Born in 1999, Mobileye has experienced a total of 5 product iterations so far, with a cumulative sales volume of 100 million chips, and has held hands with several head car manufacturers such as Tesla, BMW, and Audi. No matter what dimension is judged, if Intel wants to let Mobileye go public alone, it is reasonable to say that there will be no lack of capital pursuit.
However, on March 7, Intel chose to "secretly" file Mobileye's listing documents with the US Securities and Exchange Commission at a valuation of about $50 billion at a time when the Russian-Ukrainian conflict was in full swing. However, given the current sluggish market conditions in the US stock market, the upcoming interest rate hike policy, and geopolitical tensions, Mobileye's actual valuation is likely to be greatly reduced.
In addition, the reason why Mobileye chose to submit a listing application through a "secret" way is that on the one hand, it can avoid exposing the core competitiveness such as profit model to opponents, and on the other hand, it may be precisely because its internal confidence in the road to listing is not confident.
The so-called "secret" submission is precisely that Mobileye takes advantage of the common confidentiality document provisions introduced in the US Start-up Promotion Act. The advantage of this is that Mobileye is able to ask major investors for a price without disclosing financial data, and to determine whether the company will eventually go public based on the stock price.
Paradoxically, the document is only for emerging growth companies that "grossed less than $1 billion in total annual revenue in the most recent full fiscal year." Even though this restriction is adjusted for "inflation every 5 years," Mobileye's full-year revenue for 2021 is close to $1.4 billion.
So why did Intel have to let Mobileye "secretly" submit listing documents at this juncture?
The Death of Intel
As the world's largest chip manufacturer, Intel's development is currently facing multiple factors.
In the short term, the first to bear the brunt is the shortage of chip raw materials caused by the Russian-Ukrainian conflict. In semiconductor manufacturing, special gases such as neon, krypton and xenon are the second largest consumables after silicon wafers, and Ukraine is the main supplier of these gases.
At present, 90% of the neon gas in the US chip manufacturing industry relies on imports from Ukraine, and if tensions in the region are not alleviated in the short term, the price increase of neon gas is inevitable.
Secondly, Intel's main PC chip business is also facing the dilemma of weak growth. Relevant data show that global personal computer (PC) shipments in the fourth quarter of 2021 fell by 5% year-on-year, which is the first decline in the PC market after six consecutive quarters of growth, which means that consumers' fanatical demand for PCs in the post-epidemic era has gradually slowed down.
In the long run, Intel's woes are tied to past decision-making mistakes, pressure from competitors, and strategic planning for the future.
Earlier, Intel gradually lost absolute leadership in the chip industry due to the frequent strategic mistakes of previous leaders. In 2020, when Intel was still "grinding foreign workers" on the 10nm process, AMD, as a latecomer, took the lead in developing the 7nm process; in order to share the risk, Intel even gave up the proud chip manufacturing business for a while.
In 2021, Intel's new CEO Henry Kissinger proposed the "IDM 2.0" strategy at the beginning of his tenure, announcing a $20 billion investment in two new fabs in Arizona, USA, aiming to revitalize Intel's leading edge in manufacturing business.
Since then, Intel has invested more and more intensively. In May 2021, Intel announced an upgrade to its fab in New Mexico, and in January, intel announced that it would invest another $20 billion to build two chip manufacturing plants in Ohio. Kissinger also said that the total investment in the future may be increased to $100 billion, and a total of eight factories will be built.
According to Intel's 2021 annual financial report, Intel's annual revenue was $74.7 billion, up 2% year-on-year, setting a revenue record for 6 consecutive years. At the earnings conference, Kissinger also attributed Intel's achievements to "the advantages brought by the IDM 2.0 strategy." However, it is worth noting that the earnings report shows that Intel currently has only $11.3 billion in cash flow on its books.
Just after the release of the earnings report, Intel's stock price ushered in a sharp decline. As of the date of this writing, Intel's stock price has fallen by nearly 20% compared with the highest point at the beginning of this year, and even been surpassed by AMD at one point. Relevant people said that if Intel's heavily invested production line cannot achieve sufficient utilization, it is very likely to become a huge asset burden.
