DeepRoute.ai gets $100 million charge from auto investor

The company is one of the few from an increasingly crowded autonomous driving sector to raise new private funding in the last two years, as many of its peers make IPOs instead

Key Takeaways:

• DeepRoute.ai secured $100 million in new funding from an automaker, reportedly Great Wall Motor, following a $300 million round in 2021 that valued it at $1 billion
• The autonomous driving technology company’s map-free product promises lower costs compared with traditional map-based systems

By Hugh Chen

Two distinct approaches have emerged in the autonomous driving race that has spawned dozens of startups worldwide. One relies on high-precision maps to keep cars and trucks safely on the road, while the other operates without them.

While map-based systems remain dominant due to their navigational precision, map-free solutions, which typically cost less, have suddenly begun creeping up in the passing lane over the past year. Tesla’s (TSLA.US) Full Self-Driving (FSD) system, which primarily relies on real-time sensing and AI, has been a notable driver of this trend.

On the road to autonomous driving in China, the map-free approach is also gaining increasing momentum. DeepRoute.ai, a five-year-old Shenzhen-based startup, switched to map-free solutions in its early stages, and now touts its technology as a key differentiator compared with many of its rivals.

DeepRoute.ai made headlines last week with its announcement of a fresh $100 million in new funds from a domestic automaker, using the occasion to once more highlight the advantages of its map-free approach. While the company didn’t name the investor, multiple media reported that that state-run Great Wall Motor, a China joint venture partner with BMW, is behind the funding.

DeepRoute.ai was founded in 2019 by Maxwell Zhou, who previously worked at Baidu (BIDU.US; 9888.HK), one of China’s earliest companies to focus on autonomous driving technology. The company was valued at over $1 billion in 2021 when it raised $300 million from investors including Alibaba, Yunqi Capital, and Fosun Capital.

No valuation was given for the latest funding round, though a rocky road for recently listed Chinese autonomous driving companies suggests the latest valuation could be stagnant or even falling. As a case in point, shares of light detection and ranging (LiDAR) technology maker Hesai (HSAI.US) have lost about three-quarters of their value since the company listed in early 2023, shaving nearly $2 billion from its market value.

DeepRoute.ai’s new $100 million investment stands out as one of the few major new fundings in China’s autonomous vehicle sector over the past two years. The recent scarcity of deals reflects growing investor impatience with autonomous vehicle startups that have struggled to commercialize their technology.

In the absence of strong support from private money, many autonomous driving companies have turned to public markets lately for new funding. WeRide (WRD.US) and Horizon Robotics (9660.HK) made their market debuts last month, while others including Pony.ai, Zongmu, and Momenta have also filed for their own listings.

Map-free approach

While the map-free approach offers lower costs by eliminating the need for constant map updates and maintenance of massive mapping databases, it faces greater technological challenges. That’s because such an approach relies heavily on sensing and decision-making in real time, necessitating advanced algorithms and robust sensor systems to accurately interpret complex environments without the guidance of pre-existing maps.

After starting out with a more traditional mapping approach, DeepRoute.ai shifted gears and launched its map-free solution in March 2023. To convince skeptical automakers, CEO Zhou has frequently highlighted his company’s demonstration method, which involves transporting test vehicles by flatbed truck to client-specified locations and immediately conducting autonomous driving tests without prior mapping. Through such methods, Zhou hopes to prove his system’s ability to navigate unfamiliar environments in real-time.

Business progress has been modest, however, with only one vehicle model currently equipped with DeepRoute.ai’s system. Zhou has said two more models will reach consumers by year-end.

While the company hasn’t disclosed any financials, the fact that its technology has only been incorporated into one car model so far suggests its revenue is still quite small. Data from other firms in the sector also shows the industry remains in early stages. WeRide, which operates fleets of robotaxis and other autonomous vehicles, reported revenue of 401.8 million yuan ($55.8 million) in 2023, with a loss of 1.9 billion yuan that year.

Like its many peers, DeepRoute.ai is chasing an autonomous driving market that is likely to have huge potential for anyone who can commercialize their technologies. According to third-party research cited in WeRide’s prospectus, the global autonomous driving market is projected to reach $1.75 trillion by 2030, with China’s market alone expected to hit $639 billion.

But challenges loom, including growing competition from big companies with far greater resources. Adding to the pressure, major automakers are accelerating their in-house development of autonomous systems, with industry leaders BYD (1211.HK; 002594.SZ) and Xpeng (XPEV.US; 9868.HK) making significant investments in their own solutions. In a similar move, Volkswagen (VOWG.DE) agreed last year to pump a collective 2.4 billion euros ($2.56 billion) into recently listed Horizon Robotics and a new joint venture between the two.

DeepRoute.ai will use the latest $100 million investment to develop its Vision Language Action Model (VLA), a next-generation autonomous driving system. CEO Zhou characterizes VLA as the equivalent of having an experienced driving instructor at the wheel, contrasting it with rival systems that he likens to having an instructor merely supervising from the passenger seat.

On the commercialization front, Zhou is aiming to expand DeepRoute.ai’s market presence significantly in 2025, partnering with automakers to launch more than 10 new vehicle models. The company has set a target of equipping nearly 200,000 vehicles with its advanced driving assistance system (ADAS) by the end of 2025, which would represent a major increase from its current deployment of 20,000 vehicles.

DeepRoute.ai is pursuing a dual-track strategy, developing Level 4 (L4) autonomous robotaxis while also growing its core Level 2 (L2) driver-assistance business. The autonomous driving scale ranges from Level 0 (L0) to Level 5 (L5), with L5 representing fully autonomous driving. While the company began by developing L4 systems using mapping technology at its launch in 2019, like many of its peers it pivoted in 2022 to focus on L2 as well to start generating revenue more quickly.

DeepRoute.ai has also announced plans to expand internationally, targeting markets in Europe, Japan, and South Korea in the near future.

The company faces the same fundamental challenge as its peers: proving that its map-free solution can achieve widespread acceptance among automakers. Despite growing industry interest, the approach remains contentious. NavInfo CEO Cheng Peng recently ignited an industry-wide debate by questioning whether companies claiming to offer map-free solutions are truly operating without any map data at all.

At the end of the day, willingness by automakers to fully embrace the technology, and whose technology performs best, will be deciding factors in separating the wheat from the autonomous driving chaff. Like its peers, DeepRoute.ai must focus on commercialization as the sector can no longer rely on investors to provide unlimited funding. Profitability remains a critical challenge – though one that most companies look unlikely to solve anytime soon.

To subscribe to Bamboo Works free weekly newsletter, click here