With the acquisition of Arriver, Qualcomm and Nvidia "car" on the decider

Image source @ Visual China

Text | Value Institute

In 2016, Qualcomm offered a sky-high $44 billion price to try to acquire NXP, the world's largest automotive chip manufacturer at the time.

At that time, the intelligent driving outlet had just emerged, and major car companies rushed to the front line of competition. The semiconductor companies in the upstream of the industrial chain are far less competitive than they are now.

However, this marriage of the century, which is expected to become the largest merger and acquisition in the history of the semiconductor industry, has failed after two years of tug-of-war. After the deal was officially stranded in July 2018, Qualcomm CEO Steve Mollenkov had to issue an urgent statement and make a massive buyback of up to $30 billion to stabilize shareholder confidence.

Nowadays, chip semiconductors have become the most competitive and important part of the intelligent driving industry chain. Qualcomm, who has been coveting this piece of fat meat, has finally struck again and taken down an important prey.

In early April, Qualcomm partnered with private equity SSW Partners to complete the acquisition of Swedish automotive technology company Veoneer. On April 11, according to multiple media reports, SSW Partners is about to complete the business split of Veoneer, the latter's autonomous driving software business unit Arriver will be merged into Qualcomm, and other businesses will be sold to other buyers.

After reinforcing the shortcomings of software technology, Qualcomm is bound to set off more waves in the track of intelligent driving.

Win-win trades? Qualcomm finally reinforced the biggest shortcoming

The first thing to be clear is that although Veoneer's market value and stock price have fallen sharply after being spun off from safety equipment supplier Autoliv, its research and development capabilities are still well known in the industry. In particular, the technology related to intelligent assisted driving, whether on the software side or the hardware side, is the top level in Europe and even in the world.

Headquartered in Stockholm, Sweden, Veoneer has branches in 11 countries and territories around the world and employs more than 7,000 people, according to official figures. In fiscal 2020, its full-year sales reached $1.37 billion, and its performance has continued to grow in recent years. A large and fairly high-quality R&D team, as well as a large number of patent rights, are Veoneer's proud wealth and the most coveted resources for buyers.

The joint consortium of Qualcomm and SSW Partners also spent a lot of thought and money to acquire Veoneer.

On the one hand, there is a premium phenomenon in this acquisition. According to information revealed in October last year, Qualcomm and SSW Partners offered a purchase price of $37 per share, with a total consideration of about $4.6 billion, while Veoneer's market capitalization at the time was less than $4 billion.

On the other hand, Qualcomm and SSW Partners also face a lot of competition from their opponents.

On July 22 last year, Canadian auto supplier Magna announced that it was acquiring Veoneer for about $3.8 billion, a move that was filed earlier than Qualcomm.

According to a research report by the Royal Bank of Canada, Magna, like South Korea's LG Group, is the most likely to become the first supplier of Apple cars and the closest to signing a contract with Apple. In an exclusive report in April last year, the Korea Times also mentioned the cooperation agreement between the three companies.

Strategy Analytics analyst Mark Fitzgerald is even more blunt, Apple is magna's bid for Veoneer behind the driving force.

However, the final result is clear to everyone, qualcomm and SSW Partners with higher bids and more professional traders successfully defeated various opponents and successfully won Veoneer.

When the boots landed and looked back, the Value Institute believes that this transaction can be said to be a win-win for both Qualcomm and Veoneer.

From Veoneer's point of view, due to its limited capital strength, it is not easy to kill a bloody road in the fiercely competitive and expensive intelligent driving market.

Backed by Google's towering tree, Waymo's loss for the full year of 2021 has magnified to $5.3 billion, and its revenue in the fourth quarter has even declined slightly. In this piece of business, Google itself does not dare to set profit expectations, and can only continue to burn money under pressure.

For more independent companies like Veoneer, financing and being acquired is a better way out. In recent years, intelligent driving has become one of the hottest financing tracks, and various automotive, technology, and semiconductor giants have thrown olive branches to the head start-ups and quickly raced.

Cruise, a U.S. self-driving technology company, received an additional $1.35 billion investment from the SoftBank Vision Fund in February to advance its commercial ride-hailing trials in San Francisco. Previously, SoftBank had injected $900 million into Cruise, with General Motors, Honda and Microsoft also being its shareholders.

At the end of last year, Qualcomm Ventures participated in the A round of financing of Zhixing, and Nvidia has also made frequent moves in recent years, successively investing in star unicorns in the autonomous driving industry such as Almotive, Optimus Ride and Tucson Future. Obviously, an arms race around autonomous driving has long been quietly started among semiconductor giants.

With the acquisition of Qualcomm and SSW Partners, Veoneer naturally has a stronger capital strength and can provide support for future research and development and testing.