Therefore, whether it is the rising price of raw materials brought about by the Russian-Ukrainian conflict, the sluggish growth of the PC chip business, or the huge investment in implementing the IDM 2.0 strategy, Intel urgently needs healthy cash flow to support the entire company.
Mobileye seems to be playing the role of "fire captain".
Mobileye's change
Much like Intel's dilemma, as a chip manufacturer that was once the premier chip manufacturer in the field of ADAS, Mobileye has encountered many problems in recent years that need to be answered.
The first is the problem of insufficient computing power of Mobileye's intelligent driving chip. At present, many car companies at home and abroad have chosen driver assistance systems with multi-sensor fusion perception as the main solution. This means that higher-level assisted driving requires more sensor data, and more and more sensors need more powerful chips to coordinate the overall situation.
This directly caused the chip manufacturers to roll in at the level of computing power. However, in the comparison of many products, Mobileye's chip computing power is difficult to call excellent.
(Table 1: Comparison of chip computing power of each manufacturer)
In 2019, Mobileye's then-mass-produced chip EyeQ4 reached 2.2TOPS, which was second to none at the time, with Weilai ES8, ES6 and EC6, and Xiaopeng G3 all using this chip. In 2021, although Mobileye's EyeQ5 has increased its hash rate by 10 times, it has been surpassed by competitors' products in the same period.
In January this year, Mobileye announced its future product plans at CES, and will mass-produce the Eye Ultra with a hashrate of 176TOPS in 2025. However, NVIDIA Orin, which has a higher computing power, will officially get on the train in the second half of this year.
Second, Mobileye's hardware and software architecture is highly bundled. Mobileye has clearly stipulated that partners cannot develop their own visual perception algorithms while using Mobileye chips, nor can they touch the data generated during user use.
This means that car companies cannot personalize customization according to the needs of consumers, and it is difficult to differentiate; technical iterations and related problems can only rely solely on Mobileye.
In contrast, NVIDIA's solution provides a self-developed platform for car companies, and algorithms and data are all controlled by car companies according to their needs. In addition, Nvidia is the only chip company in the market that can provide a unified architecture and a unified software development environment for desktop, cloud and car.
As a result, mobileye's past customer barriers are also rapidly disappearing.
(Table 2: Chip selection of major car companies)
Tesla was the first to abandon Mobileye. In 2014, Tesla joined hands with Mobileye to adopt a solution with EyeQ3 as a driver assistance chip. However, the partnership quickly failed after only two years, and Tesla turned around and threw itself into NVIDIA's arms. In 2019, Tesla announced that it would start using its self-developed FSD chip.
As a long-time customer of Mobileye, BMW has also recently introduced a new partner. In November last year, Qualcomm announced a partnership with BMW, and BMW's next-generation ADAS and autonomous driving systems will use Qualcomm's Snapdragon Ride platform.
Although some of BMW's models will still be equipped with EyeQ5 chips in the future, the previous intimate relationship between the two sides has also ended with the cooperation between BMW and Qualcomm.
The eyes turn to domestic car companies. Wei Xiaoli's original models all used Mobileye's EyeQ4 chip, but Xiaopeng Automobile switched to Nvidia in 2020, and Ideal Automobile also adopted the Horizon Journey 3 chip in 2021. In 2022, the new models of Wei Xiaoli's three companies will use NVIDIA's Orin chip, which can be described as the same destination.
In addition to NVIDIA and Qualcomm, manufacturers such as Huawei, Black Sesame Intelligence and Horizon are also eating whales to swallow the market share that belongs to Mobileye. According to relevant data, Mobileye's market share in the global ADAS field has slipped from 90% in 2017 to the second echelon today.
According to the product planning, Mobileye's future market share is still likely to shrink further. Instead of testing the level of interest in the capital market at that time, it is better to take advantage of the time when it is still "OK".
In fact, as early as last December, Intel announced that it wanted to promote the independent listing of Mobileye. At that time, Kissinger also said that Intel would receive most of the proceeds of the stock offering, and would use part of the proceeds to build more chip factories. But what Kissinger may not have expected was that what had happened in just a few months had overwhelmed him.
Mobileye has to save itself first to fight a fire.