Automotive Chip Industry: Making Brains for Electric and Self-driving Vehicles
An overview of the technological trends and competitive landscape of the rapidly emerging auto microelectronics market.
As technology advances, the nature of mobility is undergoing considerable changes. While innovative ideas like the Internet of Things (IoT) – and, consequently, business models that are based on them – naturally disrupt the core methods of ‘data transportation,’ the physical ways to move things (and people) from one place to another are gradually being improved.
Major transportation methods have a long history of continuous evolution, following stages of major technological progress in the fields of mechanics, chemistry and other areas of fundamental science. Nevertheless, the mobility industry can no longer rely solely on the classical drivers of advancement. Stirred by international rivalry, car makers around the globe have no choice but to adopt the best new technologies from across the value chain in order to differentiate their final products.
At the moment, safety and comfort are the two significant directions of product differentiation in automotive. To increase these two, the world’s leading corporations are spending a lot of money on micro-innovation in four realms: electrification (as a result of the recent energy revolution), absolute automation (a long-term process widely considered the final step in the sector’s development), connectivity (a ubiquitous force that is enabled by the combination of the physical and digital worlds, which continue to enmesh, making digitalization more pervasive on a large scale) and security (including both digital and physical aspects).
As the market steps into the phase of oversaturation, the number of cars sold planetwide isn’t snowballing anymore. Moreover, the industry is projected to dwindle over the next two years, indicating deteriorating economic conditions across several key areas. Still, the struggle for innovation is ubiquitous among the most prominent car manufacturers, and all the four abovementioned emerging zones rely on micro-components.
According to Deloitte, semiconductor content per car is projected to hit USD 600 as soon as 2025. One of the auto chip market leaders, German company Infineon, predicts that driving automation level 4-5 models will, on average, contain up to USD 970 of semiconductor content per car by 2030. Another behemoth, Dutch multinational NXP Semiconductors, projects that electronic systems will account for over 50% of total car cost by 2030. Safety, fuel efficiency, comfort and infotainment are frequently considered the key drivers enabling this shift.
For semiconductor companies, Advanced Driver-Assistance Systems (ADAS) are arguably the most attractive new layers in the automotive industry: besides the powerful processors that are needed to run complex and heavy Machine Learning (ML) algorithms, the systems also require an immense number of sensors and memory chips.
In addition, there are countless opportunities in other segments of the automotive industry. Electrification and digitization – as well as other hot trends and concepts – all create more room outside of the proverbial ‘safety plus comfort’ paradigm. The scope of possible applications is expanding, and Application-Specific Integrated Circuits (ASIC) are taking over realms where Central Processing Units (CPU) were previously dominant, but already look obsolete due to their inflexibility and relatively low capacity. This trend, in many respects, resembles the story of Artificial Intelligence (AI) accelerators.
Importantly, when talking about the looming digital revolution in mobility, many think tanks and research organizations take into consideration a vehicle as a distinct subsystem or, in some cases, the nature of the interaction between different machines. The fact of the matter is that ‘connectivity’ means more than just car-to-car solutions – literally, in the multidimensional networks of the future, almost everything is projected as being linked to everything else.
Hence, in many depictions of this scenario, all the static (for example, urban infrastructure) and dynamic (all the moving objects such as cars, pedestrians, or animals) objects must be involved in the planetwide smart circuits. This, obviously, can’t be done without the use of microelectronic devices of all kinds – and this is the honey pot plenty of companies are vying for.
Clash of the titans?
As things stand, electronic systems are becoming the most intense battlefield in the technology markets. Meanwhile, there is a fine line between ‘innovative solutions’ and ‘solutions that can be easily applied to mass production.’ Sunk costs are thereby normal for projects run by both car makers and automotive chips producers. This, along with and other factors, is leading to the overconcentration of the international auto semiconductor scene.
In 2018, the five largest chipmakers held 47.9% of the global market (which had more than doubled in size since 2010). Among them, Dutch giant NXP Semiconductors (NXPI:Nasdaq), German corporation Infineon (IFX:DE) and the world’s largest maker of microcontrollers – Japanese firm Renesas Electronics (6723:JP) – crowded the top of the list, with 12%, 11.2% and 8.9% respectively.
It is important to note that this short list of the sector leaders is not just striking – it is staggering. Here’s the reason: Texas Instruments (TXN:Nasdaq), the third-biggest chipmaker listed in the United States, can boast of a market cap of USD 110 billion. NXP and STMicroelectronics (STM:NYSE) – another European trailblazer – are the fifth and seventh semiconductor firms by this indicator. The two were worth USD 32.54 billion and USD 21.84 billion as of November 19.
A significant scale and ample resources allow these and other well-established conglomerates to have more space for development, which comes from their ability to make mistakes, even big ones. But how about companies that have only one shot? The picture is apparently different for private startup firms.
The strict economics of car making, conditioned by robust market demand, place limitations on the research and development, generating limitless hurdles for private companies. It is obvious that the automotive industry is far from being a smooth journey for startups, especially in advanced economies with well-established firms that have access to state-of-the-art technology and even run in-house projects in pursuit of self-sufficiency.
However, some young players are striving to make a difference within narrow segments of their respective supply chains. Traditionally, one of the best ways to discover the most prominent of them is through industry investment footprint analysis. So, who is raising Venture Capital (VC) these days?
Sprouts of the next generation
According to online investment data aggregator Crunchbase, only 15 Private Equity (PE) financing events in the automotive chips sector have happened over the past three years. The most significant of them – a USD 50 million Series B carried out by Israeli computer vision solutions provider Oryx Vision – took place in July 2017. This round of financing was the second for the firm after a USD 17 million Series A one year earlier.
Hangzhou-based FABU.ai, which makes ASICs for ADAS, completed another significant transaction last year: global VC firm MSA Capital injected USD 25 million in August 2018. Previously, it obtained an eye-popping USD 15 million in seed investment from Sinovation Ventures, which is widely known because of its rockstar CEO – computer scientist, investor and entrepreneur Kai-Fu Lee.
Intel-backed TriEye, which raised USD 17 million in a Series A led by the American corporation, is into ADAS as well. It develops image solutions for autonomous driving systems. The company’s core product, the Raven SWIR Camera, is used to combat low visibility. Besides applications in automotive, the firm claims that the optical instrument can be employed in smart cities, industrial settings and agriculture.
Black Sesame Technologies – which is, like many other recent autonomous driving startups, simultaneously based in Silicon Valley and China’s Mainland (Shanghai, to be exact) – plays on an adjacent field. Besides image processing and light control, it also markets perception systems and facial recognition tools. SK Hynix’s Chinese branch pumped venture capital into the firm in April 2019, joining NIO’s investment arm and some other locally known names.
While there are some other names and exciting stories behind them, we should once again admit that the automotive semiconductors vertical is probably the hardest to crack among both quasi-traditional subsectors like memory and the new often-buzzword-shaped concepts such as IoT. This explains the current deficit of startups in the area and sheds much light on PE/VC investment patterns.