For Qualcomm, getting Veoneer's Arriver business can perfectly complement its shortcomings in self-driving software technology.

Qualcomm CFO Akash Palkhiwala said early in the early days before the acquisition was fully closed that Arriver's ADAS software technology would be an important part of Qualcomm's automotive business.

"The company's future potential market size will increase from 100 billion to 700 billion, and the automobile business will become an important part of Qualcomm's business model diversification reform."

Looking at Arriver's specific business, the Value Institute believes that its strength should not be underestimated - with more than 800 R&D technical employees around the world, the complete construction of the perception software Arriver Vision Perception, as well as the driving strategy software NCAP Drive Policy and the next generation of perception and driving strategy software stack that inherits the essence of Veoneer technology.

It can be said that in the field of intelligent driving software technology, Arriver's family foundation makes Qualcomm ashamed. Winning The Arriver undoubtedly makes Qualcomm more confident and more capable of competing with other semiconductor giants.

Competing for the deer: Nvidia and Intel are not afraid of Qualcomm challenges

After Qualcomm successfully took over Arrivor, Veoneer's self-driving software business unit, Nvidia and Intel will certainly not be too happy.

In the past few years, the competition between the three major semiconductor giants in the intelligent driving industry has gradually become white-hot, extending to all aspects of the industrial chain, and the confrontation is quite fierce. Chip companies such as Suri an NXP, Infineon and STMicroelectronics still firmly hold the majority of the automotive semiconductor market. But with the rise of smart driving technology, a technological revolution in software-defined cars and software subverting hardware has quietly begun.

For highly sophisticated, highly secure smart cars, technological breakthroughs in software and hardware are indispensable. In terms of software, from the underlying system software to the automatic driving software, to the intelligent cockpit, the Internet of Vehicles and high-precision mapping software, are constantly breaking through various bottlenecks.

Taking the autonomous driving technology software we are most familiar with as an example, the three core links of perception, decision-making and control require the most compact algorithm model and data processing system to support its operation, and software technology plays an irreplaceable and important role in this.

In the field of perception technology, Tesla has been studying multi-sensor fusion solutions with cameras as the core, and Google's Waymo and Baidu Apollo are mainly focused on lidar-based solutions supplemented by other sensors. But no matter which scenario, in order to achieve full autonomous driving and high-level intelligent assisted driving, software technology is a difficult problem that they must overcome.

Arriver also has a deep accumulation in vision systems, lidar and radar, which Qualcomm lacked before. More importantly, software technology is absolutely upstream of the industrial chain, and it is one of the most profitable links in the "smile curve" of the automotive industry, with strong commercial prospects.

The "2021 Software-Defined Vehicle Research Report" jointly released by a number of research institutions pointed out that the software IP licensing fee of smart cars alone is about 30% higher than that of traditional cars, and Tesla's FSD Beta software technology has completed several rounds of price increases since its launch.

Qualcomm spending so much manpower and material resources to win the Arriver is a signal that the semiconductor giant can drive the software technology war upgrade. Comparing the reserves and investment layout of Nvidia and Intel in intelligent driving software technology, in fact, the strength gap is not large, and no one dares to guarantee the final victory.

Overall, Nvidia is the temporary lead.

At last year's GPU Technology Conference, Huang Personally introduced NVIDIA DRIVE Atlan, NVIDIA's new-generation self-driving car AI computing platform, which is one of the most comprehensive systems in software and hardware.

It is worth mentioning that in terms of computing platforms, NVIDIA can unilaterally hang Qualcomm and NVIDIA. Including Audi, Weilai, Xiaopeng, Ideal, Weima, Great Wall and other car companies, are using NVIDIA's system.

In addition, its autonomous driving modular software stack Drive SDK and development platform Drive Hyperion are also at the leading level in the industry. At present, NVIDIA can provide customers with a set of more complete integrated software services for perception, planning and control, and achieve full coverage of basic software and application software, and the comprehensive strength is second to none.

In contrast, Intel is not as comprehensive as Nvidia, but it also has its own strengths - the underlying system.

Some time ago, when the news came out that Intel split Mobileye and went independently in the United States, the Value Institute had analyzed its technical advantages.

Since being acquired by Intel, Mobileye has begun to transform into a system supplier and integrated Intel's lidar and 4D radar research and development teams. In the process of upgrading from assisted driving to automatic driving, Mobileye has also gradually improved its computing power base and formed a complete set of underlying software systems.

As mentioned earlier, software technology is becoming increasingly important for the intelligent driving industry. After consolidating the technical strength, the competition of the three major semiconductor giants for middle and downstream customers may enter a new stage.

The competition for customer sources is becoming more and more intense, and open systems will become the trend of the future?

