From Sci-fi to Sci-fact with AlleyCorp’s Abe Murray

Abe Murray is a general partner at AlleyCorp, an early-stage venture fund that invests across deep tech, healthcare, enterprise software, and consumer tech.

What led you to investing in deep tech with AlleyCorp?

I’m a high school dropout who ran fishing boats and factories in my family business. That taught me how the real world works: Atoms, not bits. Forklifts, trucks, warehouses. But I’m a builder at heart. I spent a long career in tech, from engineering unmanned aerial vehicles to building businesses at Google. At Google, I built a major business in Android, ran the product team in Google Research where we shipped AI across Google including TensorFlow, and then spent five years taking AI to health tech at Verily Life Sciences where we raised $2 billion.

After more than 20 years operating, I was having more fun angel investing and supporting founders. I bumped into Kevin Ryan, who was looking to bring someone in to build a deep tech-focused investment arm for AlleyCorp, starting with robotics and automation. We both deeply felt that the future’s biggest problems—demographic shifts, climate change, comfortable living without destroying the environment—couldn’t be solved without heavy investment in automation and robotics.

Three years in, we’re investing in exactly that: robots, automation, aerospace, space, advanced manufacturing and energy. They might sound disparate, but they’re the most interesting areas in deep tech with near-term application to real-world problems. All of them combine hardware, software, and AI to do useful things in the world.

Why has deep tech suddenly become investable after years of being considered too risky?

The straightforward answer is hardware costs have come down significantly. Probably something like 10 times in the last 10 years. That’s the cost of a robot arm, a drone motor, or 3D printing. At the same time, software has had many orders of magnitude improvement in what it can accomplish.

Ten years ago, even if hardware was 10 times cheaper, it wouldn’t have mattered because the software couldn’t make the hardware do anything useful. Now cheaper hardware can be made to do many more things. We’re seeing this across our portfolio with successful 3D printing companies, drones going big, robots in more places. It used to be that you could only put a robot in a very expensive high-volume line. You needed millions of units to pay back the investment. When the equation changes, you can do more.

There’s also a generational shift. Young people want to work on problems that matter like climate change, feeding ourselves healthy food at scale, or being thoughtful about jobs that don’t just export pollution to other countries. The combination of software and hardware capabilities plus talent that cares about making things that matter is driving huge investment into problems that matter – both in financial capital and human capital. 

There have been a couple of big bets in deep tech but not a lot recently. That said, when people say they’re doing deep tech, many are secretly just doing AI or focusing on one piece like humanoids. They’re not actually investing across the sector in all the things that matter. There are still plenty of places in deep tech that aren’t seeing valuation run-ups and don’t have enough money. We need more investment, and I welcome it.

It’s worth remembering that the first billion-dollar company in the world, inflation-adjusted, was a deep tech company, Carnegie Steel. The first $10 billion company was deep tech. The first $100 billion company was deep tech. The first trillion-dollar company was deep tech. The huge enterprise value drivers in our economy have been hardware companies. This is really a reversion to the mean. If you want to have a huge impact on the world, you can’t do it with software alone. People are starting to recognize that.

When will we see more robots in our daily lives beyond the Roomba?

I had zero robots in my house when I started this. Today, I have seven including cat litter robots, cat feeder robots, and multiple vacuuming robots. But I can’t move from seven to 50 fast. We also have dishwashers and washing machines, but we call those appliances, not robots. When stuff becomes robust and reliable, something you can walk into a store and buy, we start thinking of them as appliances.

Consumer robotics still faces challenges. The best floor-cleaning robot costs $1,000. That’s not something most Americans can buy. I think we’re quite a ways from affordable and phenomenal consumer robotics. The real question is how quickly can AI close the autonomy gap? Waymo said they were five years away from full self-driving for 15 years before they actually shipped in San Francisco. The messy household is full of edge cases that will be hard.

The nonconsumer applications are where we’ll see answers first. A $10,000 humanoid can do amazing things in a factory or remote access places where you want labor but it’s hard to get labor. One of our companies, Earth Force, does forest fire prevention. They’re trying to get high-value labor into remote forests without people having to go there. They are addressing the question of whether we can build systems that go on top of excavators that provide a forestry recommendation for how to prevent forest fires and manage our ecology well without having to get people out to forests many hours away from where they live. 

Beyond robotics, what technologies are you most excited about?

Everything is predicated on abundant and cheap energy. It’s great to see solar rolling out faster every year. We have companies helping make solar “go brrr” like Civ Robotics, who is focused on outdoor construction automation for solar farms and data centers. We invest in infrastructure inspection because once you’ve built massive energy plays, you have to maintain them.

I’m very excited about Valar Atomics. I believe nuclear power will provide the baseload and a ton of the gigawatt hours humanity produces in the future. We made that bet before the recent change in nuclear thinking in this country. Valar Atomics is leading the charge in small modular reactors. These new reactors are impressive technological feats. They’re much safer and faster to manufacture. We can build factories to produce nuclear reactors at scale, which is what we need to solve climate change with cheap and clean energy.

We also invested in Portal Space Systems, who are building a 20x better satellite propulsion system so you can move satellites between orbits. That means longer lifespans and refueling capabilities. They can do this because of new manufacturing techniques, 3D additive metal printing, and AI design for heat exchangers. They’re taking ideas that were theoretically possible but not practically possible 20 years ago and turning them into something you can put on a satellite today.

What makes deep tech founders successful at commercialization?

I’m a broken record on this, and AlleyCorp has been uniquely good at it: be market-focused, customer-focused, obsessed with “Can we sell this thing?” Before we create companies, we do tons of customer interviews. What does the market need? How do we sell this? Does the world need it enough for the company to deserve to exist?

I’m a classic introverted engineer who loves tech and learned the hard way that sales and go-to-market is probably the most important thing. It’s not just sales to sell your product. It’s sales to convince investors, to convince employees to take a bet on your crazy dream, and to convince customers to buy. You need to find the thing they need and morph your company into that.

I literally give founders a mini-MBA course on becoming a salesperson. We call ourselves the department of inspiration. It’s a learnable and coachable skill. You don’t have to be born with it to be good enough to be amazing. We meet companies all the time that get go-to-market wrong and then don’t matter.

How do you define deep tech?

I have a tagline that says “sci-fi to sci-fact.” If you’re taking something that was a science fiction idea and implementing it into reality today, and most people would say “I don’t think that’s possible yet” or “That seems crazy,” you’re probably doing deep tech.

I think there’s a lot of money and societal benefit to be made on the edge of what was deep tech and impossible and is now possible, but people don’t realize it’s newly possible. We take engineering risk, not science risk. We’re making bets where there’s an engineering challenge that’s tractable. 

The companies I mentioned—Valar Atomics, Portal, Eyebot—they’re all doing things that are really hard. You need a special team, well-trained, and who’ve spent time working on the problem space. At Eyebot, they spent 10 years working on eyeglass prescription optics before realizing the moment was now when you could build their solution. Does the world need it? That’s the section of deep tech I focus on.

SPONSORED

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

Learn more