Specific scenarios (ODD) or even fully autonomous driving, when will it enter the real commercial mass production landing? Everyone and every business in the automotive industry has their own answer.
In fact, as early as 2009, a classic Audi A6 Avant had achieved automatic follow-up in congested road scenes at a speed of 70 km / h under the supervision of Audi engineers (disengaged), and could automatically change lanes under manual guidance.
At the time, Audi called it: Piloted Driving. Eight years later, in 2017, Audi officially released the first L3 level automatic driving system that can be mass-produced, but unfortunately, due to various factors such as regulations, this mass production system eventually "died".
By 2018, Waymo, a global leader in autonomous driving startups, announced a long-term strategic partnership with Jaguar to order 20,000 I-PACE models to expand its autonomous driving test fleet in the United States. To this end, Waymo also built a new retrofit (quasi-front) factory for L4 autonomous driving a year later.
Waymo's initial goal was to roll off the production line with 20,000 self-driving modified vehicles by 2020. But the results are certainly an impossible task, and the data shows that waymo will still have fewer than 1,000 test vehicles in the United States by the end of 2021.
As the most aggressive Tesla for the private autopilot market, it was originally planned to release the official version of the full autopilot FSD in 2020, but until the end of 2021, it is still in the stage of small-scale user internal testing.
By April 2021, Waymo's first CEO, John Krafcik, suddenly announced on his personal LinkedIn platform that he had officially left Waymo and remained only as an advisor.
John Krafcik previously said in a public interview, "It's an extraordinary chore[on autopilot]." It's more challenging than launching a rocket... Because it has to be done safely over and over again. ”
One
In the past few years, the field of autonomous driving has received some very subtle changes while it has received almost fanatical attention from capital. For example, some companies began to "withdraw", and there began to be a small number of mergers and acquisitions, some companies began to choose special scenarios to find "cash flow", and the remaining companies insisted on moving forward.
Optimists believe that according to data given by government agencies, 93% of road traffic accidents are caused by the human error of drivers. In theory, autonomous driving systems can circumvent these problems.
At the same time, based on vehicle-road collaboration, automatic driving and electrification, travel will be more energy-efficient, urban traffic congestion can be better managed, and automatic driving can even optimize the use of road space.
But John Krafcik admits that he and his colleagues rely on their experience in the auto industry to judge how fast autonomous driving will grow. "The idea was that if we had a prototype, it could be mass-produced in a few years."
As a result, it seems to have been lower than expected.
However, the capital market is still pushing the "forward" of autonomous driving, but it has also allowed more people, especially non-industry practitioners, to have a clearer understanding of the commercialization of autonomous driving.
In April 2021, Tucson Future achieved a U.S. stock IPO, becoming the first stock of self-driving trucks. Then, over the past 8 months, the company's stock price has gone through a roller coaster, with the $40 offering price rising to around $70 at one point, but then falling sharply and falling below the issue price several times.
Tucson's future prospectus discloses that the cost of L4 self-driving trucks is still very high, and it may be necessary to find more financing solutions to help the company's users in the future or buy (and then rent) vehicles to operate on their own.
In addition, until mass production of custom self-driving trucks by 2024, the company said it also needs huge capital expenditures to maintain and upgrade its existing fleet. "Any investment is highly speculative and may result in the total loss of the investor's investment."
This week, After a series of executive turnovers, Tucson Future announced that it is ready to sell its business in China to focus on the US market due to a series of data security regulations and other issues. Meanwhile, the company's market capitalization has fallen to around $3 billion, down nearly 70 percent from its IPO price.
According to the financial report data, Tucson's future revenue from 2018 to 2021 is $0.9 million, $710,000, $1.84 million and $6.26 million, respectively; net losses are $45.03 million, $84.88 million, $178 million and $732 million, respectively, and the losses continue to increase.
In early November 2021, Aurora's share price, known as the world's "Robotaxi first stock", soared 51% on the day of listing, closing at $17.11 per share, with a market capitalization of nearly $20 billion. At the same time, it has once again set off a round of automatic driving "warming tide".
"I don't shy away from the possibility that companies may need to raise more capital." Aurora's co-founders said that a larger amount of capital is needed to continue to invest to maintain the commercialization of the landing, for example, for mass production-level software and hardware improvements, and even the fixed asset investment of the pre-operating fleet.
According to the plan, the company will launch the Aurora Horizon truck self-driving subscription service by the end of 2023, and expects to launch the Aurora Connect taxi service a year later, but it is not expected that at least until 2024 there is a glimmer of profitability.
In the past 5 months, Aurora's stock price has fallen from the highest point to about $5, and the latest market value is only about $6 billion, down nearly 70% from the highest point.
At present, the annual R&D investment of leading companies like Aurora is still at a high level of hundreds of millions of dollars, and the amount is still increasing as commercialization approaches.