The mutual customer snatching between Nvidia, Intel and Qualcomm has long been not news. Only now, the object of competition between the three giants has spread from traditional consumer electronics companies to automobile companies. It should be noted that although software technology is becoming more and more important and the competition is very fierce, it does not represent the full strength of the semiconductor three in the field of intelligent driving. Their old bank, chips, is also very critical for downstream car companies.

Especially driven by Intel, packaging hardware such as chips and software technologies such as intelligent driving solutions has become the mainstream of the industry.

Long before being acquired by Intel, Mobileye adopted the so-called "black box solution" to integrate software and hardware technologies such as chips, operating systems, and intelligent driver assistance systems into packages and sell them to customers. Since 2007, almost all models of BMW, GM and Volvo have been inseparable from Mobileye's services.

Inspired by Intel, even Qualcomm, which started relatively late, is also taking the route of "packaging software and hardware services".

At the 2022 CES International Consumer Electronics Expo, Qualcomm announced the launch of the latest combination of the Snapdragon Ride platform, the Snapdragon Ride Vision System, which will integrate Qualcomm Snapdragon's Ride SoC autonomous driving technology and Arriver's next-generation visual perception software stack.

However, the practice of packaging and selling hardware and software cannot satisfy all customers: the biggest problem with this practice is that the software and hardware ecology is too closed to leave more development space for car companies.

The most obvious negative impact of this closed ecology is also Intel.

In 2021, Weilai and Volvo have abandoned Intel and switched to Nvidia's Orin chip; even BMW, which has worked with Intel for many years, has joined hands with Qualcomm shortly after. In addition, the cooperation between Tesla and Intel has long been broken, the former because of the earlier feud with Mobileye and switched to its own FSD chip, ideally preferring to change to Horizon Journey 3 rather than Use Mobileye's EyeQ5.

A large number of customers abandon Intel, in addition to looking at the chips with higher computing power of semiconductor giants such as Nvidia, a more open ecology also has a fatal attraction for them.

We can look at two of the most stark examples: the ideal move to horizon, and BMW holding hands with Qualcomm.

As a start-up, although Horizon has reached a high standard in the industry in terms of chip computing power and underlying software system technology, it does not have many advantages over Intel. What really attracts customers is its highly open cooperation model.

According to the service rules currently provided, Horizon can open the underlying software OS to cooperative car companies, car companies can develop more software services according to their own needs on the basis of the system, and Horizon will also provide other services such as money collection chip IP.

Ideal Car founder Li Xiang said bluntly that Intel and Mobileye cannot meet their own needs for self-research in the field of intelligent driving, and Horizon's services can meet this requirement.

As for bmw that has cooperated for more than a decade from Intel to Qualcomm, it is also indispensable to the latter's compromise.

Although Qualcomm has not given up selling its own software and hardware integration services to customers, they still open the door of convenience for BMW, a big customer.

Last November, Qualcomm's open automatic software stack to BMW was already the first to be used in the innovative BMW iX model. In the future, Qualcomm, Arriver and BMW will also jointly develop next-generation autonomous driving technology solutions based on BMW's autonomous driving software stack.

In the final analysis, from closed ecology to joint development, the semiconductor giants and car companies are competing for dominance over autonomous driving technology. We must understand that the relationship between the two is both interdependent and mutually constraining. Car companies compete for more autonomy, emphasizing the self-development of the full boardwalk, but also to fight for their own future technology insurance, to avoid being stuck by upstream semiconductor manufacturers.

How to find a satisfactory balance in this game - that is, to determine the degree of technological openness - is a difficult problem that the semiconductor big three need to solve.

Write at the end

As we all know, semiconductors are a highly cyclical sector, and the performance in the capital market has been fluctuating considerably. However, due to the outbreak of the epidemic, many links in the production chain have been seriously affected, and the stock prices and market value of semiconductor giants seem to have shown a counter-cyclical trend in the past two years.

However, historical experience tells us that after any super cycle, those sectors and leading stocks that lead the market will usher in adjustments, and the semiconductor sector may not be an exception. A series of recent indicators can also see signs of decline in the return rate of the semiconductor industry, such as the VanEck Vectors Semiconductor ETF market, which tracks semiconductor production and equipment companies, according to the Wall Street Journal.

Intel missed the outlet of pc transformation to smartphones and AI, and watched Qualcomm and Nvidia catch up later, which is a painful lesson. The semiconductor giants that have exhausted the dividends of the times naturally need to take precautions and grab a ticket to the future for themselves.

As mentioned earlier, the future of the intelligent driving industry is not short of imagination space, whether it is Nvidia, Qualcomm or Intel, there are still opportunities to break through. For the peripheral melon eaters, we certainly hope that the technology roll of these semiconductor giants can lead the entire industry forward.