Scaled and Failed #9 – Plus and Starsky Robotics

Rubber Hits the Road, Startups Hit the Highway

Mar 19, 2021

Welcome to Scaled and Failed! My name’s Amil Naik and I’m an aspiring VC and founder at The University of Texas at Austin. I write about startups that scaled and startups that failed to draw insights about the patterns of startup failure and how to avoid them. Everything is clearer in hindsight, so it’s worth looking back.

If you haven’t subscribed already, do it here:

TLDR

Today’s Topic: Autonomous Trucking
Scaled: Plus; Accelerated its arrival to the market by focusing on China and built out extensive partnerships so it could focus entirely on its primary product over anything else.
Failed: Starsky Robotics; Investment failed to materialize due to changes in expectations and industry factors that pushed its vision out of favor with VCs.
Lessons Learned: Your go-to-market may not be suited for the locale you assume it is, so invest in a strong market analysis to determine the best place to launch rather than the most convenient or simple location to start in. In slow, capital-intensive industries, you will be at the mercy of investors for a long time. Make sure to keep them invested in what you’re doing and pivot to follow the larger vision for the industry to avoid losing interest.

Today’s Topic: Autonomous Trucking

When trucks are no longer driven by people, many of us will miss pumping our arms and getting the drivers to blast their air horns. Hopefully, autonomous systems will be programmed to respond the same way, but it’s impossible to deny the supply chain improvements that could be realized with autonomous trucks. There’s been a severe shortage of truck drivers in the last several years, and it’s only getting worse. Autonomous trucks pushing Level 4 autonomy are meant to be tested without supervision on the road this year, which may start alleviating the demand for drivers while cutting costs in a big way. It will be some time before regulations and infrastructure catch up to the possibilities of the technology developed, but incremental improvements have already been seeping into the trucking industry.

Trucking handles around 72.5% of the freight in America and the industry approached $800B in revenue in 2019. Autonomous trucks won’t be required to take breaks as required by law, can save on fuel, make better time, and are a safer alternative over long distances; the savings potential is enormous, with McKinsey estimating a 45% reduction in costs once Level 5 autonomy is achieved. The pandemic has only hastened the demand for autonomous technology because of its safety concerns, accelerating the adoption timeline. The entire supply chain will feel ripples as goods can be moved faster and more cheaply, which has great implications for consumers. The transition will take some time though since won’t be ready to go completely unsupervised for a while, and legal issues will likely keep people in there for a while longer. Until then, many semiautonomous improvements will be added to ease the job of truck drivers just as they’ve aided average consumer vehicles in recent years. Plus is one of many companies tackling autonomous trucking, having raised $200M in February to support the deployment of its product in the US and China this year. Starsky Robotics also attempted to bring autonomous trucking technology to the market but shut down early in 2020 amid a failure to raise its next round.

Scale: Plus

Plus was founded in 2016 by several classmates at Stanford and is one of the leading firms launching autonomous trucking technology. They’ve had several impressive milestones in the last few years that put them on the map; most notably, in 2019 the company ran the first cross-country Level 4 autonomous commercial pilot (though it was supervised by safety drivers and had stops to comply with federal regulations) with Land O’Lakes. 2,800 miles of road from California to Pennsylvania while hopping hub to hub is an impressive feat for a self-driving vehicle, though it is still far from the dream of Level 5 autonomous trucking. The company managed to be the first autonomous trucking company to get a California Autonomous Vehicle Testing License back in 2017, and they’ve pressed that early advantage to accelerate a full commercial launch in the near future. In 2019, Plus made serious moves into the Chinese market by partnering with FAW Jiefang, one of China’s largest truck makers, to create autonomous trucks for the country This joint partnership launched a Level 2 autonomous truck as its first product, cementing Plus in the Chinese marketplace. Plus also had a relationship with the Chinese startup Full Truck Alliance that deepened in late 2020, bringing the autonomous FAW-Plus trucks for sale on the FTA platform and providing a wealth of data and real-world testing opportunities to prove the safety of its technology. The FAW-Plus trucks had reached Level 3 autonomy at this point and were gaining commercial traction with this new sales channel through FTA. These trucks passed the national certification test and met all the safety and performance requirements needed to start hitting the road in China full throttle that same year, becoming the first automated truck to pass the test.

Mass production of the FAW J7+ is meant to begin this year, and while Plus has not reached the same stage in the US, it had many commercial pilots in 2019 and 2020 besides the one with Land O’ Lakes. They have been aiming to test their system across all states that allow testing to build experience for their system across different terrains and weather. They also partnered with the Transportation Research Center to conduct independent testing in complex, multi-car scenarios to demonstrate adherence to safety standards in the US and Europe. Early customers in the US can get their hands on their autonomous system, PlusDrive, this year. The startup is planning on starting pilot operations in Europe this year as well, making inroads for Plus to add value to supply chain logistics globally. Their recent $200M raise is meant to support this expansion, with the goal of providing support for fleet integration and more rapid deployment in current and new markets.

