By Marc Amblard, Founder & Managing Director, Orsay Consulting
The ongoing crisis is expected to be have a severe impact on the automotive industry. The significant drop in sales in recent months and a lengthy projected economic recovery has accelerated pre-crisis trends in the mobility space. The shift towards autonomous driving will be significantly impacted, whereas the electrification trend may slow down marginally overall — and actually be accelerated in some countries.
Eighteen to twenty-four months ago, the industry at large recognized that autonomous vehicles (AV) would not be deployed across multiple use cases in the early 2020s as previously anticipated. The tech was simply not maturing fast enough. Projects were massively scaled down or postponed indefinitely, e.g. Waymo’s commercial autonomous ride hailing service in Phoenix and Cruise’s in San Francisco, or even cancelled, e.g. Audi’s A8 Level 3.
Startups began having difficulties raising more funds, some opting to pivot away from Level 4 applications to favor other use cases for their tech. This trend was apparent for companies across the whole spectrum of the AV landscape, whether hardware or software. Incumbent players have been increasingly engaged in cooperations among peers or with tech companies in order to access technology with reduced risk exposure, e.g. Hyundai-Aptiv, or Ford-Argo-Volkswagen.
The crisis impacts AV funding and startups’ livelihood
OEMs and suppliers are adjusting their capital allocation, prioritizing projects that can positively impact their top or bottom line in the next two to three years. However, players that really have a long term perceptive, such as Toyota, continue to invest in ambitious, multi-annual projects in-house or with partners.
Some incumbents have announced they were getting out of partnerships and into new ones. Daimler and BMW paused their AV-focused collaboration, however Daimler now works with Nvidia (though scopes differ). Volvo and Veoneer are dismantling Zenuity, their 50-50 JV — the OEM picks up the AV SW activity and Veoneer the ADAS-focused resources — as Volvo just engaged in a partnership with Waymo. Incumbents are fine-tuning the collaboration strategy to best serve their needs and means.
An increasing number of startups are reaching the end of their financial runway, with no additional funding in sight — startups typically raise funds necessary to last for about 18 months. Some companies have shuttered, like autonomous trucking startup Starsky Robotics.
Venture capital and private equity firms, which have accumulated very large funds overall, now tend to use their “dry powder” to double down on their most promising portfolio companies rather than invest in new ones. As a result, early stage startups will have much more difficulties getting off the ground. This will result in a healthy concentration, but will stifle the emergence of new ideas. This is recently benefited Waymo, which has raised $3B this year.
This money tightening has created opportunities on the buy side, which will benefit startups with advanced tech and/or strong teams. Zoox, with its moonshot mission to develop a vehicle, a full AV stack and a ride-hailing service, was reportedly having difficulties raising additional funds beyond the $1B already collected. The project had become overly ambitious. Zoox was just acquired by Amazon for $1.3B, at significant discount vs. previous valuation of over $3B. Similarly, Caterpillar acqui-hired Marble Robot, which had raised $10M to develop a delivery robot, and Cruise just bought Astyx, a radar startup of which ZF previously controlled 45%.
Fewer players will have to drive the effort to Level 4
Many startups that provide picks and shovels to the AV industry, e.g. those working on Lidars, image annotation / labelling, localization / mapping or simulation, are suffering. There are just too many of them (e.g. over 50 for Lidar) for a decreasing number of potential customers. The most mature startups have already concentrated much of the investment pool, e.g. training and validation data startup Scale which raised $100M at $1.4B valuation mid-2019.
The less “lucky” startups are quickly pivoting to generate revenue in order to extend their runways. They target other use cases either within ground mobility or in other industries. Alternative applications for their tech include trucking, sidewalk delivery robots, automated guided vehicles, drones, automation of mining or agricultural vehicles, surveillance of public spaces, surveying, etc. — Level Fives supplies identified 100 application for Lidars.
Similarly, OEMs will likely shift some Level 4 R&D resources towards ADAS / Level 2. The latter tech is mature and can therefore contribute to the top line in the near term. Most OEMs in this situation will maintain a L4 development team at least to stay abreast of progress. This is the case for OEMs with a “fast follower” strategy, akin PSA.
AV use case priorities reshuffled
We are facing the combination of two key factors in the deployment of AVs. First, reaching tech maturity has proven much more difficult than anticipated, in particular for the initial core use case, i.e. autonomous ride hailing in urban areas. Second, the COVID-19 crisis has significantly boosted e-commerce, which will result in more ground transportation, especially for the last mile.
In what order is AV tech likely to be deployed now? This will depend on tech maturity for each use case, the infrastructure and the regulatory context.
A less visible set of use cases provide low hanging fruits for the industry in the short term, as they are technically less demanding. They include industrial logistics robots such as Autonomous Guided Vehicles, or autonomous vehicles for local hauling, e.g. moving trailers at logistics centers or hauling luggage carts on airport tarmacs. Here, these vehicles operate in a work environment on private property, which make things easier. Whereas this set of use cases may not result in a massive market, they are already happening now. EinRide is a good example here.
The next wave of AV deployment will likely address middle and last mile deliveries, where the speed is limited and smooth driving behavior is not critical. In addition, some deployments can serve exclusively pre-set routes for which the system can be more easily optimized. Gatik and Neolix are good examples of these applications. One major issue here will be the business case. For reference, food delivery is currently a money losing business.
Next, we will likely see autonomous trucks with a focus on highway driving. Several truck OEMs and startups have been focusing on this use case, e.g. TuSimple. More recently, a couple of large AV startups, which previously prioritized passenger car automation, have increased their interest in autonomous trucking. This is case for Waymo and more recently Aurora Innovation, which now seems to put trucking ahead of passenger vehicles. The absence of a driver combined with high utilization rate will likely make for an easier business case.
As the tech matures, the cost of sensors and compute platforms drops and significant learning is gained from early AV deployments, the time will come for a significant deployment of autonomous shuttles (beyond currently existing way-points solutions) and ride-hailing. At that point, we may also see Level 4 technology carried across to high-end passenger vehicles for private use. By then, Level 2 / 2+ will have become widely deployed across the market.
The current crisis will impact the development and deployment of AVs on multiple fronts. An industry consolidation is unavoidable — and healthy — which will allow strong players to emerge. Goods transportation will become the leading target in the mid-term, whereas more efforts will be placed on deploying ADAS / Level 2 solutions in the passenger vehicle space. AV-related startup founders will need a very convincing value proposition to raise their first money!
This article was orginally published on the Orsay Consulting blog.