(Bloomberg) -- As a founding partner at the San Francisco venture capital fund Trucks, Reilly Brennan has met with thousands of entrepreneurs with ideas to revolutionize transportation. From tractors to scooters, if it has wheels, he’s probably heard a pitch about how to make it driverless, electrified and internet connected.

The fund, founded in 2015, has about 40 start-ups in its portfolio, including Roadster, an online sales platform for car dealerships, and the self-driving shuttle company May Mobility. Trucks aims to be the first investor in a new venture other than friends and family. Before forming the fund, Brennan and his partners invested in Cruise Automation, the self-driving car company that General Motors acquired in 2016 for more than $500 million. That same year, Trucks took a stake in Starsky Robotics, the automated trucking company that folded earlier this year.

Brennan, 41, grew up in Detroit and has been working in the transportation industry since his undergraduate days at the University of Michigan, when he was a “motor gopher” at now defunct Automobile magazine. “I was washing and gassing cars for $7 an hour,” he says. Later, he worked at Stanford University raising money for its automotive research program. He currently teaches classes in transportation entrepreneurship at the university.

Brennan recently spoke with Hyperdrive about the effect the pandemic has had on investing, timelines for autonomous vehicles, SPAC fever, and the failure of Starsky. The interview has been edited for clarity and length.

What has changed in the last six months, from your perspective as an investor?

In March, I thought we’d see a huge drop off in company creation. And since I’m an early stage investor, that for me is the seed corn. You need great people to want to take a chance to start a new company. And my fear was that we would lose that source. That hasn't been the case. In July, we saw 94 new companies. That's pretty healthy for us. August was 80. We usually see a little under a thousand companies a year, a little under a hundred a month. That's companies when we hear about them for the first time. So the numbers have been really strong. We’ve made three brand new investments, where we didn’t know the teams before COVID hit.

What do you look for in a start-up now that six months ago didn't matter as much?

I think that one of the areas that will probably become a big, horizontal business is cleaning and sanitation, particularly for anything that touches shared resources. We need a lot more technology in that space. Most consumers, up until COVID, really believed cleaning was about no visible dirt or odor. Now the cultural perception of how something is clean is a lot different. Knowing that you’re the first person to use a space is really important. You know when you go to a hotel and the toilet paper roll has the little fold in it, just to point out to you that you’re the first person to use a bathroom. Those types of visual cues that confirm for you, the end user, that something’s been sanitized are going to become really important in transportation. I did some research on an interesting company that had this product that was used for relief workers, to make sure that their hazmat suits were clean. They were trying to figure out a way to prove to everybody that they had cleaned a suit with enough bleach. They put this highlight powder as an additive into the bleach. When you spray the bleach, it’s got blue in it. So if you're a frontline worker battling an Ebola outbreak, you can actually see how the thing is getting cleaned.

Do you have a sense of when a vehicle without a driver will be something that most Americans have experienced? Will that day actually come?

I think September 2nd, 2035. You can put me down for that, 15 years from now. We’ll set a Google alert to come back to this.

There’s this perception that autonomy is going to take too long or it’s not worth working on and investing in anymore, which I find really funny because none of the autonomy engineers that I know ever made promises that we’d have self-driving cars in 2020. It was the auto executives and the people who get up at TED and CES who would make those proclamations, knowing full well that they would probably be playing golf by the time 2020 actually came around. So while I think autonomy as the car companies define it, this robo-taxi feature, is going to take quite a long time, decades, there are a lot of really large markets with important applications where you need autonomy now, like in logistics and construction and mining and agriculture. We would love to have the best company in each one of those segments.

Do you have a roadmap in your mind to the driverless future? Does it start with long-haul trucking? Or do you see all these things happening at once?

I used to believe that there was this caterpillar-into-the-butterfly approach, but I think less of that idea now. I believe that these [driverless vehicles] are mostly going to do particular jobs well countless times before we see expansion into the broader public and robo-taxis. I’m also of the belief that we’ll never actually achieve what the SAE would call level five [any road, anytime autonomy]. Humans are dumb enough to drive in awful conditions, but machines aren’t. The fascinating question is what are the areas where level four [driverless in defined areas and conditions] gets better and better? Is it high-density areas, because you’ve run them so much and you’ve been able to hang sensors in the infrastructure that allow you to get better? Or is it that you’re able to run in low density areas because the risks are lower and it’s easier? I’m not smart enough to answer those questions, but I think in the next decade we’ll see more and more of those types of decisions being made.

It’s so far out, fifteen years from now, that every single one of these companies will have multiple management changes in the meantime and the market’s going to move and the world’s going to change. The commitment required is almost greater than the engineering task itself. It’s a $20 billion gambit. You need to have the internal approval and the board alignment to be able to go through multiple generations of leadership and many billions of dollars. How many companies are willing to do that? Maybe a few. 

Is there still a place for venture capitalists in this?

The timeframe for venture funds is usually max ten years and most funds have a three to four year investment cycle. So if you’re getting a check from a venture fund on the tail end of their investment period, you really are expected to get that big return within six or seven years. And with the timeline here, that’s not gonna work out. The big acquisitions, for robo-taxis-as-systems within car companies, those are more or less gone. The bigger opportunities are in components or in application-specific things like trucking or agriculture. It would surprise me if Walmart, for example, made big investments in technology around autonomy in the next five years, whether it’s last-mile or middle-mile or long-haul. In things that are off-road, like agriculture, you could think of John Deere or Caterpillar. At the component level, there’s a lot that the big car companies and suppliers are not that great at, that start-ups are doing a better job of, whether it’s computer vision or driver monitoring or over-the-air software.

What do you make of the current trend of special purpose acquisition companies, or SPACs, taking transportation start-ups into the public markets through reverse mergers?

For EVs, I think there’s probably too many. There’s a lot of SPAC fever at the moment. Just because you’re able to do a SPAC doesn’t necessarily mean the company is a worthwhile investment. Some of the companies that have taken that route, I’m not necessarily convinced that a SPAC is going to get them enough money. Both Silicon Valley and Wall Street underestimate how much it costs to take a car through validation and get ready for starting production. If you look at the new Mercedes Benz C class sedan, which is just a nice sports sedan, Daimler spent five and a half billion dollars getting it ready for production. The amount of capital required to just bring a normal car to the green light is really breathtaking. The SPACs that are going to do really well are companies that are already profitable or just at the edge of profitability. I’ve talked to a lot of people in the last couple of months who were thinking about doing SPACs and they already have an evaluation list in hand and are saying ‘Can you help me look at this list?’ Few of the companies on those lists are anywhere close to profitability. What did you learn from the failure of Starsky?

I am not sure which segment of logistics is going to be most valuable in the near term for autonomy. We knowingly made three investments in logistics autonomy. Starsky was one of them, but we also invested in Gatik AI and Refraction AI because we weren't smart enough to know which of these three is really going to be the best at the moment. So we at least had a bit of a portfolio approach. One of Starsky’s main issues was that they wanted to be both an autonomy company and have their own operations. They effectively wanted to become a carrier like a UPS. I often thought that was an advantage, that you could say, not only am I going to move your crate of oranges autonomously, but since I run the truck, I will have my own level of service requirements. We thought that was really important for Starsky, because if they had to pull over on the side of the road for 20 minutes to debug something, that was their own network that they could control. It turns out that was actually quite a big leap for investors not only to understand but also to get behind. There were VCs downstream who thought it was too challenging for them to own all those trucks and run that service. So their next round didn't come together.

©2020 Bloomberg L.P.