Plus has weathered a tumultuous journey for the autonomous trucking space that has left only a few startups well positioned to advance. Competition is fierce, and Plus has framed itself as a technology partner rather than a fleet operator, unlike many competitors. This allows for faster deployment of their product, as they aren’t building trucks themselves or running a transportation network. Testing and deployment in China were also beneficial to the acceleration of Plus’s development, as the US and individual states tend to move a lot slower on regulations and testing for technology like this. With strong sales channels open, the regulatory go-ahead in China’s massive market, and a manufacturer working with them to deploy their system, Plus has the pieces needed to become commercially viable in the mainstream. Only a few competitors have a similarly strong setup, with TuSimple being the most notable. TuSimple is working with an American truck manufacturer and is building out its own transport network for hub-to-hub trucking in the US, contrasting with Plus’s more hands-off approach. Both autonomous trucking heavyweights have a lot of support behind them, but I’m a fan of Plus’s choice to stay out of the fleet operation aspect and simply provide a system that is easy to integrate. Regulations are still far behind the technology of both of these companies, but autonomous trucks will dominate the landscape very soon.

Fail: Starsky Robotics

Starsky Robotics was founded in 2015 and sought to combine autonomous capabilities on highways and allow truck drivers to remotely control vehicles through more difficult situations, such as the first and last mile of operation. Their approach was a more moderate take on the eventual capabilities of autonomous driving, as they kept humans as a part of their vision for success rather than striving to remove people from the equation entirely. This was seen as more feasible than achieving a business case of fully autonomous trucks that met all regulatory and safety standards. Beyond just the driving, Starsky wanted to automate the entire dispatch and administration side of trucking to solve inefficiencies in the industry. In regards to Starsky’s business model, the CEO, Stefan Seltz-Axmacher said this:

Humans are better at navigating many of the nuances of driving than even the most advanced computer systems. This is why we use remote drivers to help our trucks at their most contextually complex junctures. We are not eliminating drivers' jobs. Instead, we are moving them from a truck to a safe and comfortable office where they utilize their years of long-haul trucking experience but remain close to their families and go home between shifts.

The startup operated much like a regular trucking company, with some regular drivers moving freight and several autonomous trucks operating with remote drivers supporting them. Starsky had its own milestones in the industry; it powered the first unmanned truck drive on a public highway with nobody onboard in 2019. That same year, the company launched an API called Hutch that allowed shippers to purchase capacity directly on its autonomous trucks without any human involvement with the goal of eventually having all loads brokered through the API so that employees could focus primarily on the autonomous technology. Using the Hutch API, they were able to integrate with an automated dispatch API from LoadSmart, a digital freight brokerage, and handle load shipping from quote to book to delivery without any humans involved. Things appeared to be on track for Starsky, but the company was unable to raise additional funding early in 2020 and was forced to shut down.

Seltz-Axmacher wrote a post-mortem breaking down what led to the closure of Starsky, focusing on macro factors in venture capital and the industry overall that led the round to fall through. He believes that the promise of full autonomy was overhyped and on a much longer timeline than many expected, as improvement becomes more difficult using supervised machine learning as the AI got better and better. With a significant slowdown in progress, interest in the space cooled as VCs realized it would be extremely difficult to build a business on technology that has a far horizon to deliver on the human-level decision-making that was expected. Additionally, he pointed out that Starsky’s business model that was more akin to fleet operation rather than software provision is more unattractive to VCs as it has lower margins, even if the market is still sizeable and holds the same potential for growth. He also found that demonstrating all the work that went into safety engineering didn’t provide the proof VCs wanted compared to additional features:

The problem is that all of that work is invisible. Investors expect founders to lie to them — so how are they to believe that the unmanned run we did actually only had a 1 in a million chance of fatality accident? If they don’t know how hard it is to do unmanned, how do they know someone else can’t do it next week? Our competitors, on the other hand, invested their engineering efforts in building additional AI features. Decision makers which could sometimes decide to change lanes, or could drive on surface streets (assuming they had sufficient map data). Really neat, cutting- edge stuff.

All in all, to Seltz-Axmacher it came down to the timing that brought all these factors together that caused investor interest to fizzle out. He closed with a grim outlook for autonomous trucking in the near future:

The current companies who are [betting a business on safe AI decision makers] will continue to drain momentum over the next two years, followed by a few years with nearly no investment in the space, and (hopefully) another unmanned highway test for 5 years.

