GM’s Autonomous Car Impresses, But Big Hurdles Remain.
“Courageous of GM to allow investors to peek into its AV world. The ride was entertaining. It mostly worked, and will surely improve, but we believe that the last 1 per cent will be extremely difficult to perfect”
General Motors’ demonstration of its commercial autonomous vehicle (AV) capabilities and its declared aim to launch within 2 years was received positively by many investors, although the most negative reaction suggested GM’s plan was far too aggressive.
GM’s demonstration in San Francisco to the media and investment bankers used a Cruise 2nd generation technology Chevrolet Bolt on a 2-mile trip through busy streets. The trip was variously described as a bit defensive and slow, or like being driven by a tentative 15-year old learner driver.
GM said it expects to deploy fleets of autonomous cars in dense urban areas by 2019. It didn’t say where, but GM is currently testing 3rd generation self-driving vehicles in San Francisco, Scottsdale, Ariz., and Warren, Mich. GM said it expects cost-per-mile of under $1 by 2025, the key to achieving profitable scale. The current cost in similar cities is $2.5, according to GM.
Morgan Stanley was the most negative investment bank of those polled.
“It was courageous of GM to allow investors to have a private peek into its AV world. The ride was entertaining. It mostly worked, and will surely improve, but we believe that the last 1% will be extremely difficult to perfect. GM’s urgency speaks louder than its targets,” said Morgan Stanley analyst Adam Jonas.
As autonomous cars are developed, the remaining problems get harder to solve, not to mention the need to develop a legal framework. It would be hugely impressive to develop an AV which worked 99% of the time, but nobody is going to buy one that is not 100% reliable.
Barclays Equity Research said it had heard GM’s targets were going to be unrealistic, similar to plans presented by Tesla, but it emerged from the demonstration believing the technology was real, with a fail-safe mode monitored by a control center.
“It was impressive. There no disengagements. While the vehicle drove conservatively – taking time, strictly obeying speed limits, checking carefully at both sides before turning at an intersection – we’d also note that the driving environment was quite complex,” Barclays Equity analyst Brian Johnson said.
Investment researcher Jefferies said the demonstration showed GM and its Cruise technology are leaders in the race for AV mobility on demand, but it was less sure that it would be able to dominate this new market. The demonstration showed GM’s lead.
“Riding in the AV Bolt was slightly jarring at times, like being driven by a hesitant learner, but one who can only improve rapidly. Road and traffic conditions were challenging, safety was prioritized and it felt at times that the main threat came from hesitant or annoyed human drivers in other cars,” Jefferies analyst Philippe Houchois said.
The news GM would launch in 2 years surprised Houchois, who also pointed to barriers like regulation. He was also worried that GM might find profits hard to make as other companies entered the market.
“Our main concern remains that industry fragmentation and competition may drive pricing to lowest marginal cost levels, resulting in attractive services for consumers but relatively low ROIC (return on invested capital),” Houchois said.
Investment researcher Evercore ISI didn’t think the 2019 launch date was cast in concrete, with much work still to be done, while it also wasn’t clear yet if GM had a lead on the competition.
Critical path to safety
“(The Gen 3 car) is also fitted with the redundancies, parallel compute and dual power supplies which GM believes are needed in order to go to the market without a driver. With the vehicle and hardware theoretically in place, GM believes the critical path to market from here is improving the vehicle’s safety,” Evercore ISI analyst George Galliers said.
Galliers marked the performance of the Gen 2 vehicle on the San Francisco route as 5.5 out of 10.
“Clearly there is still work to do,” Galliers said.
But Morgan Stanley’s Jonas believed GM’s timetable was too aggressive and the rollout of AV’s will be much slower.
“In fact, the more we study and experience this topic, the more convinced we are that the inflection point of the adoption curve will be far later than the market may anticipate,” Jonas said.
“We believe that stock market expectations for commercialization at scale and profitability may need to be dialled back significantly. GM is targeting mass commercial deployment by 2019. In our opinion, if even 1% of global miles travelled were executed by cars without a steering wheel by 2030, that would be an extremely impressive number. We are talking with many clients who expect 50% Level 5 (total computer control) penetration. While we cannot be specific, we’d add 10 to 15 years to such levels of penetration. Seriously,” Jonas said.
Amongst hurdles to be jumped, Jonas pointed to the cost of the Gen 2 or 3 at between $250,000 and $300,000, and he worried about operating costs. Also the ethical, legal and regulatory issues surrounding artificial intelligence might be the biggest hurdle of all.
UBER, Waymo pressure
Reuters Breaking Views columnist Antony Currie said GM is the only player that has been testing self-driving vehicles in bustling cities, and that puts pressure on likely competitors Uber and Alphabet’s Waymo, as well as Ford and others. Currie pointed to the possible impact on GM’s bottom line from this business, which could achieve margins of between 20 and 30%.
“Much could go wrong. But shareholders and competitors alike now have to reckon with a company that is no longer their parents’ General Motors,” Currie said.