TSMC Reportedly Rejects Samsung's Proposed Exynos Mass Production Request

[T0@st]

Samsung's native foundry operations have wrestled with the 3 nm Gate-All-Around (GAA) process—these problems have persisted since the first reports of "missed production targets" emerged late last year—online speculators floated a very disappointing yield figure: only 20%. Last December, industry moles proposed that the South Korean technology giant had devised plans to form an Exynos-centric "multi-channel partnership" with rival chipmakers. Speculation pointed to TSMC being the only valid ally. Semiconductor industry tipster—Jukanlosreve—believes that negotiations have taken place, and the answer was a firm "no." TSMC's most advanced node process order books are likely filled up with more important customers—industry watchdogs reckon that Apple usually gets first dibs.

Taiwan's top semiconductor manufacturer leads the market with its cutting-edge lithography techniques. Insiders believe that Samsung was impressed by TSMC's 2 nm trial production runs achieving (rumored) 60% yields. The higher-end Exynos chipsets are normally produced with the best node process available, but missed manufacturing goals have caused Samsung to drop in-house tech. In the recent past, Qualcomm's most powerful Snapdragon mobile chipsets have been deployed on flagship Galaxy S smartphones. Jukanlosreve believes that TSMC rejected Samsung's proposed Exynos deal due to a fear of revealing too many "trade secrets." Potentially, the South Koreans could have learned a thing or two about improving yields—courtesy of TSMC's expert knowledge.



View at TechPowerUp Main Site | Source

 

[Wirko]

So go to Intel, they said they want to be a foundry too.

 

[Selaya]

intel what? 7nm? 10nm?

might asw diy at that point

 

[bonehead123]

Oh Sammy, Oh Qually....how many years have you been making chips now ?

Yea it's hard, but if you haven't figured it out by now, then you deserve to fail....

Of course this will leave TSMC as the only mega-sized, uber-efficient chip mfgr, which seems to be their ultimate goal, cause as soon as they get all their US fabs up & running, they will be pushing out chips faster than stink on dog poo, and then....ya'll can kiss your market share goodbye !

So long, it was nice knowin ya

And yes, I realize that both Sammy & Qually have other revenue streams to rely on, but once you fall behind in the chips race, it's nearly impossible to recover.

 

[Daven]

So go to Intel, they said they want to be a foundry too.

The same 'trade secret' conflict of interest would exist. However, as someone pointed out to me, you would think that TSMC would have the same issues manufacturing Intel chips as they would Samsung chips. That might tell you something about how the industry views Samsung business practices.

 

[Sound_Card]

They need to reject Intel too.

 

[Wirko]

The same 'trade secret' conflict of interest would exist. However, as someone pointed out to me, you would think that TSMC would have the same issues manufacturing Intel chips as they would Samsung chips. That might tell you something about how the industry views Samsung business practices.

There's an interesting point mentioned in the article: it's not only the foundry that can potentially learn the customer's business secrets. It can also be the other way around, and Samsung the chip designer+foundry could extract too many business secrets from TSMC the foundry.

intel what? 7nm? 10nm?

might asw diy at that point

They said they wanted to offer Intel 3 too. But that was long ago, they make their roadmaps out of jelly, so they move constantly.

 

[AsRock]

They need to reject Intel too.

Loosing Samsung or intel regardless if you like either company or not, if they fail prices will go even higher.

 

[kondamin]

How much of this is true and not just some guys shorting Samsung and spreading fud?

If SMIC can poach TSMC engineers so can Samsung

 

[tfp]

percentage yields are meaningless with out knowing chip sizes.

 

[kondamin]

They need to reject Intel too.

Amd and apple should start building their own fabs

 

[Daven]

Amd and apple should start building their own fabs

So AMD had their own fabs when they almost went bankrupt, sold them to become more successful than Intel and now you want them to build them again? Why?

 

[kondamin]

So AMD had their own fabs when they almost went bankrupt, sold them to become more successful than Intel and now you want them to build them again? Why?

I wrote a dumb reply to a dumb comment.

amd went bankrupt Because they build to many fabs at once and also bought ati right before the financial collapse.
just like intel did the last couple of years.

amd would have been fine if they didn’t buy ati and hadnt tried to expand operations as much as they did.
heck they would have gotten away with it if bankers hadn’t given loans for mansions to people who could only afford an apartment.