For example, Tucson's future R&D spending in the fourth quarter of last year was as high as $82.393 million, compared with only $31.799 million in the same period last year. Aurora lost $214 million in 2020, of which up to $179 million was spent on research and development.
Interestingly, Musk is also "waiting" for the last promise to be fulfilled, and in an investor conference call in the fourth quarter of last year, he predicted that "the final version of the FSD will be ready this year, and I will be shocked if we can't achieve full autonomous driving that is safer than humans." ”
This means that the prospect of commercialization must and still needs to continue to "burn money".
Two
This week, GM announced a $2.1 billion acquisition of its stake in self-driving subsidiary Crusie from SoftBank Group, a subsidiary of the Vision Fund, and an additional $1.35 billion in cruise (the amount of a phased investment previously signed by SoftBank Group).
After the transaction closes, GM will hold about 80 percent of Cruise. However, this is not GM's "active" will. It is reported that the cost of debt default insurance for SoftBank Group has risen, hitting the highest value in two years, and its bond yields have also climbed.
At the same time, the market value of the technology companies held by SoftBank Group in the past six months has also plummeted, for example, Alibaba and Didi, which have fallen by 35% and 64% respectively since the end of last year.
"We're definitely going to sell some of our assets." SoftBank Group, which began last year looking to sell Arm (but Nvidia ultimately failed), now plans to list the chip IP leader in the mobile and automotive sectors independently. This also means that SoftBank does not have enough funds to continue to "bless" Cruise.
As we all know, cost is the key to whether L4 autonomous driving can be truly deployed at scale (and at the regulatory level, relaxation is already the main theme. Previously, for example, U.S. regulators planned to allow vehicles to be free of manual controls, such as traditional steering wheels, brake accelerator pedals, etc.).
Previously, Mobileye expected that the cost of L4-level hardware and software solutions would be expected to drop below $5,000 by 2025. The CEO of GM hopes to "graft" Cruise's autonomous driving capabilities to GM, for example, Ultra Cruise (which supports automatic and on-demand lane changes, achieving 95% of driving scenario coverage) and provides users with a monthly subscription fee.
Or because of the "immaturity" of L4 autonomous driving, Cruise's former CEO was removed at the end of last year for insisting on pursuing an independent IPO. In fact, whether it's Waymo or Crusie, "executives have promised roadmaps with different time nodes, but none of these milestones have been achieved." ”
In Waymo's case, it has been constantly adjusting its strategy, from the early days of seeking to sell lidar to reduce internal costs (which ended up in vain) to the two-line layout of commercial trucks to self-driving to obtain cash flow.
On the one hand, from a hardware perspective, the objective challenge facing all businesses is whether the sensor can work properly in all weather conditions. At the same time, vehicles are still very cautious about the unpredictable behavior of pedestrians.
On the other hand, even though Waymo is doing its best to reduce hardware costs, or even invest heavily in its own hardware, the company estimates that the cost of hardware and software modification on the basis of existing vehicles is still as high as tens of thousands of dollars (this does not take into account the amortization of upfront research and development costs).
And even for industry leaders like Waymo, who are backed by a financial backer (and Google is a parent company), lack of money is still the primary problem. Previously, the company had raised a total of $5.5 billion in foreign financing twice.
In order to share the cost, GM decided to introduce Cruise's autonomous driving capabilities to the private market, and increase the software payment revenue of new car purchases as much as possible, thereby maximizing the dilution of the cost of upfront technology research and development. "You know, whether it's Tucson Future or Aurora's stock price, it's already deterring many companies."
For these companies, this timing is correct, on the one hand, the commercial landing has begun to dawn, and the early stage of technology research and development has entered the key stage of "quantitative change to qualitative change". On the other hand, investors in the primary market have entered the recovery period and are in urgent need of secondary market takeover.
Now, almost the vast majority of self-driving companies have pointed to the "sweet" time point of mass production on a large scale to 2024. This means that in the next two to three years, sufficient funds, deep integration with OEMs, the gradual promotion of policies and the release of specific scenario needs will be carried out simultaneously with the elimination and integration of the industry.
No matter which track it is, constantly burning money and waiting for the dawn of commercialization is the current situation of the industry. "Both anxious and full of expectations," industry insiders admitted, the hope is still there, but do not know who will "live" to the end.
However, a clear change is that car companies are becoming commercial protagonists. From Mercedes-Benz's L3 to Volvo's L4, it will promote the autonomous driving of specific scenarios into the private consumer market. More L4-level autonomous driving companies are also looking to get a piece of the front-loading mass production of car companies.
For example, after receiving investment from SAIC motor and receiving orders for the development of front-loading projects, Self-driving startup Momenta has established a joint venture with BYD to focus on the joint development of high-level intelligent driving solutions.