Is a 28nm chip more expensive than a 3nm chip? This is no joke.

A friend from a chip manufacturer revealed a shocking fact after a few drinks. The flagship chip in your phone, which is 3nm, costs only a few dozen dollars. But that “outdated” 28nm chip in your car? It can easily sell for over a hundred dollars.

This is not an IQ tax. It’s an industry secret you don’t know.

Today, we will break this layer of glass and see why automotive grade chips are so “arrogant” and why they dare to sell for more than mobile chips. No exaggeration, just the facts today.

1. The “Extreme” Requirements of Automotive Chips: What Does 15 Years of Zero Failure Mean?

What happens if a mobile chip fails? Just get a new phone. But if an automotive chip fails? That could be life-threatening.

Let’s start with the most basic difference. Your mobile chip lasts at most three to five years. By then, it’s either lagging or broken, and you just get a new phone. But automotive chips? They are required to last 15 years with zero failures.

What does 15 years mean?

The chip in your current car must work reliably from 2025 to 2040. It must endure extreme cold and heat, countless bumps and vibrations, and electromagnetic interference without any issues.

The temperature requirements alone can drive ordinary chips crazy.

What is the working temperature for mobile chips? 0-70°C at most. Exceeding this range will cause them to crash. But automotive chips must withstand extreme temperature ranges from -40°C to 150°C.

Lei Jun once mentioned that during tests in Xinjiang, even tissue boxes melted. Can you imagine what would happen to ordinary chips in such an environment? They would fail in an instant.

In summer, the dashboard can get hot enough to fry an egg. In winter, in Northeast China, it can freeze solid at minus dozens of degrees, yet the car must still start normally. In such environments, automotive chips must be as stable as a rock.

Vibration and electromagnetic interference are even more critical.

The roar of the engine and the bumps on the road are torture for chips. A mobile phone can break if dropped, but automotive chips must work for over a decade in such conditions.

Just designing redundant circuits and protective modules adds 20% to the cost. And that’s just the beginning.

2. The Bloody Cost of Production: What Does a Yield of 70% vs. 90% Mean?

This is the harshest reality. It’s also a cost black hole that consumers will never see.

The defect rate requirement for automotive chips is one in a million. It’s stricter than the sterile standards in an operating room.

For the same 28nm process, mobile chips can achieve over 90% yield. Automotive chips are forced down to 70%. Do you know what that means?

The cost of wafer production explodes.

Originally, a wafer could produce 100 qualified mobile chips. Under automotive standards, it can only produce 70. The remaining 30 are all waste.

The wafer cost skyrockets to $15,000 each. Spread across each chip, the cost becomes outrageous. This doesn’t even account for subsequent testing costs.

The testing phase is an even bigger money pit.

Thermal cycling tests, vibration table shaking, electromagnetic compatibility tests—each of these costs money. The testing cost for a single chip is five times that of consumer-grade chips.

YouTube blogger @TechBusters conducted a real test. The failure cost of a chip from a certain car manufacturer is enough to buy 20,000 mobile phones. This is not an exaggeration; it’s a bloody reality.

The certification threshold is frighteningly high.

Car manufacturers often require signing a ten-year supply agreement. Chip manufacturers must reserve production lines specifically. The depreciation of equipment for TSMC’s 28nm automotive capacity is shared among several car manufacturers.

All these costs are ultimately passed on to the price of the chips. Each Renesas MCU selling for $15? That’s already a reasonable price.

3. Hidden Costs Are the Real “Killer”

The surface production costs are not the scariest part. What truly makes automotive chips sky-high are the hidden costs you can’t see.

The R&D cycle starts at three years. Certification fees can start at several million dollars. All this money must be spread across each chip.

The risk of recalls is a Damocles sword hanging over your head.

If a mobile chip has issues, users might complain a bit and get a new phone. If an automotive chip has issues? That could mean recalling the entire vehicle, costing millions of dollars in compensation.

This risk premium directly raises the price of chips.

The monopoly structure keeps prices “stiff”.

The automotive chip market is tightly controlled by oligopolies like Infineon, Renesas, and NXP. Want to enter this market? First, pay a “protection fee” of several million dollars.

In contrast, the mobile chip market. Apple’s A17 sells for $150, which seems expensive, but with sales in the hundreds of millions, the manufacturing cost per chip is only $0.3.

The huge gap in scale effects.

A mobile chip can have annual production in the hundreds of millions. Automotive chips? A single model selling a few hundred thousand units a year is already good. Each car uses different types of chips.

This scale difference directly determines the vast disparity in cost structures.

4. Will Prices Drop in the Future? Autonomous Driving Chips Offer New Hope

Everyone is now focusing on autonomous driving chips. The introduction of 7nm technology may change the game.

Tesla’s FSD chip, Horizon’s Journey series, NVIDIA’s Drive series—all are using more advanced processes. As production scales up, costs will naturally decrease.

But for ordinary automotive MCUs? Those chips controlling wipers, windows, and seats? For the next five years, we will likely continue to pay for this “sense of security”.

Can domestic chips break the monopoly?

BYD is developing its own automotive chips. Horizon is also making strides. But to truly shake the position of foreign oligopolies will take time.

Chips are not just about being made. They need to be validated over time and gain the trust of car manufacturers; this process cannot be completed in less than ten years.

Where is the hope for cost reduction?

First, scale effects. With the popularization of electric vehicles, the demand for automotive chips will increase significantly. Second, technological advancements. The maturity and widespread adoption of more advanced processes. Third, intensified competition, breaking the existing monopoly structure.

But to be honest, automotive chips will never be as cheap as mobile chips. The safety requirements are there, and this premium is necessary.

To summarize: There’s a reason automotive chips are expensive.

Is 28nm more expensive than 3nm? On the surface, it seems unreasonable, but it actually makes complete sense.

The requirements for 15 years of zero failures, 70% yield, five times the testing costs, millions of dollars in certification fees, and the huge compensation for recall risks—all these add up to make it not surprising that they are expensive.

However, this “expensiveness” is also a guarantee of our safety. Would you risk brake failure just to save a few bucks?

So here’s the question: Do you think the price of automotive chips is reasonable? Or are car manufacturers collecting an IQ tax? Feel free to leave your comments below!

We look forward to domestic chips providing us with more choices while ensuring safety, and we hope the entire industry becomes more transparent, allowing consumers to make informed purchases.

What do you think? Is this money well spent?

Give it a thumbs up

Let’s move on