Intel says it can fix its 7nm problems, will outsource some manufacturing to external...

nanoguy

Posts: 1,355   +27
Staff member
Why it matters: Intel's latest earnings report is the last with Bob Swan at the helm and looks surprisingly positive given the company's recent manufacturing hurdles. Incoming CEO Pat Gelsinger says the company is on track to deliver its first 7 nm chips in 2023 but at the same time acknowledged the need to outsource some of its chip manufacturing while it catches up.

Intel's latest earnings report is in, and it looks like the company made a record $77.9 billion in revenue in 2020, which an 8 percent year-over-year increase compared to its 2019 performance. At the same time, Intel says it generated a profit of $20.9 billion and paid $19.8 billion to shareholders.

For the three months ending in December, the silicon giant posted revenues of $20 billion and net income of $5.9 billion, both of which are lower than the results from Q4 2019. However, it is also a bit above expectations for public investors and certainly better than Intel's prior projections. The company attributed the results to strong PC sales, particularly at the low end, fueled by a rapid transition to working and studying from home.

Intel's data center group saw a 16 percent drop in sales compared to Q4 2019, but overall sales for the full 2020 fiscal year were 11 percent higher than the previous year. Interestingly, the company's Mobileye self-driving car division saw 39 percent more sales in Q4 2020 than the same quarter in 2019. And while it's still a relatively small part of Intel, Mobileye is quickly laying the groundwork to become Tesla's biggest competitor.

In a statement during the investor call, incoming CEO Pat Gelsinger said Intel is on track to solve its 7 nm manufacturing woes fast enough to allow mass production of a majority of its portfolio of chips in 2023. Gelsinger also warned that Intel would likely have to outsource some of its chip manufacturing to external foundries. We don't know what models the company will choose to manufacture using external process technology, but this could be a gradual transition that will first involve Core i3 production.

Recently, Gelsinger told senior Intel staff that the company has to "deliver better products than certain lifestyle company in Cupertino." He was referring to Apple, of course, but he has yet to develop a plan on how to make that happen. However, Gelsinger is an Intel veteran who previously spent 30 years at the company in various roles, including CTO.

His return is encouraging other veterans to follow in his footsteps. A notable example is former Intel Senior Fellow Glenn Hinton, who previously worked as the lead architect of the company's Nehalem CPU core project, among other things. He'll be coming back from retirement to help on Intel's upcoming projects, and Gelsinger says we can expect more key people to make a return as he shifts the company's focus back to engineering.

In any case, Gelsinger says the use of external foundries is only meant as a way to close any gaps while the company solves its internal manufacturing issues. The goal is to no only close those holes but also to resume "that position of the unquestioned leader in process technology."

To that end, Intel's incoming chief executive outlines four important principles moving forward. The company will have to design and build better products, leveraging existing partnerships with OEMs. It also must become more agile so that Intel may regain its customers' trust. In the case of Apple, that's too late—the Cupertino company is moving to its own silicon and likely never coming back.

The last two guiding principles have to do with company culture, first by ensuring a renewed focus on engineering innovation and then by establishing a data-driven culture that is more in tune with what its customers needs. Gelsinger will take on his role starting on February 15 and has told investors he'll provide more details about the 2023 roadmap once "I fully assess the analysis that has been done, and the best path forward."

Image credit: Drserg

Permalink to story.

 
So many business buzzwords here, reminiscent of every senior leadership statement released to shareholders stating that, "We're definitely gonna do new stuff. Some new stuff, and other stuff. But for now we have no realistic roadmap or requirements available so we're gonna pump this release full of really good sounding words instead."
 
So which external foundry would be stupid enough to give up capacity for a short term gig? Not sure why TSMC would do that.
 
So which external foundry would be stupid enough to give up capacity for a short term gig? Not sure why TSMC would do that.

not only that they have to keep a lot of secrets from Intel over how their foundries work that they might otherwise share with AMD etc, because Intel directly competes with them in The foundry business. I'm not really sure who would be stupid enough to work with Intel.
 
Apple already moved to TSMC's 5nm. Between now and 2023 I'm sure TSMC will have a healthy supply of 7nm wafers. AMD will be onn 5nm in 2022, is it? And with Huawei gone or going, that opens up even more wafer orders for all parties involved. Intel needs a crutch, not an ICU - yet.
 
Its called money and Intel have a lot of it. In business money talks over almost anything. TSMC has deals with tons of companies, if they can help Intel, believe me they will. Again, money can make a lot of things happen.
But there‘s smart business (securing money long term) and less smart business (need a quick sale to boost the quarterly result in order to get a bonus).

Intel is a TSMC competitor in the sense that each Intel die sold means one less die made at TSMC sold.
 
But there‘s smart business (securing money long term) and less smart business (need a quick sale to boost the quarterly result in order to get a bonus).

Intel is a TSMC competitor in the sense that each Intel die sold means one less die made at TSMC sold.
TSMC is already over subscribed, so I doubt they need Intel's business currently.
 
