Intel Reportedly Ramps "Arrow Lake" Orders at TSMC Amid Internal Foundry Struggles

[AleksandarK]

According to Taiwanese media Commercial Times, Intel is significantly increasing its outsourcing of "Arrow Lake" CPU production to TSMC, a strategic move as it grapples with persistent issues in its own foundry division. This decision to outsource a substantial portion of Arrow Lake's production is a significant shift in Intel's strategy, showing the company's rising reliance on external partners to meet quality and performance demands. The Arrow Lake Core Ultra 200 series is Intel's first major outsourcing initiative, in which Intel gave its core IP to third-party foundries, more specifically for a 3 nm node at TSMC. However, it clearly indicates the performance gaps in Intel's own Intel Foundry and the high demand expectations for the new CPUs. Originally intended to use Intel 20A node, Intel shifted focus of 18A node for its products and upcoming foundry customers.

Intel's recent orders with TSMC extend to its upcoming Lunar Lake chips and next-generation Falcon Shores AI GPUs, both of which will use TSMC's 3 nm process. Although Intel's 18A node remains promising, the company relies on current products to sustain its revenue streams, making TSMC's support crucial in ensuring timely shipments. This increased outsourcing reflects Intel's need to maintain competitive performance in the short term. Once its Foundry division meets performance and capacity targets, Intel aims to bring more CPU manufacturing back in-house. However, if anything goes wrong, Intel could face challenges securing sufficient volume from TSMC, as the foundry has longstanding commitments with major clients like Apple, NVIDIA, Qualcomm, and AMD.



View at TechPowerUp Main Site | Source

 

[Redwoodz]

Ok so how many times are we supposed to believe this reply of "the process is still on track"? Think we have been hearing it for like 10 yrs now? Trying to figure out what's really been going on in those fabs.

 

[usiname]

Its not like they were going to release Arrow Lake on their node in first place so why this is new?

 

[Ravenmaster]

Dafuq are they talking about? Arrow Lake has been a complete disaster. It can't beat last gen performance and its full of software related bugs. Nobody is buying these chips. Why make more when its a failure? It should never have been released in the first place.

 

[Carillon]

Pat wants to put more eggs in the taiwanese basket, because it is unstable, just like raptor lake.

 

[Ferrum Master]

Pat wants to put more eggs in the taiwanese basket, because it is unstable, just like raptor lake.

Maybe he is a red impostor.

 

[Daven]

IFS will probably go the way of Glofo: stuck at a particular node. In this case, Intel 3 will be their best.

 

[tfp]

How is this news? Arrow Lake was already stated to be TSMC produced and not on Intel 20A for weeks now. Increasing orders implies they are selling more than expected not that Intel is producing less. They aren't producing less than zero.

 

[TristanX]

Intel do not gave IP to any foundries. Intel just send masks to TSMC and that's all.
Article on this Commercial Times as well as all broadcasters show only anti-logic fantasies and lack of any understanding reality.

 

[kapone32]

Intel do not gave IP to any foundries. Intel just send masks to TSMC and that's all.
Article on this Commercial Times as well as all broadcasters show only anti-logic fantasies and lack of any understanding reality.

Since when have broadcasters have been right about anything to do with PC?

 

[Eternit]

Ok so how many times are we supposed to believe this reply of "the process is still on track"? Think we have been hearing it for like 10 yrs now? Trying to figure out what's really been going on in those fabs.

Maybe it is on track, but on the wrong one. I stopped trusting their promises when they postponed 7nm once again and renamed 10nm with some plusses to Intel 7.

Its not like they were going to release Arrow Lake on their node in first place so why this is new?

Initially they were supposed to have compute tile in Intel 20A, then they were supposed to use TSMC for the higher end and 20A for the lower end, but now everything will be from TSMC.

IFS will probably go the way of Glofo: stuck at a particular node. In this case, Intel 3 will be their best.

I think they invested a lot of money into 20A/18A so eventually they will have it like with 10nm and 7nm(Intel 4 Intel 3) but it will take them 5 years if they will survive that long

 

[john_]

I thought it was supposed Intel to start using it's fabs with Panther Lake after they scraped 20A and gone 18A only. I don't see an indication of having fab problems, only that they have high demand from OEMs for their current CPUs that are made on TSMC.

 

[Daven]

I think they invested a lot of money into 20A/18A so eventually they will have it like with 10nm and 7nm(Intel 4 Intel 3) but it will take them 5 years if they will survive that long

I'm starting to get the feeling that everything below the so called '3 nm' naming conventions from all three big players (Intel, TSMC and Samsung) is going to be delayed up to 5 years from now.

