Why some VCs are investing in hardware startups

Hardware has made a comeback in Silicon Valley, as evidenced by the crowded Maker Faire events (coming up in San Mateo, California from May 18 to May 20) to the recent HardwareCon event in San Jose.

Conductive Ventures recently raised $100 million to invest in enterprise hardware and software. One reason is that hardware giants have been acquiring innovative startups when those larger companies aren’t able to invent something in-house, which makes hardware startups a good investment.

I moderated a panel on hardware’s resurgence at HardwareCon. My panel included Jennifer Gill Roberts, cofounder of Grit Labs; Mike Edelhart, managing partner at Social Starts; Ashish Aggarwal, principal at Grishin Robotics; and Ephraim Lindenbaum, managing director of Advance Ventures.

Our focus was on hardware’s resurgence, how to do pitches, and how to elevate a startup beyond the Maker stage to get real funding and traction.

Here’s an edited transcript of our panel.

VentureBeat: I moved to Silicon Valley in 1994, and I remember when I arrived that people said that hardware startups were done. That was obviously wrong. We’ll get into that, but first please introduce yourselves. 

Jennifer Gill Roberts: I’ve been in Silicon Valley my entire life. I grew up here. I went to Stanford and worked at Sun and HP as a hardware engineer. After business school I went into venture. In the ‘90s I was funding optical and wireless equipment, up and down the stack. That was very capital-intensive. Some of those companies raised hundreds of millions of dollars and had only a handful of customers. Today actually seems a lot easier.

When I look at businesses now, to me they have some of the same elements: finding customer and market fit, trying to get launched on a reasonable amount of capital. I’m excited to talk about that.

Mike Edelhart: I’m a managing partner of two sets of very early stage venture funds. One is Social Starts, which tracks the impact of social mobile technology. That’s been active since 2012. The other is Joyance Partners, focused on the technological vectors of happiness. I look at both hardware and software. I was the founding editor of pretty much all the Ziff-Davis computer magazines, like PC Magazine and PC Link, the labs and benchmarks and things like that behind them. We tested every product, hardware and software, on the PC.

Ashish Aggarwal: I’m part of Grishin Robotics. We’re a $100 million VC fund, primarily investing in early stage companies, in hardware robotics and consumer.

Ephraim Lindenbaum: I’m managing director of Advance Ventures. I’m a recovering entrepreneur. I’ve actually sat in your chair before. I started in Silicon Valley as well, growing companies all through the ‘90s, not to date myself. Exited in the first dot-com boom. I came back to Silicon Valley after some time as a chief strategic officer. It’s pretty binary when you get back to Silicon Valley. You’re either starting a company or becoming an investor. I didn’t have any great ideas, so I became an investor.

We spun up Advance Ventures in 1999 and we’ve now invested through three successive funds. We invest in information technology, mobility, IOT, and what we call sustainability, which is pretty much agriculture and food technology. We’ve been involved in the connected device and hardware space since the ‘90s. With our entrepreneurs we built the first connected gas pump for Chevron, the first connected terminals in airports. Like Jennifer we were in telco and moved into hardware. Today about 20 percent of our portfolio is hardware-component. We just closed a deal this morning, a hardware-connected diagnostic device company. We’re quite happy in the hardware space.

VentureBeat: We have all the big companies – Intel, Cisco, Seagate. Why should there be hardware startups at all now? At the same time, we also have so many makers out there, people working in their garages and going to the Maker Faires. What seems to be going on, from your perspective, as far as why the hardware startups still exist?

Roberts: With every technology wave, you create new market leaders. It’s not in the DNA of the prior market leaders. I’m not saying these companies won’t be dominant for some time to come, but certainly we saw companies emerge in the PC wave, in the telecom wave, in the mobile social wave. I’m excited about the power of robotics. I think we’ll see completely different DNA.

The companies we’re funding through Grit Labs are coming to us with deep domain knowledge, with deep tech backgrounds. They’re recent PhDs. We’re not seeing as many companies iterate out of the dominant companies you mentioned. To me it’s that new DNA and new technology that drives new market leaders.

Edelhart: It’s always the best time to start new technology and it’s always the worst time to start new technology. There are always enormous entrenched players who look like they have infinite advantage and can’t be displaced. And they’re always displaced. Human behavior changes. Technology changes. When you put those two together, there are requirements for new capabilities that the existing companies aren’t set up to understand, respond to, or provide. There’s always a need for the new, and the old order will always change.

Aggarwal: The fundamental theory—when you think about it, what is robotics made of? Math, physics, all of these fundamental technologies. A lot of research has been going on for the past few decades, and in the past decade, a lot of technologies have started to come together, things that weren’t available 10 or 15 years ago. What AWS makes possible today—I used to manage hardware data centers for Yahoo. It used to take almost a year of planning to spin up a hardware data center. Today it’s so easy.

Similarly, the amount of data people can capture from devices—15 years ago, none of us used to have—how many connected devices did you carry around with yourself? Today everyone has at least one or two, all the time. Fundamentally, a lot of things have changed that allow new types of problems to be solved and new data sets to be captured. That has led to new types of advances and allowed new realms of technology to come together for people to build and create new types of robotics. New AI and machine learning are being applied to these larger problems in ways that weren’t possible before.

Lindenbaum: If this was Jeopardy, this would be the daily double. The most valuable company in the world is a hardware company. It’s Apple. It all starts there in terms of that conversation. To Jennifer’s point, hardware has been out of favor in the venture space for a while. That creates a drought. It creates the opportunity to innovate in particular categories.