Intel is starting to sound a little bit like IBM before they lost their significance in the marketplace. Unclear direction and a leadership that is .... lacking. They very well may spiral down to 2nd or 3rd if they don't get their goals clearly set and the appropriate leadership to make it all go forward. I will be watching them, but it's far too early to make any investment moves in their direction.
 
not only that they have to keep a lot of secrets from Intel over how their foundries work that they might otherwise share with AMD etc, because Intel directly competes with them in The foundry business. I'm not really sure who would be stupid enough to work with Intel.

...and get slapped with multi-million copyright infringement suite, sure, they are forming a line already...
Please don't open a business. You'll go broke.
 
...and get slapped with multi-million copyright infringement suite, sure, they are forming a line already...
Please don't open a business. You'll go broke.
If it‘s a multi million lawsuit, that would be a good deal.

And TSMC would bring the lawsuit where ? In the US as a foreign company ? Good luck with that.

It seems that larger companies who ignore rules (with the exception of VW) tend to do a lot better than those who abide by them.
 
Some decent signs of progress there. I am still decently impressed with my 11th gen laptop chip for the money but Intel definitely needs to get 10nm to the desktop and get 7nm moving to compete against 5nm. AMD hasn't done anything dumb with there money coming in yet so prob no massive performance swap like what we saw in 2006 coming in the near future. But they gotta get the mass core chips moving at affordable prices and on silicon that doesn't suck down 300 watts.
 
But there‘s smart business (securing money long term) and less smart business (need a quick sale to boost the quarterly result in order to get a bonus).

Intel is a TSMC competitor in the sense that each Intel die sold means one less die made at TSMC sold.
Believe it or not, most of that means nothing in business. Money is the root and if deals need to be made for the benefit of helping you make money, a deal will get done. Just ask Apple and Samsung. They compete in the same space for a decade and have billion dollar deals together. So yea Intel and TSMC can get a deal done if they choose to.
 
If it‘s a multi million lawsuit, that would be a good deal.

And TSMC would bring the lawsuit where ? In the US as a foreign company ? Good luck with that.

It seems that larger companies who ignore rules (with the exception of VW) tend to do a lot better than those who abide by them.
intel would bring the lawsuit it doesn't matter where was something taken from...it could be from garbage can, it's copyrighted...
 
Nah :) I was thinking if someone steals Intel's tech and uses it, Intel would bring pain against everyone involved. I mean, for posterity reasons, AMD would do the same. Anyone would really....
Whether they would be succesfull or not - who knows?! (rhetorical question)
 
For those saying TSMC would be stupid to work with Intel, money is more important. Apple and Samsung compete and were locked into various lawsuits, about to rip each other’s heads off, yet Apple sources its iPhone panel displays from Samsung. Billions of dollars.

Also, Apple and Qualcomm were involved in a bitter legal battle, world wide, with Qualcomm trying to get an import ban on iPhones, yet they squashed that beef suddenly when it became clear that Intel couldn’t deliver a 5G modem in time. Apple signed a deal with Qualcomm, once it’s mortal enemy, for its 5G modems, worth billions of dollars.

Yes it is true that Intel and TSMC are competing foundries, but Intel serves itself while TSMC serves many different companies. If an Intel deal guarantees TSMC billions of dollars, money they can then use to build 3,2,1 nm why not take it?
 
For those saying TSMC would be stupid to work with Intel, money is more important. Apple and Samsung compete and were locked into various lawsuits, about to rip each other’s heads off, yet Apple sources its iPhone panel displays from Samsung. Billions of dollars.

Also, Apple and Qualcomm were involved in a bitter legal battle, world wide, with Qualcomm trying to get an import ban on iPhones, yet they squashed that beef suddenly when it became clear that Intel couldn’t deliver a 5G modem in time. Apple signed a deal with Qualcomm, once it’s mortal enemy, for its 5G modems, worth billions of dollars.

Yes it is true that Intel and TSMC are competing foundries, but Intel serves itself while TSMC serves many different companies. If an Intel deal guarantees TSMC billions of dollars, money they can then use to build 3,2,1 nm why not take it?
Apple signed the deal with Qualcomm because they effectively had no choice - their attempt at kneecapping QC failed because Intel could not deliver and QC did not back down.

If TSMC had excess capacity, sure why not sell it to Intel.
As TSMC is fully booked, they do not need to do this, even if Intel paid more than others as it‘s just a short term gig.

Samsung supplying Apple was simple - they had the extra capacity and if they had not delivered the parts, someone else would have.

Speaking of Samsung, I see them as a more likely outsourcing candidate for Intel.
 
Longest TV show in history. Meanwhile they will release another tweaked 10nm series that requires a new MB socket.
 
...and get slapped with multi-million copyright infringement suite, sure, they are forming a line already...
Please don't open a business. You'll go broke.
Oh, you mean Intel might get another fine that barely affects their profit margin? I'm sure that's EXCELLENT deterrent as their history well demonstrates!
 
Back