 

[ThomasK]

Intel is currently worse in the two areas it used to thrive: engineering and manufacturing.

It engineers worse products with a more advanced node than its competitor (3nm vs 4nm) and basically can't manufacture anything on 18A yet, so it has to outsource to TSMC.

Let's see how Patty saves this sinking ship.

 

[Vayra86]

Ok so how many times are we supposed to believe this reply of "the process is still on track"? Think we have been hearing it for like 10 yrs now? Trying to figure out what's really been going on in those fabs.

Last I heard they were demoing their Chips

 

[qcmadness]

Last I heard they were demoing their Chips

View attachment 371472

They produce fish and chips now.

 

[Vayra86]

I'm starting to get the feeling that everything below the so called '3 nm' naming conventions from all three big players (Intel, TSMC and Samsung) is going to be delayed up to 5 years from now.

When has a node not been delayed or somehow changed to meet a projected time to market, an acceptable yield, etc.

There's always trouble.

 

[phanbuey]

honestly this is sad but at least they have a plan B this time-- they needed this for Rocket Lake back during the 14nm+++++ debacle.

 

[ScaLibBDP]

Ok so how many times are we supposed to believe this reply of "the process is still on track"? Think we have been hearing it for like 10 yrs now? Trying to figure out what's really been going on in those fabs.

>>...Trying to figure out what's really been going on in those fabs....

Totally unacceptable Yield Rates of chips ( possibly well-well below 30% ). There is also another problem: Intel is Unable to solve management problems and it affects all the rest units ( first of all fabs ).

 

[Darmok N Jalad]

Dafuq are they talking about? Arrow Lake has been a complete disaster. It can't beat last gen performance and its full of software related bugs. Nobody is buying these chips. Why make more when it’s a failure? It should never have been released in the first place.

They have to release something. Releasing nothing is even worse than just launching something mediocre. Shareholders calls are really rough when no new products are out.

 

[TheinsanegamerN]

I'm starting to get the feeling that everything below the so called '3 nm' naming conventions from all three big players (Intel, TSMC and Samsung) is going to be delayed up to 5 years from now.

That'd be good for us. Last time TSMC had major fab delays we got Nvidia Maxwell, which was a fantastic optimization of 28nm.

>>...Trying to figure out what's really been going on in those fabs....

Totally unacceptable Yield Rates of chips ( possibly well-well below 30% ). There is also another problem: Intel is Unable to solve management problems and it affects all the rest units ( first of all fabs ).

Intel is long overdue for a management purge, middle management on up is horribly bloated and inefficient and repels industrious individuals and those with new ideas, according to glassdoor.

 

[Eternit]

>>...Trying to figure out what's really been going on in those fabs....

Totally unacceptable Yield Rates of chips ( possibly well-well below 30% ). There is also another problem: Intel is Unable to solve management problems and it affects all the rest units ( first of all fabs ).

I don't thing it was only problem with poor yields. You can't improve yields without producing anything and producing at your fab with 20% yields is not a lot more expensive than producing in your competitor high end process. Even if it is a bit more expensive, you are investing it in your own fabs. If you don't produce anything in your fabs, you won't improve their process and you still have to pay salaries of these factories workforce. I think the problem is the same as it has been since 22nm. The early CPU have problems with high frequencies. With every next process its getting worse, the early products are worse and takes more years to fix. Pat expected they will manage to go in one year from 7nm (4/3) to 5nm (20/18) and the result is something like Cannon Lake.

 

[Wirko]

They produce fish and chips now.

That used to be flash nand chips, until it wasn't.

I'm starting to get the feeling that everything below the so called '3 nm' naming conventions from all three big players (Intel, TSMC and Samsung) is going to be delayed up to 5 years from now.

Indeed... And on top of that, a new node only brings a small PPA (power, performance, area) advantage. The concept of "a new full node" has disappeared.

 

[PaddieMayne]

Dafuq are they talking about? Arrow Lake has been a complete disaster. It can't beat last gen performance and its full of software related bugs. Nobody is buying these chips. Why make more when its a failure? It should never have been released in the first place.

You say nobody is buying these chips, yet at alot if not most outlets the 285k is sold out. Sadly not everyone reads tech reviews, they just see the Intel logo and think automatically it's the best and most reliable. If most people did read tech reviews before there purchase AMD would have a much larger market share.

 

[Broken Processor]

Oh boy Intel is going to miss that 40% discount they used to have.

I'm guessing demand is coming from intel partners because they know the average person buying a pc doesn't care about cutting edge gaming and the brand recognition is still strong.