might be the moment where hard tech and venture capital finally find a sustainable relationship. Lucas: I hope so. Because if we're going to meet the energy demands of AI and electrification, we need more than just software solutions. We need real, physical infrastructure - and that means venture capital has to adapt.

Luna: Yeah. And it's fascinating to watch that adaptation happen in real time.