News Intel's new CEO warns employees about 'tough decisions', but Wall Street cheers

Page 2 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
There is no reason to make an arm cpu, for either amd or intel, why on Earth would they make a competitor to their own products?! And one that has to compete against so many other companies especially.
Also anybody can get a license from ARM as long as you pay for it.

Intel want's that fab money from arm though (and anybody else) ,no questions there.
That's the genius part about not having all your eggs in one basket. When something new comes out that disrupts the market, you won't be left in the dust. Just look at how long it took Intel to come out with GPUs when Nvidia was milking it with crypto-NFT-AI.
 
  • Like
Reactions: bit_user and King_V
That's the genius part about not having all your eggs in one basket. When something new comes out that disrupts the market, you won't be left in the dust. Just look at how long it took Intel to come out with GPUs when Nvidia was milking it with crypto-NFT-AI.
Yeah, it took them so long because they had to make them at tsmc during a time of maximum demand where tsmc was booked out for many months, just as they are now, that's one of the mayor reasons that intel makes the new fabs in the first place, to be able to produce anything they want anytime they want.

Also what are the eggs and what is the basket in your case?
Nvidia only produces at tsmc so what eggs do they have in what baskets?!
 
If he's coming in to implement an austerity plan, then Intel is dead. You can mark the time of death here - it may not result in a bankruptcy/shutdown/takeover for 5-10 years yet - but it's a zombie that just hasn't fallen over yet.

If Intel was going to rise again, it would need to retake and sustain process leadership - which is not going to happen if the #1 priority is cost-cutting. They are not TSMC, they are not even AMD or other failed competitors like Cyrix. They are Intel, and it's highly unlikely that Intel will ever exist in a version that knows how to survive on thin margins in 2nd or 3rd place. That's a different company - maybe the one that buys their assets after they go belly-up.
 
I still think Gelsinger was good for Intel, but not perfect and Tan's complaints when leaving seem valid: https://www.reuters.com/technology/...nces-over-chipmakers-revival-plan-2024-08-27/ and addressing them seems like it will be good for at least the foundry part of the business if not the whole as well. 1 page link and it has some good stuff.

I imagine Tan will try to fix the things he stated as problems. Some of which, like bloated, bureaucratic middle management that impedes progress and does not contribute to engineering efforts may be difficult to get rid of if you are playing mr nice guy. And his goal of making the foundry business more customer centric also seems good. How well he will perform his execution of these goals is yet to be seen though.

Maybe we will see some Adamantine cache? Maybe we will see some high profile customers of 18A?
Maybe we will see an Intel that can adapt faster.
 
Funny that 99% of the time when theres an executive blunder or the higher ups f.. up the companies..

employees are always the ones having to "toughen up" and get the brunt of everything.
 
If he's coming in to implement an austerity plan, then Intel is dead. You can mark the time of death here - it may not result in a bankruptcy/shutdown/takeover for 5-10 years yet - but it's a zombie that just hasn't fallen over yet.

If Intel was going to rise again, it would need to retake and sustain process leadership - which is not going to happen if the #1 priority is cost-cutting.
It seems to me that your statements mostly apply to IFS, where I tend to agree that their main focus has to be delivering their roadmap on schedule. If you don't do that, then it doesn't matter how much money you did or didn't lose, because you're not going to attract any customers except maybe some bargain hunters / bottom feeders.

I think they'll probably always be a world-class design house. At least the cores are world-class. Maybe SoC design could stand to improve.
 
Becoming Intel CEO at this point is setting oneself up for failure. Intel can't be saved anymore. Lip-Bu Tan most likely knows that, but he won't care when he gets fired 3-5 years from now, after he has raked in all those millions of $ in salary and bonuses.
 
Becoming Intel CEO at this point is setting oneself up for failure. Intel can't be saved anymore. Lip-Bu Tan most likely knows that, but he won't care when he gets fired 3-5 years from now, after he has raked in all those millions of $ in salary and bonuses.
They are almost ready to produce 18A, at the end of the year/start of next year they are going to steamroll the market with CPUs and GPUs.

That will be the point where we will see if they will make tons of money or loose tons of money, but go on and show me one person that doesn't like a good deal, intel won't have to pay tsmc like everybody else does so their products will be cheaper than anybody's else.
 
