Semiwiki on LinkedIn Semiwiki on Facebook Semiwiki on Twitter Register / Log in 
Array
(
    [content] => 
    [params] => Array
        (
            [0] => /forum/threads/samsung-to-produce-tesla-chips-in-16-5-billion-multiyear-deal.23248/
        )

    [addOns] => Array
        (
            [DL6/MLTP] => 13
            [Hampel/TimeZoneDebug] => 1000070
            [SV/ChangePostDate] => 2010200
            [SemiWiki/Newsletter] => 1000010
            [SemiWiki/WPMenu] => 1000010
            [SemiWiki/XPressExtend] => 1000010
            [ThemeHouse/XLink] => 1000970
            [ThemeHouse/XPress] => 1010570
            [XF] => 2021770
            [XFI] => 1050270
        )

    [wordpress] => /var/www/html
)

Samsung to Produce Tesla Chips in $16.5 Billion Multiyear Deal

K

kevin01

Well-known member
TL;DR: Samsung Electronics secured a $16.5 billion semiconductor supply contract starting July 2025 through 2033, likely with Tesla for Full Self-Driving (FSD) chips. This deal strengthens Samsung Foundry's position against TSMC amid growing demand for advanced automotive chips in the evolving electric vehicle market.

Read more: https://www.tweaktown.com/news/1066...033-rumored-be-for-tesla-fsd-chips/index.html

 
I have to wonder, where is IFS?

This kind of chip can probably be manufactured on Intel 3 and 18A. Why did Intel not pursue it aggressively? Is it because: 1) Intel 3 and 18A PDKs are not suitable for this kind of chip, 2) IFS is not aggressive in terms of pricing, or 3) something else?

If I were IFS, I would offer a price that would allow me to barely make money to secure the first big customer. Foundry is a fixed-cost business; even if a contract doesn't make money, it at least pays the employees and keeps the factory running.
 
kevin01 said:
I have to wonder, where is IFS?

This kind of chip can probably be manufactured on Intel 3 and 18A. Why did Intel not pursue it aggressively? Is it because: 1) Intel 3 and 18A PDKs are not suitable for this kind of chip, 2) IFS is not aggressive in terms of pricing, or 3) something else?

If I were IFS, I would offer a price that would allow me to barely make money to secure the first big customer. Foundry is a fixed-cost business; even if a contract doesn't make money, it at least pays the employees and keeps the factory running.
Click to expand...

I am trying to remember my A-level economics , covering your fixed costs was the basics for a company to build from?

So as long as fixed costs were covered Intel , would Intel have "been laughing"
 
kevin01 said:
I have to wonder, where is IFS?

This kind of chip can probably be manufactured on Intel 3 and 18A. Why did Intel not pursue it aggressively? Is it because: 1) Intel 3 and 18A PDKs are not suitable for this kind of chip, 2) IFS is not aggressive in terms of pricing, or 3) something else?

If I were IFS, I would offer a price that would allow me to barely make money to secure the first big customer. Foundry is a fixed-cost business; even if a contract doesn't make money, it at least pays the employees and keeps the factory running.
Click to expand...
Maybe the profit margin isn't high enough, the CEO said IFS would not do anything if it didn't make them 50% profit margin.
 
Even if it is simply profitable or profitable, automotive semiconductors also need reliability that is different from ordinary semiconductors.
It's the same in the manufacturing process
Intel also needs to gain experience in manufacturing processes for high-reliability products

In the first place, it takes time for a vehicle with a Tesla chip made of that Samsung Fab to actually come out.
Really state-of-the-art new process nodes are not ready to be produced immediately

For example, let's talk about TSMC, do you know the process called TSMC N3A?
This is a variation of N3 nodes for automotive products and will be in production in 2026.
In 2026, nearly three years have passed since the first release of N3, and N2 has also appeared. and A16 is also in sight.
Anyway, in-vehicle semiconductors require reliability
 
1753690955672.png
 
tomatoma said:
Even if it is simply profitable or profitable, automotive semiconductors also need reliability that is different from ordinary semiconductors.
It's the same in the manufacturing process
Intel also needs to gain experience in manufacturing processes for high-reliability products

In the first place, it takes time for a vehicle with a Tesla chip made of that Samsung Fab to actually come out.
Really state-of-the-art new process nodes are not ready to be produced immediately

For example, let's talk about TSMC, do you know the process called TSMC N3A?
This is a variation of N3 nodes for automotive products and will be in production in 2026.
In 2026, nearly three years have passed since the first release of N3, and N2 has also appeared. and A16 is also in sight.
Anyway, in-vehicle semiconductors require reliability
Click to expand...
Well, even if a new advanced process emerges, it will take some time before it can be used for automotive products.
It's the same for any manufacturer
Moreover, long-term supply must also be provided.
 
Xebec said:
View attachment 3398
Click to expand...

Personally, I do not mind that Elon collaborates with Samsung in the future. Whatever comes out, Elon doesn't seem to mind burning cash,
He is the guy of overpromising and under delivering, unlike CC Wei. I do not think TSMC minds Elon leaving the TSMC orbit at some point.

