Why Hardware Beats Software with Wave Function Ventures’ Jamie Gull

Jamie Gull is the founder and solo GP of Wave Function Ventures, a pre-seed and seed fund investing in hardware deep tech companies across aerospace, defense, energy, and robotics.

What’s your journey from aerospace engineering to venture capital?

I studied aerospace engineering at Stanford, got my undergraduate and master’s there. After that, I went to Scaled Composites in the Mojave Desert, the best-known prototype aircraft company in the world. Most people know them for Spaceship One and Two that turned into Virgin Galactic, but they churn out tons of prototype aircraft yearly. When I was 24, I was part of a program called Firebird, a 5,000-pound optionally manned aircraft for electronic warfare testing. We designed, built and flew that in nine months. I was responsible for the tails, the booms and all the control surfaces. Really got to cut my teeth rapidly there and build hardware that actually flew.

After a few years, I joined SpaceX in 2010 and was there until 2016 in the structures group. My biggest contributions were the aft thermal shield on the bottom of Falcon 9 that enabled the reentry and reuse of that rocket. I was there from the earliest days when the rocket would reenter and blow up over the water until we got it to actually work. Also did a lot of work on Falcon Heavy structures, holding all the engines onto the rest of the rocket and holding those three boosters together.

After SpaceX, I started two deep tech companies as a co-founder. One builds space deployable systems that fold out for antennas at low mass and low volume for launch. The other was Talyn, an electric aircraft company that went through Y Combinator in 2019, raised a seed, got eight contracts with the Air Force, built an airframe, did flight tests, and got acquired by Ampaire in 2023. Along the way, I started angel investing and doing work with Pioneer Fund, which is all fellow YC founders that invest in YC companies. I really enjoyed that part and decided to launch Wave Function Ventures last year.

What makes Wave Function Ventures unique as an early-stage deep tech fund?

Wave Function Ventures invests in pre-seed and seed hardware companies in aerospace, defense, energy, robotics, and industrials. Anything with a primarily hardware component. I highly leverage my experience as an engineer and a founder for finding, evaluating, and helping founders once I’m in.

On the tech side, I don’t love investing in science projects or lab spinouts. I don’t want to fund somebody to go see if a technology is even possible in a lab for two years to find out it’s not. That makes no sense for a small fund. It’s more about finding tech that’s definitely doable but requires excellence in execution and a really strong component around why now. You’re combining existing technologies or techniques in some new way at the right time with a great team. I’m generally not doing a lot of research on whether a technology is even possible to gain conviction. It’s more around the team, the timing, the market and the why now.

Everything I invest in, whether the founder admits it or not on day one, has a government component to it. It might be defense, it might be something else. But there’s almost always a heavy regulatory component, and what they might not realize right up front is the Department of Defense has a use for this technology and there’s likely available non-dilutive funding and a real customer there.

Why is understanding the government component so critical for deep tech companies?

If you do it correctly, you can use government funding to develop your tech on a roadmap that you would have done either way. Why wouldn’t you take non-dilutive funding while leading toward a potential marquee customer? You develop your commercial tech and can sell to commercial customers off the back of your development work with the government. That can bridge you over the longer timelines it takes for government acquisitions or procurement.

But it’s super opaque, so my piece of advice is get advice. Don’t try to go at it yourself. If you don’t have a national security background, you will flounder. You need to work with either internally or externally experts who can guide you through the process. Get language written in the NDAA. That’s a long-term process, and a lot of folks think “I need to develop my technology and then I’ll go do that.” That’s the wrong attitude.

The government doesn’t know how to help you, which is especially odd. If you get a Phase One SBIR with the Air Force to go talk to customers, your goal is to get an end-user and customer memorandum within the Air Force so you can move on to Phase Two. Your contracting officer doesn’t know who to talk to. So you have to find somebody, and then you actually have to educate them on how the SBIR process works. They don’t get training. The founder has to say “I can provide this benefit to you. Is this of interest?” They say yes, and then you have to educate them on getting a customer memo, signing on as a technical point of contact, and working through this with you. They have no idea unless they’ve been through it before.

