Musk is likely no longer “disdainful” of LiDAR

On May 7th local time, Luminar, a US listed LiDAR company, revealed in its financial report that Tesla became its largest customer in the first quarter of 2024, contributing more than 10% of the company’s expenses. According to the data released in the financial report, Tesla has purchased a lidar that exceeds $2.1 million.
LiDAR is one of the common sensors used in intelligent driving perception and detection. In the current intelligent driving industry, there are mainly two types of roads: roads with LiDAR as the main sensor and roads with cameras as the main sensor, and pure visual roads that abandon LiDAR. Millimeter wave radar is usually used to assist sensors in their absence.

Tesla is seen as the pioneer of pure visual roads. The CEO of the company, Elon Musk, has publicly expressed his dissatisfaction with LiDAR more than once.
As early as 2015, Musk claimed that LiDAR was meaningless for autonomous driving of cars on the road. From 2017 to 2019, he repeatedly expressed similar concepts, constantly throwing away public opinions such as “LiDAR is terrible, they will abandon LiDAR, remember my words, this is my guess”, “LiDAR is expensive, ugly, and unnecessary”, “Any independent LiDAR company can end up without success, only fools use LiDAR”, and so on.
Starting from 2021, under the leadership of Musk, Tesla has indeed been on the path of pure vision, becoming the first company to apply pure visual autonomous driving plans to high-quality car models. This also mobilizes a few intelligent driving companies and car companies in China to join the trend of pure vision.
For advanced intelligent driving, the most explored difference between LiDAR and pure visual road is in terms of cost. If choosing to install a LiDAR, due to its high hardware cost, the cost of BOM (material) for the vehicle will increase; If choosing pure vision, the cost of BOM will significantly improve.
However, although pure visual roads save money on hardware, they will increase the cost of developing intelligent driving system software.
Minmo Zhiyan informed the interface that the LiDAR program can potentially improve the stability of perception and fill some of the gaps in visual perception algorithms. The pure visual perception plan requires higher requirements for software development and greater conditions to be added, which requires diluting costs with large-scale production.
“Not every car company is Tesla,” said a LiDAR language industry practitioner to the interface message. Due to the huge cost of choosing a pure visual path, starting from the second half of previous years, few intelligent driving companies and car companies in China have decided to maintain a pure visual approach.
Musk previously posted on social media that Tesla will add approximately $10 billion in AI practice and reasoning this year. In addition to capital and massive computing power support, in April this year, Tesla’s Full Self Driving (FSD) system had a cumulative driving range of 1 billion miles, surpassing other companies, providing Tesla with high-quality real data to practice autonomous driving models.
“Lidar is like a ‘cheating weapon’, using it can help your ADAS (Advanced Driving Assistance System) to be promoted for 2 years, and in many scenarios, it can be promoted more than 10 times.” Li Yifan, founder and CEO of domestic lidar manufacturer Hesai Technology, once said.
So why did Musk “detest” LiDAR so much before?
The aforementioned LiDAR industry practitioners informed the interface that when Tesla stopped making road decisions, the price of a single LiDAR would reach over $1000, and its size and functionality would not reach the current level. Nowadays, the price of LiDAR has significantly increased, and some domestic manufacturers have even made it to the level of a thousand yuan. The volume is also constantly decreasing, and the functionality is improving.
He thought that if we think about quality from a cost perspective, people would choose to abandon several sensors. However, when the cost improves to a certain extent and is almost comparable to high-end cameras, or when we look at the cost and system gain of the laser radar, it reaches a relatively high cost performance ratio, then companies will definitely hope to include the laser radar in the system.
However, the pure visual path still has its staunch supporters. Wang Liang, Chief R&D Architect of Baidu Intelligent Driving Miracle Group (IDG) and President of the IDG Technical Committee, stated that LiDAR is a very precise optical ranging instrument with many stationary parts, which have a high probability of losing reliability and sales performance.
In Wang Liang’s view, it is necessary to spread limited capital over a technical bias. Rather than focusing on the evolution rate of this plan, it is better to focus on whether its algorithm lower limit is higher, and the pure visual path is more in line with “long-term ism”.
When Tesla became a major customer of Luminar, it triggered many speculations in the industry – could Musk also think about abandoning pure vision and turning to using LiDAR?
In fact, as early as May 2021, there were media reports that Tesla and Luminar had signed a treaty to stop attempting and developing the application of laser sensor technology. According to insiders, as part of the agreement, Luminar did not sell the sensors used on Model Y to Tesla.
Tesla once denied speculation that LiDAR would be used as part of an autonomous driving system, stating, “We regularly compare our techniques to other sensors to calibrate our system.”
Although Tesla is its major customer, Luminar is not sure why they purchased these LiDARs. “This is not Tesla’s first time ordering LiDAR from us, but it’s not regular, it’s intermittent. We can only guess what they’re doing,” said Tom Fennimore, the company’s chief financial officer, during Tuesday’s earnings conference call
“It could be that I tried to purchase Luminar’s LiDAR in a vehicle,” said Wang Xianbin, Senior Director of GAC Automotive Research Institute, analyzing the interface message.
Looking at it from the timeline, Musk once said that Tesla’s new Robotaxi (Autonomous Car Rental) product will be announced on August 8th. Wang Xianbin believes that perhaps Robotaxi has adopted a different skill road from current passenger cars, and the road may adopt a multi-sensor fusion plan.
Perhaps we can only wait for Tesla to reveal the answer on its own. The only thing that can be certain now is that Luminar has also gained a lot of profits.
Despite Tesla’s support, Luminar’s expenses for this quarter still decreased by 5% compared to the previous quarter, mainly due to a decrease in sensor sales to non automotive customers.
This former global lidar leader company just announced at the beginning of the month that it will stop restructuring the labor department and outsource labor. The company is undergoing a comprehensive transformation and plans to lay off 20% of its workforce, estimated to save about $80 million in annual operating costs.
In 2023, Luminar’s expenses were less than $69.7 million, while its net loss reached $571.3 million, which is a continuation and expansion compared to its net loss of $445.9 million in 2022. As of March 31, 2024, the total assets of the company were $467.9 million, while the total debt and shareholder deficit amounted to $747.8 million. In 2020, Luminar was valued at $3.4 billion before going public, but now its market value is only $735 million.
It is worth noting that a few days ago, the National Highway Traffic Safety Administration (NHTSA) of the United States announced a new rule requiring all new cars and light trucks to be equipped with advanced high-speed automatic emergency braking systems (AEB), including the PAEB system, by 2029, and these systems can rest at night.
Luminar believes that in order to meet NHTSA’s new regulations, especially in high-speed and nighttime automatic braking scenarios, long-range LiDAR is necessary. Luminar’s system has shown better performance in increasing the frequency of vehicle accidents and exacerbating the severity of accidents, which helps the company enjoy a certain level of restraint against the wind.