The back office ecosystem of technology has come around to the point where these pieces work. Robotics without great machine vision isn’t great robotics. Robotics without great AI isn’t great robotics. The biggest challenge we see in this space is solutions looking for problems. Being able to drive into that, we’ve seen some wins in hardware in the venture world, companies like Tesla that have been successful, which spurs us to look back into that space.

A question in the green room was, with the Intels of the world out there, why are you all doing this? Innovation is value. It started to happen at the big companies, in the late ‘80s and through the ‘90s. It became an M&A process very quickly. These big companies couldn’t innovate purely on their own. M&A became a huge part of the R&D strategy at all of these companies, and it still is today. Look at Salesforce. Look at Apple. Look at everybody. They’re buyers. That gives you, as an entrepreneur, a clear trajectory to grow these businesses.

The next piece of the story is, there are great sectors where hardware is super meaningful. We’re a big player in food and agriculture. These are huge areas. There was a very meaningful robotics exit earlier this year in the space. In the food space there’s a ton of traction. If you look at some of what we call frontier markets, like cannabis, there’s tremendous movement in cannabis in hardware.

VentureBeat: How many cannabis investors do we have in the audience? We have one!

Lindenbaum: The reason why you want to look at that is adoption curves. When you have industries that have huge margins and incremental improvements translating into huge movement, you get high-speed of adoption. When you’re moving into the retail market where you have perhaps a three percent margin, getting any kind of technology adoption in rapid fashion is very challenging.

We’re investors, for example, in financial technology, where there’s a lot of retail movement. We’re investors in health tech, which is a huge IOT and connected device category. Food and agriculture are very meaningful at this point.

VentureBeat: You’ve already partially answered my second question, then, which is: What are the interesting categories you’re looking at? Do you want to talk more about what you like and what you might not do anymore?

Edelhart: We’re investors at the very beginning. We tend to see these trends when they emerge. Some of the things we’re seeing related to hardware—things like throwaway hardware, where the hardware is a smaller component of the overall solution. It used to be the hardware was the biggest, heaviest, most expensive part. Now a sensor is very inexpensive. It can be part of accumulating data sets that can then be exploited in other places. IOT sensors are one example.

We have one in agriculture where the whole system is taking advantage of the sensors that are already present on farm equipment. The sensors are there, the equipment is there, and the system is taking new advantage of what’s already there. Or taking advantage of capabilities in the phone and things like that, where the new hardware is an incremental element.

The other thing we’re seeing in hardware is hardware as a service. If the data sets are changing, or the usages are changing–the traditional hardware just does what it does. It does it the same way forever. It’s becoming a bit outmoded to the use. Hardware needs to become more flexible. The interaction between hardware and software needs to be more flexible. The whole system has to interact with the person more flexibly. We’re starting to see hardware less as the big piece of steak in the middle of the plate and more as a part of a mixed stew.

Roberts: We like verticals that are getting disrupted. Someone already mentioned agriculture. If you look at logistics, transportation, and construction, you can see that many Fortune 500 businesses are at risk. What’s going to happen to UPS? What has already happened to U.S. automakers? You’ve probably all seen them arrive in Silicon Valley and set up their venture arms and innovation centers. They’re starting to be very acquisitive. I like that these verticals are not only ripe for innovation, but also for acquisition.

I love robotics as a service because you can lower the barrier to entry, to getting it deployed. You can constantly upgrade it. You can delight the customer. You have a better lock. Our companies are very capital-efficient because it doesn’t need to be perfect and pretty like a consumer product. We’re looking for this intersection of things that lead us to believe that these companies are a prospect for a nice multiple or have the opportunity to be independent in these verticals.

Aggarwal: Among the things we look at, food tech is one of the big things, where you start seeing innovation happening in how food is prepared, how food is delivered, how producers are meeting demand. There’s a lot happening, mostly on the back side rather than front side. You see a lot of places where people are thinking about automation, like Eatsa is doing in downtown San Francisco. We see a lot of innovations where you’re trying to take the human factor out of how food is prepared.

A lot of stuff is happening on the agricultural tech side. Because of labor shortages and what’s happening with climate, people are looking at different ways of getting access to labor to pick different types of stuff – apples, other kinds of produce. People are trying to figure out ways to solve that problem.

There’s a whole bunch of things we see around self-driving. Everyone knows about Uber and Waymo and all these other companies building different stacks of software, both for getting people and getting goods from point A to point B. There are competing applications in almost any space: security, smart homes, transportation, logistics. All of these have big applications for computer vision.

We’re seeing companies try to build Amazon Go-style stores for almost every other retailer. Every retailer is feeling pressure from Amazon in terms of trying to improve their margins. They have to think about reducing costs as well as increasing revenue. Everybody’s trying to figure out things like inventory management at scale that can help reduce the cost of labor. That’s becoming a lot more prevalent.

On the consumer side, we’re focused on how people spend time. You cook, clean, drive, sleep. Companies are looking at how to automate some aspect of life. One of the first companies in the space was iRobot, because cleaning needs to happen in a constrained environment and it’s easy to build a robot to do that. Today we’re seeing robots for more complex applications in the home, like taking care of the elderly.

On the enterprise side, we’ve started seeing companies that are focused on either increasing revenue or decreasing cost. On the industrial side, the focus is primarily on understanding how to take humans out of hazardous environments, like using drones for the surveying of critical structures. That’s very difficult for humans to do.