What are the advantages of working with a solo GP in deep tech?

Speed and conviction. If you’re trying to get an investment committee on board, the chance of somebody getting hung up on capital requirements or government sales is multiplied. If you have two, three, four people that have to sign off, it takes a long time and you’ll likely run into roadblocks. I think that’s a better fit for a seed round when there’s more progress to show, more concrete data points.

My goal is to provide super high-leverage advice, but I’m not going to do the work for founders. Some people ask if I sit down and help with engineering. God no, I don’t have time for that. Why would I invest in somebody I have to help with engineering? That’s a terrible investment. You’re investing primarily in the team. It’s more like what we talked about with the government stuff: “Here’s how you need to go do this. Here’s three people to go talk to for help.” Or for first-time founders, hiring is brutal. Here’s a process you can build. You need to allocate probably 50% of your time for the first few months. People will be gobsmacked by that, but unless you sit down and do it, you’re not going to go anywhere.

Is there a playbook emerging for building hard tech companies?

As it becomes more popular, there is more of a playbook forming. It’s your government momentum, your hiring momentum, how much product momentum you have, and the ability to translate some ideas from the software world to hardware, which is iteration. You can’t “move fast and break things” in hardware with the same cavalier attitude, but you do need to bring some of that to the table.

That’s why it’s harder to invest in somebody coming out of Raytheon, for example. They’re used to spending a year getting their requirements perfect and building the perfect product on day one. But SpaceX showed that’s a terrible idea in hardware. The right way is to iterate quickly, blow stuff up and learn. You have to balance that with your funding and time lost, but people still under-index on time lost and money lost. When you iterate quickly on hardware, it’s still cheaper to iterate and have failed tests both time-wise and money-wise than to get it right from day one.

How do you view the current state of venture capital in defense tech?

It’s going to be really interesting to watch it play out. When I started Talyn in 2019, there were maybe 10 to 20 SpaceX alum companies. It was a small community. Now I think there’s close to 200, and there’s a lot of people piling into the venture world here. You get a couple problems. One is dumb money chasing dumb ideas.

You also get a player like Anduril, and now you have to ask every time you look at a company: Is Anduril just going to step into this sector? They’re rolling out a new product every two to four weeks it seems. Just because it’s a good idea and a good team, the person with the massive lobbying team that gets all the language written for themselves, even if you have better superior tech, you could probably lose. So you can go downstream or upstream. Maybe it’s a supplier of a key technology that sells into Anduril or the primes without going too much into component land, which I don’t like because you don’t capture value.

What belief about deep tech will be common knowledge in five years?

Hardware is not too expensive. That’s just not true. The reason it’s not true is that these days building something can be cheaper than hiring a software engineering team. You can hire a bunch of pretty killer engineers for $150,000 a year in Los Angeles. Say you raise a $3 million seed round, you can hire four or five people and then go spend a million dollars on hardware and build out a great prototype.

If you go to the Bay Area building another SaaS company, how many engineers can you get for $3 million for a two-year runway? Like a few. They’re probably asking for $300,000 a year. You don’t have to spend any money on hardware and can maybe iterate a bit faster, but especially with AI, SaaS has become commoditized. The only thing that wins in SaaS is distribution, and distribution costs a ton of money.

In hardware, you have to pay more upfront, but once you have customers settled, it’s more organic and more paced growth that’s super sticky. You have a built-in moat because the hardware is hard and because you have a relatively limited customer base with lock-in. It’s a lot harder to swap out a major piece of hardware than a piece of software.

How do you define deep tech?

For my fund, I define it as hardware, and I don’t do software for hardware. I’ll just leave it at that for the fund. If you’re looking in general, it’s all over the map. If a normal VC gets scared away by it, it might be deep tech.

SPONSORED

HAUS specializes in public relations and creative services for deep tech startups.

Learn more