Kodiak Robotics’ founder says tight focus on autonomous trucks is working

Kodiak Robotics is one of the last private autonomous vehicle companies focused on trucking that is still standing. Nearly all the rest have been wooed by the public marketplace and the capital it can provide. But co-founder and CEO Don Burnette says the three-year-old company’s strategy of staying focused and small(er) is paying off.

It will be able to deploy a commercial-scale operation for about $500 million in funding, he says in the interview below. To put those go-to-market costs in perspective, that’s 10% of what Waymo has raised in external fundraising and less than 25% of newly publicly traded company TuSimple’s total fundraise.

Kodiak’s strategy is to take a specialized, hyperfocused approach to autonomous trucking that outsources a lot of tech, like data labeling, lidar, radar and mapping, to existing companies. Burnette, who was one of four founders of the self-driving truck startup Otto that Uber acquired, thinks this is a faster, cheaper and more efficient path to commercialization versus building out your own systems and teams.

The company is moving freight for commercial customers, dipping its toes in the market by working with technology partners within the existing ecosystem. Burnette says Kodiak’s Driver technology has achieved a level of maturity where it can handle anything the highway throws at it. In December, the startup achieved “disengagement-free deliveries” between Dallas and Houston, meaning the autonomous system didn’t have to be switched off for safety reasons.

The following interview, part of an ongoing series with founders who are building transportation companies, has been edited for length and clarity. 

You previously told me that Kodiak would need about $500 million in total funding to get to commercial driverless. You also said you’ve had some undisclosed funding rounds, but publicly, you’ve only raised $40 million. Can you still execute on your vision this far off?

Absolutely. We are always, as startups are, in fundraising mode. We’re always talking to investors. And there’s a lot of great things happening behind the scenes currently that we haven’t yet announced. We are growing, we’re hiring, if you can look to that as an indicator of the health of a company.

Our tech and our plan is really sound, and we are building up our commercialization efforts in a way that I think is going to be very exciting to the overall industry and to the market. We will need to raise more money, as you pointed out, that’s certainly no secret, but I think that we have multiple options to do that.

How do you intend to close that gap? Are you looking at venture capital, or maybe going for an IPO or SPAC?

We’re considering all of the above. It’s a constant conversation internally on what is the best path for Kodiak, what is the appetite of the various forms of investors and strategic relationships. Nothing is excluded.

The stock market is obviously very attractive and exciting. I think TuSimple has demonstrated that an IPO with the right set of metrics and the right set of momentum and partners is possible and can be successful. I think there’s also lots of opportunity within the VCs and the private markets. Kodiak is one of the only remaining serious AV trucking companies still in the private sector, and so I think that gives us some advantages in a lot of ways.

What’s your sense of the venture funding environment right now in autonomous? Is it harder now than it was, say, four years ago?

The appetite has changed. In particular, investors are more skeptical of timelines and promises. There is not this sense of Wild West excitement like there was four or five years ago, and that was the Golden Age of raising capital, certainly for earlier stage companies.

Kodiak was at the tail end of that age, and now the goalposts have changed, and the target investors have changed. It’s no longer the early-stage VCs that companies like Kodiak and others are talking to. It’s more of the growth-stage funds, and growth-stage funds look for different types of metrics. They look for commercial traction, product-market fit, users, efficiencies, etc.

The target demographic for strategic investors has also shifted. They’re looking for third-party validation and commercial partnerships in order to make determinations about the likelihood of success of a company. And so for us, that’s why I think it’s so important that independent of the market, we need to keep focusing on our fundamentals, on the technology. We have a really polished system now and it’s been great to be able to give demo rides. I think seeing is believing so focusing on that vision is really important. Building industry and commercial partners is obviously an incredibly important part of the story, and we are now actively working to improve our position in that area.

What would you say really makes your approach to the tech unique?

One of the unique aspects of our tech is that it’s highly customized for a specific goal. We don’t have this constant requirement that we maintain really high truck highway performance while at the same time really high dense urban passenger car performance, all within the same stack and system. Theoretically it’s certainly possible to create a generic solution for all driving in all conditions under all form factors, but it’s certainly a much harder problem.

We’ve taken a map-lite approach. HD maps are prone to go stale pretty much as soon as you build them, and they’re also hard to build at scale for highway driving. So I think that has given us an advantage over some of the legacy players that depend on this concept of HD maps, because our philosophy is that our system drives based on what it sees live. It’s always able to handle changing environments.

And then there’s our sensor-fusion approach. We are sensor agnostic. We don’t rely on any one sensor, so we’re not a lidar company and we’re not a camera company. We build a cohesive view of the world using all of our sensors as first-class citizens, which we think makes for more efficient real-time decisions.

Kodiak Robotics has stayed a small company because you rely on tech partnerships rather than building it all in-house. Is this how you’re able to close your funding gap? 

It’s certainly how we’re able to do more with less. It allows us to achieve our goals without having to spend as much money as some of our competitors.

