Car chips: new balance between OEMs and manufacturers

The rise of self-driving vehicles is changing the demand for automotive chips and prompting OEMs to design their own solutions in-house. The value chains and strategies of all players in the supply chain are thus set to change.

125
Oem automotive chips

Edited by Giorgia Andrei| Based on an article by Ondrej Burkacky, Johannes Deichmann, Larissa Rott and Alexander von Falkenhausen. McKinsey & Company

It is important for companies in the automotive sector and those that manufacture or distribute semiconductors to understand the impact of new technologies for self-driving vehicles on demand for chips. Since 2010, OEMs and start-ups have invested $106 billion in this sector, with much of the investment going into improving advanced driver assistance systems (ADAS), which manage braking, object detection and other critical vehicle functions. The focus on autonomous vehicles has already altered demand patterns for automotive semiconductors, with strong growth in sales of special silicon chips, tailored for specific applications. These are chips that are only available from a handful of semiconductor companies, and some OEMs are designing them in-house to reduce development time and gain more control, launching a trend that may be followed by others.


The full article was published

on number 11 of Electronics AV


An evolving market for chips

The economic rebound in this post-pandemic period has led to a surge in orders for semiconductors, with factories busy increasing their output to meet rising demand. Most of the growth in this area has come from the shift to autonomous driving systems: Adas have the unique ability to process data instantaneously, a capability that requires more interconnections within the vehicle and high-performance chips. Their electrical and electronic (E/E) architecture is more centralised than in traditional vehicles, including more sensors and computing solutions. However, autonomous vehicles have different levels of automation: a common classification system, defined by the Society of Automotive Engineers (Sae), divides them into six categories, ranging from 0 (no automation) to 5 (full automation with autonomous driving capability). Level 2, in turn, is divided into two parts: entry level (vehicles with some autonomous features, such as braking, but which still require drivers to keep their hands on the wheel at all times) and advanced (vehicles with more extensive autonomous features, which in some cases allow drivers to take their hands off the wheel).

Level 0 to 2 (entry) vehicles achieve adequate performance with standard chips, but level 2 (advanced) to 5 vehicles are expected to require an increasing share of specialised silicon, i.e., more efficient chips that enable rapid performance increases within vehicle systems and the execution of complex software and analysis capabilities, such as those that make it possible to fuse sensors between cameras, lasers, lidar solutions and other devices.

However, many OEMs find it difficult to obtain silicon that corresponds exactly to their requirements and thus find themselves revising their development programmes. In addition, there are few solutions that fit their software base. Faced with these constraints, some OEMs have already started to design chips in-house, a route that offers numerous advantages: optimised performance for specific algorithms, shorter development time for feature improvements, software definition, and the possibility, in essence, to create customised solutions that could differentiate their autonomous vehicles (e.g. by allowing earlier time-to-market or providing greater feature availability). Some Tier One suppliers also claim ownership of certain software systems. The value chain is thus becoming increasingly disaggregated, with a division of competences that leads companies to seek more and more partners.

Opportunities for chipmakers

The increased involvement of OEMs in chip design could reduce the profits of semiconductor companies, but the spread of self-driving vehicles could grow the market for automotive chips and offset much of the business.

Revenues for 'autonomous chips', a major sub-category, are expected to rise to around $29 billion a year by 2030, or about $350 per vehicle, a sharp increase from $11 billion in 2019.

High-performance central processing chips, such as control units (CUs) and sensors, are expected to grow the fastest: revenues are expected to increase by around 12% per year from 2025 to 2030, while chips for electronic control units (ECUs) and decentralised sensors will grow by around 6% per year from 2019 to 2030.

Control units are expected to account for a higher percentage of revenues than sensors, due to the growing trend towards more central computing in autonomous driving operations, bypassing processing in sensors.

It is important for semiconductor companies to understand how the possible changes might affect their market share, but it is difficult to quantify the exact impact of OEMs choosing to design their chips in-house. However, three scenarios can be considered: of the approximately $6 billion market for central DCs in 2030, OEMs would account for about a third of the value if there was a small shift in this direction. The figure would rise to around 50% in the medium scenario and 80% in the highest scenario. Chipmakers therefore need new strategies, starting with hiring the best talent for chip design, E/E architecture and software, and working with OEMs to better understand the particular silicon requirements, bearing in mind that needs vary from region to region. It will also be crucial for semiconductor companies to build strong relationships with other players along the value chain, including specialty software players, integrated device manufacturers and Tier One suppliers. Tier One suppliers, for their part, could develop flexible offerings to avoid being excluded from the market: for example, Adas applications can be configured to run on any chip. Tier One vendors could also consider investing in semiconductor capabilities to complement their electronics and systems integration expertise and actively pursue partnerships with OEMs, semiconductor companies and technology players within their ecosystem, again to avoid being left on the sidelines.


The full article was published

on number 11 of Electronics AV


 

Previous articleAnalog Devices attendance algorithm
Next articleCybersecurity: strategic cooperation agreement between congatec and Sysgo

LEAVE A COMMENT

Please enter your comment!
Please enter your name here