  • Like
Reactions: rtoaht
Joking aside, there are some obvious ways he can do the chopping, but with very significant impact. One immediate one is do the "Alphabet" strategy and make the foundries and design houses their own spin-off Companies under the same parent, still Intel. Kind of like Altera is right now. Pros and Cons, but at least should be the path of least resistance.
Wasn't this done a few years ago? Lot's of articles about Intel separating design from manufacture etc. as part of the Intel fab-for-hire plan.
 
It's necessary to counter the high cost of building new fabs, it's not a rash decision it's continuing what has started when the fab constructions began.
Also counting your beans is extremely important unless you own a bean farm that can produce infinite beans at any moment.
The new CEO is supposed to be some finance guy, so I assume he is going to do what he is good with to try to reduce bleeding. However, deep cuts in headcount have severe repercussions that may help the P&L look good for a time. Most of the cuts will impact people doing the job and so resulting in knowledge loss. Even at the management level, morale will be low and people looking to bail out of the sinking ship. So yeah, Intel may have money to construct their fab but will eventually have to pay more to hire people back to do the work.
 
They are almost ready to produce 18A, at the end of the year/start of next year they are going to steamroll the market with CPUs and GPUs.

That will be the point where we will see if they will make tons of money or loose tons of money, but go on and show me one person that doesn't like a good deal, intel won't have to pay tsmc like everybody else does so their products will be cheaper than anybody's else.
Being able to "mass produce" 18A does not mean they can produce enough and with good yields. For example, Samsung said they are mass producing their 3nm, and we know now that the yields are bad. And if clients like Broadcom already sound the alarm that the yield is poor, it does not sound good to me. I do agree that it will be great Intel find their footing in the foundry space so that we don't pay TSMC through our noses. TSMC can charge whatever they want now, and the cost will be passed over to us.
 
  • Like
Reactions: bit_user
The new CEO is supposed to be some finance guy, so I assume he is going to do what he is good with to try to reduce bleeding. However, deep cuts in headcount have severe repercussions that may help the P&L look good for a time. Most of the cuts will impact people doing the job and so resulting in knowledge loss. Even at the management level, morale will be low and people looking to bail out of the sinking ship. So yeah, Intel may have money to construct their fab but will eventually have to pay more to hire people back to do the work.
You talk as if it's guaranteed that he will fire like half the people or something, relax a bit.
If they really need a huge amount of money they could spin off smaller parts of the business and sell them as whole businesses.

They could also just get rid of useless departments and make working there nicer than before boosting the morale.

It doesn't make sense for us to talk about something we have no idea about how it's gonna go.
Being able to "mass produce" 18A does not mean they can produce enough and with good yields. For example, Samsung said they are mass producing their 3nm, and we know now that the yields are bad. And if clients like Broadcom already sound the alarm that the yield is poor, it does not sound good to me. I do agree that it will be great Intel find their footing in the foundry space so that we don't pay TSMC through our noses. TSMC can charge whatever they want now, and the cost will be passed over to us.
Yes, as I said, that's when we will know if they are going to win or loose a lot of money, obviously if the yields are lower than they need to get any margin then they will lose.
 
He's not wrong about middle management. Too many companies have too many layers between the chiefs and individual contributors that largely just results in internal bickering over turf
This is exactly Intel's problem. For years, middle managers have curated fiefdoms for their areas of responsibility. They reward those loyal to them and punish those who have 'radical' ideas like improving cooperation with other groups. Surprised Keller didn't die from a stroke trying to fix this problem for CPU IC uarch, design, simulation, implementation, etc - each were kingdoms unto their own.
 
Wall Street analysts see Lip-Bu Tan's appointment as Intel's CEO as good news, but Intel employees are preparing for a major overhaul.

Intel's new CEO warns employees about 'tough decisions', but Wall Street cheers : Read more
As past Intel employee, this is nothing new for Intel employees; working under rumors of layoffs. Pat was supposed to be their saving grace. IMO, too many bosses and not enough workers. Intel is more worried about its hiring targets.

I worked at Folsom, CA., the pandemic sent non essential workers home which, and from what I saw was over 2/3 of site capacity. Cube city was no longer, it was desolate. You would walk through buildings and could hear a pin drop.