If you built up an entire lidar and radar team in-house, as well as machine learning infrastructure, data labelling with a team of 500 people overseas, detailed HD maps that have to be annotated by hand — these things all cost a lot of money that you could be spending to hire the expertise to build up your tools. So there’s a lot of inefficiency in that. I would rather work with a company like Scale AI for our data labeling needs. We pay as we go, we pay for the labels that we need and we don’t pay for any labels we don’t need. To do it all in house, you also don’t get the learnings from working from the entire ecosystem.

What if you did have the money? What if someone was like, “I believe in you. Here’s a billion dollars.”? 

I think we would still work with partners. I actually think we’ve made faster, better progress with the path that we took versus if we had had more money and tried to build everything from scratch. Building things from scratch takes a lot of time, and you better hope that you built it right because building it twice, that can really be difficult. It’s not only technically difficult, but it’s politically difficult to admit to yourself, to your team, that what you built wasn’t the right product, and that you’re going to tear it down, take some learnings from it and build version 2.0.

5 fundraising imperatives for robotics startups

That’s where I think a lot of teams start to struggle. They have a lot of technical debt for things they built early on, but it didn’t in the end actually solve all of their needs. And then, as you grow you have management challenges and organizational communication challenges. You have to make sure you keep everybody on the same page. You have to keep your hiring bar extremely high. There’s a lot of complications that come with that type of model so I would say, even if we had unlimited money, the path that we are taking actually makes a lot of sense and is producing verifiable results.

Tell me as briefly as possible what your commercial business model is.

We have two initiatives for commercialization. The model we work within today is where Kodiak trucks are used to move goods for others, like carriers and shippers, within their ecosystem. We own the trucks, we outfit the trucks, we operate the trucks, and they pay us to move freight from A to B. We want to be able to offer this as “flex capacity” to existing industry players, so working with partners, carriers, fleets who have their own trucks, but can call upon Kodiak’s capacity as they need it. We’re going to be building that fleet over the next couple of years as we march toward our driverless launch.

Once we’ve launched driverless, we’re going to continue to operate that fleet, but really we’re going to do it in the spirit of working with existing carriers and fleets. This is when we’re going to go to the second initiative, which is to sell the technology directly to the fleets, basically to have Kodiak driver technology on the assets that other fleets and carriers own directly. We will continue to operate the truck, so we’re not asking them to have their own command center or remote monitoring services, but they will own the assets and do all of the logistics and manage customers and schedules.

Where do you expect to be a year from now? How far along in these developments?

A year from now, we will have increased our fleet size, we will have continued to grow the company. We’re aggressively hiring across the entire company in pretty much every department, so I think we’re gonna have a much stronger and more capable team a year from now.

I think that the system itself will be much farther along from a safety case perspective. We’re still going to be a couple years away from being able to launch an actual driverless system, but I think we will be in a much stronger position to be able to claim that our technology is going to get there.

People may not know that you have been a founder before, and it’s a pretty dramatic story. I don’t want to go into that now, but I’m more curious about what you have learned not to do from that? 

Dramatic, that’s a good word. What I have learned not to do is to not use HD maps! I don’t think that’s the point of your question. I’ll first address maybe some things that I have learned that are very important: Treating people with a lot of respect and building a culture that comprises compassion and effective communication.

I think of Kodiak as family, it’s not just transactional and when you really care about people like that, I think it creates an environment where people really like to be. When people care and they’re motivated and excited to come in each day, it ultimately produces a lot better results than environments where people are somewhat disgruntled or distrusting of leadership or their peers.

Many other founders come from legacy companies like Google. What’s something they should expect?

One thing that’s so easy to get lulled into when you’re at a company like Google is that so many things are provided for you. Google has a lot of great infrastructure, maybe some of the best infrastructure for code development anywhere. You get lulled into that just being the norm, and then you go to a startup and you have zero lines of code in your database, and you realize, “Wow, I have to build everything from scratch” — tooling, infrastructure, support. Those are all things you take for granted when you come from a large company.

But in another way, you’re not locked into the system because Google’s tools, Uber’s tools, they’re built generically, they’re not built for your specific use case. And so you have the opportunity to specialize and that often can create synergies and efficiencies that you would otherwise not have.

You really have to have a certain kind of attitude at the start, a willingness to kind of leave all that comfort behind and recognize that things might be not as good in some areas because you just don’t have that mothership there to support you.

From the perspective of someone who did leave, what causes someone to want to leave the bosom of a Google and spark out on their own? 

Everybody’s different. For me, I just fundamentally disagreed with the direction that the leadership was taking the project. I really believed that specialization was going to win the day over generalization.

Google loves to aim big, and that’s motivating for a lot of people and it certainly was motivating for me. They want to solve the hardest challenges, which is great from a cultural perspective, but it doesn’t always produce the best results. And so for me I wanted to really focus on a narrow problem. I wanted to focus on trucking and highway driving which I saw as the future of the industry.

Experts from Ford, Toyota and Hyundai outline why automakers are pouring money into robotics