laitimes

Will chips account for 20% of the cost of automotive BOMs in 2030? | Principal Zhu's column

Will chips account for 20% of the cost of automotive BOMs in 2030? | Principal Zhu's column

At last year's German International Automotive and Smart Mobility Expo, Intel CEO Pat Gelsinger delivered a keynote speech and put forward an interesting prediction: by 2030, chips will account for more than 20% of the cost of high-end automotive materials (BOM), an increase of 5 times over 4% in 2019; and about 12% in 2025, this process of increasing the proportion of value will continue to occur (the demand for chips in various industries around the world is also growing)... Today we will talk about this matter.

Will chips account for 20% of the cost of automotive BOMs in 2030? | Principal Zhu's column

Figure 1 Will chips directly occupy such a high vehicle manufacturing cost according to this trend?

Part 1: Can chips account for 12% of BOM costs by 2025?

Let's take the 1764kg Model 3 for reference (before there were also many car companies to take the price of a vehicle per kilogram and the price of crayfish). From the perspective of sub-systems and sub-parts, when we think that 60-80kWh batteries account for about 40% of the BOM cost of the Model 3 vehicle, if the semiconductor accounts for 20% of the BOM cost in the future (even if it is 12% in 2025), then other components account for less than half of the cost, and the weight of this part needs to account for about 70% of the vehicle.

Note: To be honest, at the moment, people do invest the incremental part of the value in semiconductors and software.

Will chips account for 20% of the cost of automotive BOMs in 2030? | Principal Zhu's column

Figure 2 Can we accept that the value of traditional components is gradually reduced to a small proportion?

In several previous UBS reports, there is a blueprint for the chip value of existing Model 3 electric vehicles - in the overall cost of Model 3, the estimate of all semiconductors in the vehicle is 1516 US dollars (about 9500RMB), if we take the average selling price of the Model 3 model of 42000 US dollars, the BOM cost is estimated at 25,000 US dollars, then the current chip accounted for about 6%. Of the $1516 chip cost, the most is the electric drive part, followed by the ADAS part, then the entertainment system and PD part, and the least is the body control part.

Will chips account for 20% of the cost of automotive BOMs in 2030? | Principal Zhu's column

Figure 3 Tesla's cost structure ratio

In other words, the chip cost around the entertainment system is $147, ADAS is $510, body control is $115, PDS is $107, and the electric drive part (including BMS, high-voltage control, OBC, DC-DC and power semiconductors) is a total of $636.

Note: Here the estimation of inverter SiC seems to be less.

Will chips account for 20% of the cost of automotive BOMs in 2030? | Principal Zhu's column

Table 1 Table of chip values decomposed at that time

In this subdivision of $636, some disassembly is also done, as shown in the following figure:

Will chips account for 20% of the cost of automotive BOMs in 2030? | Principal Zhu's column

Figure 4 Powertrain related chip 636 US dollars cost breakdown table

Will chips account for 20% of the cost of automotive BOMs in 2030? | Principal Zhu's column

Figure 5 Value decomposition of chips in the corresponding powertrain (supplier)

Before replacing the Nvidia chip, Tesla's allocated chip cost, follow-up bicycle 200 US dollars, Tesla also used its own team to innovate this part, that is to say, the procurement part of this part is decomposed into chip development and chip FOUNDry costs - team manpower and chip IP costs are included in the development cost, effectively reducing costs.

Will chips account for 20% of the cost of automotive BOMs in 2030? | Principal Zhu's column

Figure 6 The overall chip cost of a non-powertrain

Part 2: The value is not worth it

I think auto companies will eventually follow the path that mobile phone companies have taken. Chips do account for a large part of the cost, but it is also necessary to see that players like Tesla are also trying to reduce the cost of chips through the simplification of their own design and architecture.

Note: Tesla's biggest feature before was that it could replace a large number of low-level MCU chip designs.

And we can see Intel Mobileye's quarterly revenue - growth is slow, each quarter stuck at about 300 million US dollars, in other words, in the current L2 era, luxury car (non-electric) the most important chip investment is still in the cockpit and intelligent driving these two pieces, electrified powertrain may increase by 600-800 US dollars, but the overall proportion is still very limited.

Will chips account for 20% of the cost of automotive BOMs in 2030? | Principal Zhu's column

Figure 7 Quarterly revenue of Intel's Mobileye

That is to say, Intel's CEO estimates that in the next L4 era (2030) that Mobileye expects, high-end electric vehicles equipped with multi-lidar and high computing power may cost $3,500 under multi-perception and high computing power, and then continue to increase to the expected $5,000 (20% of the BOM cost). The largest incremental part is in the ADAS piece.

Will chips account for 20% of the cost of automotive BOMs in 2030? | Principal Zhu's column

Figure 8 UBS's growth projections for sub-regional semiconductors

brief summary:

Judging from the amount of value, which can be constantly updated and disassembled, I mainly think about how to build the overall investment framework next. This time, I revisited these reports, and I can draw very interesting points from them and keep trying to think about the value in them.

Figure | network and related screenshots

About author:Zhu Yulong, senior electric vehicle three-electric system and automotive electronics engineer, author of "Automotive Electronics Hardware Design".

Write a message

2030 Mobility Research Laboratory

The first in China to consist entirely of PhDs

In-depth research organization for new mobility in automobiles

·

Zhihu private message ID: Fish is not fish

Read on