Elon seems not an easy guy to collaborate with. Once you put some 10-12 kids in the world, because you think you are the greatest and best the male-world offers genetically, time to take a distance from the guy. The world is big enough to do interesting and valuable stuff and not getting entangled into Elon's orbit.

His X-CEO finally saw the light and left X, Elon burns through people too quickly and too fast:
https://www.forbes.com/sites/tylerroush/2025/07/09/musk-loses-x-ceo-linda-yaccarino-resigns/
 
user nl said:
Personally, I do not mind that Elon collaborates with Samsung in the future. Whatever comes out, Elon doesn't seem to mind burning cash,
He is the guy of overpromising and under delivering, unlike CC Wei. I do not think TSMC minds Elon leaving the TSMC orbit at some point.

Elon seems not an easy guy to collaborate with. Once you put some 10-12 kids in the world, because you think you are the greatest and best the male-world offers genetically, time to take a distance from the guy. The world is big enough to do interesting and valuable stuff and not getting entangled into Elon's orbit.

His X-CEO finally saw the light and left X, Elon burns through people too quickly and too fast:
https://www.forbes.com/sites/tylerroush/2025/07/09/musk-loses-x-ceo-linda-yaccarino-resigns/
Click to expand...

I think $ USD Billions of business probably matters more to TSMC, Samsung, and Intel than some of the public's perception of Tesla's CEO.
 
Xebec said:
I think $ USD Billions of business probably matters more to TSMC, Samsung, and Intel than some of the public's perception of Tesla's CEO.
Click to expand...
I knew that Samsung gave very very good terms for Tsla's Hw3 manufacturing, even though it was a fairly small business initially.

The HW3 chip was custom-designed by Tesla and built on Samsung's 14nm process around 2019. The HW4 was on Samsung 7nm. The Hw5 should be on TSMC N3(?) and is scheduled to release soon.

Interestingly, according to wikipedia, Musk stated HW5 will be ten times more powerful than HW4. Musk also stated that it will use up to 800 watts when processing complex environments, versus up to 100 watts for HW3 and 160 watts for HW4.

Tesla Autopilot hardware - Wikipedia

en.m.wikipedia.org en.m.wikipedia.org
 
kevin01 said:
I knew that Samsung gave very very good terms for Tsla's Hw3 manufacturing, even though it was a fairly small business initially.

The HW3 chip was custom-designed by Tesla and built on Samsung's 14nm process around 2019. The HW4 was on Samsung 7nm. The Hw5 should be on TSMC N3(?) and is scheduled to release soon.

Interestingly, according to wikipedia, Musk stated HW5 will be ten times more powerful than HW4. Musk also stated that it will use up to 800 watts when processing complex environments, versus up to 100 watts for HW3 and 160 watts for HW4.

Tesla Autopilot hardware - Wikipedia

en.m.wikipedia.org en.m.wikipedia.org
Click to expand...

Does Samsung has also advanced packaging developments to handle the 800 Watt? Do they also move to photonics-co-packaged at Samsung's leading edge foundry processes?
 
Xebec said:
I think $ USD Billions of business probably matters more to TSMC, Samsung, and Intel than some of the public's perception of Tesla's CEO.
Click to expand...

Perhaps a nice move by TSMC to let the world "know" they are not running a (quasi)-monopoly ;) Samsung is taking business from them!
 
Knowing how Elon/Tesla works, it was probably a combination of price and control. Tesla will strong arm it's suppliers and wants a lot of control over how they do things. Intel probably wasn't willing to bend enough for them.
 
count said:
Knowing how Elon/Tesla works, it was probably a combination of price and control. Tesla will strong arm it's suppliers and wants a lot of control over how they do things. Intel probably wasn't willing to bend enough for them.
Click to expand...

Unfortunately this is the story of how Intel lost it's Foundry business.

The same exact thing happened when Otellini wasn't willing to fab Apple's ARM chips right before iPhone took off..
 
A little more on perhaps why Samsung:

Curious how Musk will react if Samsung yields in TX follow their pattern of low yields in SK on advanced nodes.

1753712316416.png
 
Xebec said:
A little more on perhaps why Samsung:

Curious how Musk will react if Samsung yields in TX follow their pattern of low yields in SK on advanced nodes.

View attachment 3399
Click to expand...

Yes, I can imagine CC Wei (and all of TSMC customers) would not appreciate it when Elon is messing around TSMCs Fabs in Arizona and pissing off all his engineers, like also the new (star)-CEO of TSMC-Arizona, who's job it is to make Arizona a well-run and yielding TSMC-Gigafab, as if it was located in Taiwan :p
 
This situation marks four wins and one loss:
(1) Tesla successfully diversified its supply chain, cutting costs;
(2) Samsung landed a major deal for the Texas facility and is positioned to support GAA development;
(3) TSMC, though losing the contract, relieved itself of the strain of enlarging investment in U.S.
(4) The U.S. saw a boost in semiconductor production.
The outlier? Intel Foundry
 
You must log in or register to reply here.
Back
Top