news

한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina

21st century business herald reporter yi silin reports from beijing

thirteen days ago, at the second quarter earnings call, leapmotor chairman zhu jiangming announced that in order to focus more on algorithm research and development, leapmotor's intelligent driving algorithm research and development would be separated from the electronic and electrical architecture research and development to establish a professional intelligent driving research institute.

21st century business herald learned from multiple independent sources that the head of leapmotor's intelligent driving algorithm research and development is about to take up his post. he uses the pseudonym "cao bo" within leapmotor and is responsible for coordinating the development of intelligent driving algorithms and reporting to wang yaonong, vice president of leapmotor and head of intelligent driving.

leapmotor rarely hires technical experts from external teams. most of the current mid- and senior-level technical talents are technical personnel brought over by zhu jiangming from dahua international, including cao li, senior vice president of vehicle development, and wang yaonong, vice president of intelligent driving. cao bo, who is about to join the company, is a rare technical r&d talent hired from outside in recent years.

"wechat for business has not yet displayed cao bo's specific information." an insider close to leapmotor told the 21st century business herald reporter that the person in charge may have worked in the us r&d team of xiaopeng motors before. he speculated that after cao officially joined, his rank would most likely be above deputy general manager, two levels lower than the senior vice president of the intelligent technology research institute.

in the algorithm development team led by cao bo, leapmotor has recently set up a separate end-to-end r&d team to match ideal auto. "the entire end-to-end team currently has about 10 people," a headhunter serving leapmotor told the 21st century business herald reporter, "but the organizational structure is not yet clear and recruitment is still ongoing.

the 21st century business herald contacted leapmotor to verify the above information. as of press time, the other party has not responded.

in leapmotor’s intelligent driving planning, algorithm research and development occupies an important position.

in order to reserve more computing power, tesla set up the dojo supercomputing center, and xiaopeng motors joined hands with alibaba cloud to set up "fuyao", but zhu jiangming believes that the computing power of supercomputing centers in china is excessive, "heavy assets, and depreciation is also fast", in this regard, leapmotor prefers to "rent computing power rather than invest in computing power". therefore, compared with the heavy investment in supercomputing centers, leapmotor chose to expand the size of the intelligent driving team and develop its own underlying algorithms.

"recently, leapmotor has released nearly 200 hcs (headcounts) in the algorithm department alone, and has also poached a number of p7-level engineers from the ideal intelligent driving team," the headhunter told the 21st century business herald reporter. according to zhu jiangming's plan at the earnings call in august this year, the size of the intelligent driving team will be expanded to about 500 people by the end of 2024. at present, the number of ideal, nio, and xpeng intelligent driving is about 900, 1,500, and 2,000 respectively.

in the intelligent business, leapmotor's past strategy was to be a "follower", which also meant that "it was difficult to widen the gap," zhu jiangming reflected at the second quarter earnings conference this year.

leapmotor was founded in 2015, but its intelligent team did not take shape until 2020. that year, leapmotor was still worried about selling cars and had no spare energy for intelligent r&d. the intelligent team was very small, with only a few dozen people, and was placed under the "electronic and electrical department" as a secondary department.

after listing on the hong kong stock market and the sales of extended-range models took off, leapmotor has the confidence to accelerate its intelligentization. since last year, intelligence has begun to affect users' car purchase decisions, and the route has become clearer: the intelligent driving routes of various companies have gradually moved from pure visual solutions and high-precision map solutions to map-free solutions, and at this stage, they have begun to catch up with tesla to develop end-to-end intelligent driving solutions.

in order to catch up faster, leapmotor not only invests money and people, but also accelerates cooperation with relevant hardware suppliers.

by the end of 2023, the size of leapmotor's intelligent driving team will reach about 200 people, with its core being self-developed algorithms and solutions based on high-precision maps.

in the first half of 2024, leapmotor's r&d expenses on intelligence increased by 50% year-on-year, "mainly due to increased investment in intelligent driving," zhu jiangming further explained at the second quarter earnings conference. he added that in 2024, leapmotor's r&d investment will maintain a growth rate of 50% throughout the year.

compared with the past, although leapmotor has always boasted of self-developed products in all fields, its r&d investment is relatively low. from 2022 to 2023, leapmotor's r&d investment in two years is about 3.33 billion yuan, while ideal's r&d investment in the same period is 17.366 billion yuan.

since the beginning of this year, leapmotor first signed a strategic cooperation agreement with hesai, and then signed a strategic cooperation agreement with ambarella semiconductor, a chip company focusing on ai visual perception. at the same time, two of its papers were selected for top international academic conferences. leapmotor is accelerating its catch-up in intelligent driving.

but the latest and most fundamental change is organizational. 21st century business herald learned that in the first half of this year, leapmotor established the intelligent technology research institute, which was upgraded from the original electronics and electrical department. the person in charge is senior vice president zhou hongtao.

previously, leapmotor's first-level departments only included the three-electric departments, such as battery, electric drive, and electronic and electrical. according to a former leapmotor employee, in the organizational system of other new forces, the three-electric departments are generally placed under the vehicle department and are third-level departments. in his opinion, the original leapmotor "is not like a car company, but more like an electronics factory."

after the establishment of the intelligent driving research institute, "zerorun's intelligent driving business will usher in a new starting point this year," a headhunter serving zerorun commented to 21st century business herald. he said that currently, about 10 r&d personnel join zerorun every week, and the heads of various departments of the intelligent technology research institute will also take up their posts one after another, and offices will be established in beijing and shanghai in the future.

zhu jiangming's goal is to "strive to have leapmotor's smart driving enter the first echelon by the end of 2025."

at present, leapmotor has realized high-speed nap and urban nac intelligent driving functions, and will carry out ota upgrades on the existing l2+ assisted driving functions at the end of july. further functional optimizations will be launched in the second half of the year, including point-to-point commuting capabilities on open roads and parking lot memory parking, etc. it also plans to launch urban intelligent function nap based on the end-to-end intelligent driving big model within 2025.

but other car companies are moving faster in intelligent driving. in july, ideal launched the map-free noa to all ad max users, which can be used across the country. its official announcement also stated that it will launch an end-to-end intelligent driving solution to all users by the end of this year at the earliest and the beginning of next year at the latest.

leapmotor currently has a long way to go to catch up with the first echelon. however, tesla has led a new end-to-end technology trend, which has brought some opportunities to overtake others.

the best example is ideal. before this year, ideal's intelligent driving capabilities lagged slightly behind those of xiaopeng and huawei. however, after launching the map-less city noa in july this year and launching the end-to-end version of intelligent driving with thousands of people in internal testing in august, it received unanimous praise.

leapmotor also wants to achieve overtaking in intelligent driving through end-to-end technical architecture.

some investors also questioned leapmotor's input-output ratio in high-end intelligent driving, after all, it is not easy to do high-end intelligent driving. "leapmotor sells well in third- and fourth-tier cities, but whether intelligent driving can become an important factor in users' purchasing decisions in third- and fourth-tier cities remains to be seen."

however, the trend towards intelligence and ai is clear. a headhunter in the automotive industry who once served ideal auto believes that leapmotor’s current intelligence is very similar to the situation when ideal auto first started to focus on intelligent driving three years ago. “in fact, the difference between everyone’s algorithms is not that big. it’s just a matter of who is more diligent and makes more decisive choices. on the contrary, the underachievers don’t have too many burdens.”