Not everyone agrees with that analysis, however. The writer of this article provides a rebuttal of many of Seltz-Axmacher’s perspectives. He argues that although there was a decrease in investor and automotive manufacturer interest in autonomous vehicles, many companies hit major milestones for going driverless in 2019 and 2020 and the time period was merely a small barrier. He also pointed out that though supervised machine learning has its limitations, other techniques are available to help spur progress in the space. Additionally, he spoke with representatives from Plus and TuSimple for their views:

Not so, say others in the truck ADS space who I spoke with. Maybe those breakthroughs “failed to appear” for Starsky, but a spokesperson for TuSimple said “We believe our solution is advancing well and the AI approach we’ve taken is delivering well on its promise.”
Plus.ai agrees that Starksy’s experience is not representative of the entire truck ADS space. Shawn Kerrigan, COO and co-founder, said “The capital markets are becoming more selective when investing in autonomous trucking companies, which is wise considering that several early market entrants over-promised and under-delivered. While the number of autonomous trucking companies is now smaller as the market matures, the amount and pace of investment going into that smaller group of companies is increasing. Most importantly, the broader ecosystem’s commitment to autonomous trucks remains unchanged —the world’s largest OEMs, Tier 1 suppliers, and regulators continue to invest in and prepare for the launch of autonomous trucks in several years.”

Both of these companies have continued to raise enormous amounts of capital to fuel growth and are launching commercial autonomous solutions despite any reduction of VC interest in autonomous trucking. The author brings up many more large raises in the space along with large internal investments by automotive companies to dispel the interest of waning investor interest. He concedes that many VCs in the space were drawn by features rather than safety, but believes that as the space has grown, smarter investors focused on the important values for becoming commercially viable have emerged. Despite his disagreements, he praises Starsky for being a pioneer in the space of driverless trucking and doing a lot of great work for the industry. Whatever the reasons are for the sun setting on Starsky, the startup has left a legacy in autonomous trucking.

Lessons Learned

Plus’s strategy for market positioning and rapid development gave the company a clear advantage in a crowded space that saw a shakeup and consolidation. Though the company is based in the US, it is set to dominate the Chinese market this year and soon expand further outward. In America, the company is still undergoing testing and pilots though it has plans to roll out its PlusDrive system this year. Had Plus not gained a significant foothold in China, they might be behind many competitors that are also launching their autonomous trucking products. Go-to-market strategies should begin in the best locale, not just the closest or easiest to access. Don’t skimp on market analysis and assume where you are is the right place to gain a foothold; it could be anywhere. Regulation, ease of relationship building, the concentration of human capital, networks, and a host of other factors can make one geography more attractive than another, especially with something as complicated and surrounded in red tape as autonomous vehicles. Additionally, Plus avoided diving deep into the nitty-gritty aspects of manufacturing its own vehicles or being a fleet operator and simply created software products that could be integrated into new vehicles or outfitted on old ones. This reduced friction in actually getting into trucks and working with fleet operators, as it could forge partnerships with companies that already had the expertise and experience to develop what it needed. Plus could then solely focus on making its autonomous driving system better rather than worrying about building a truck meant to handle it or diving deep into the more logistical side of trucking. If someone else is already doing exactly what you need, why recreate the wheel? Many startups try to take on too much to capture a greater market share, but there is a tradeoff with how thinly you spread yourself and the greater number of competitors. If an opportunity is already enormous as it is, carving out a niche and focusing solely on that piece of the ecosystem narrows the scope of competition and concentrates energy into a single vision. The tip of a spear should be sharp and thin, not wide and blunt.

In a very slowly advancing, capital-intensive industry, you are likely to be at the mercy of investors. When bigger companies are collectively throwing billions of dollars at a problem, bootstrapping is like bringing a knife to a very large gunfight. It’s almost impossible at that point to win without money constantly pouring in for a very long time, which is why many companies raise massive rounds in certain industries. You have to sustain investor belief in your company by moving towards a vision that will make the long burn worth it for everyone with a stake in the company. While some opportunities in the present may be tempting and could be profitable, striking away from the vision could kill the startup in the long term by shaking faith in the company’s foundation. After all, if a founder is having doubts and is branching out to find ways to make money, why should I believe that in ten years the current path will work? The autonomous vehicle industry is firmly planted in the vision of being free from human intervention, and Starsky diverged from that belief as a company. While remote driving has the capability to make trucking a safer, more efficient industry, many see it as just a bridge towards a future without drivers. The focus on remote driving had them fall out of favor as tides turned, and autonomous trucking moved on with significant progress towards the future Starsky didn’t expect to see for a very long time. In industries like this, build solutions in line with a vision investors see long-term potential for the industry in, even if it’s a bit of a pivot from what you set out to do. If you set out on tangents, even if to help add to your own runway, make them move your company towards the vision rather than perpendicular to it.

Scaled and Failed