Intel spent tons of money to renovate work areas to get people to come back and that didn't work at all. It was such a waste! That place was so toxic, people quit before returning. I drive by from time to time and it's still like a ghost town in the parking lot compared to pre pandemic. Split the company off to AMD and TSMC.
 
Last edited by a moderator:
More garbage from Intel. How about you people kindly STOP putting useless thunderbolt ports EVERYWHERE and begin put USB 3.1 ports on to laptops. Literally NO ONE cares your paid-to-put stuff.

My ASUS laptop have two thunderbolt ports, instead of that hot trash that could been have more USB ports instead. Not everyone is a Graphics Designer. Some of us just wish to have more ports aka USB.

Keyboard/Mice and that's it. Cannot wire in my external SSD. So thanks.
 
They are almost ready to produce 18A, at the end of the year/start of next year they are going to steamroll the market with CPUs and GPUs.

That will be the point where we will see if they will make tons of money or loose tons of money, but go on and show me one person that doesn't like a good deal, intel won't have to pay tsmc like everybody else does so their products will be cheaper than anybody's else.
Intel server CPUs are still made 100% by Intel, like always. Recently, Intel did some price-cutting on these (rather uncommon, for them), which managed to squeak them under AMD's launch prices on Turin.

So, let's see how they're faring in the market. The 128-core Zen 5 (no 5C) EPYC 9755 can be purchased through an authorized reseller for $9462. Meanwhile, the 128-core Xeon 6980P costs $13275 from the same reseller.

If we go down to 96 cores, the EPYC 9645 goes for $7431, while the Xeon 6972P cost $10889. Even if you use Intel's repriced MSRP instead of this reseller's Xeon prices, the AMD models are still considerably less.

Nodes are similar: TSMC N4P vs. Intel 3.

Looking ahead to Intel 18A, we've heard that Backside Power Delivery will be more expensive, to the extent that not all 18A chips will use it. That doesn't sound to me like 18A is going to be dramatically more economical for Intel. In fact, if it can't find more external customers quickly, fab costs seem likely to drag down the company even further - far from being a bargain!
 
Last edited:
If they really need a huge amount of money they could spin off smaller parts of the business and sell them as whole businesses.

They could also just get rid of useless departments and make working there nicer than before boosting the morale.
Which parts could be spun off? Which departments are useless? Gelsinger already made all the easy cuts (plus some hard ones).
 
  • Like
Reactions: thestryker
Intel server CPUs are still made 100% by Intel, like always. Recently, Intel did some price-cutting on these (rather uncommon, for them), which managed to squeak them under AMD's launch prices on Turin.
Are you still living 20 years ago when x86 was making big money in servers?!
Intel makes 2-3 times the money from consumer that they do from server.
Intel sees server the way nvidia sees gaming, just release anything at any price just to keep a foot in the door, people will still buy it.
 
Are you still living 20 years ago when x86 was making big money in servers?!
Intel makes 2-3 times the money from consumer that they do from server.
Intel sees server the way nvidia sees gaming, just release anything at any price just to keep a foot in the door, people will still buy it.
2024 revenue:
$18.6b - DCAI + NEX
$30.3b - CCG

To suggest Intel only wants to "keep a foot in the door" is silly at best.

Intel didn't used to separate out enterprise/client because it was all under product type, but it wasn't until the 2010s that enterprise started catching up to client. It has swung the other way again more recently due to a competitive AMD.
 
  • Like
Reactions: -Fran- and bit_user
More garbage from Intel. How about you people kindly STOP putting useless thunderbolt ports EVERYWHERE and begin put USB 3.1 ports on to laptops. Literally NO ONE cares your paid-to-put stuff.

My ASUS laptop have two thunderbolt ports, instead of that hot trash that could been have more USB ports instead. Not everyone is a Graphics Designer. Some of us just wish to have more ports aka USB.

Keyboard/Mice and that's it. Cannot wire in my external SSD. So thanks.
TB is completely compatible with USB so that's a weird rant to make.

Perhaps you're upset over the shift to USB-C ports instead of having USB-A?

If that's the case get used to it as all high speed/power USB ports are exclusively USB-C. This is what low cost USB hubs and adapters are for.
 
  • Like
Reactions: King_V and bit_user
Are you still living 20 years ago when x86 was making big money in servers?!
Intel makes 2-3 times the money from consumer that they do from server.
Your point was about product pricing and the advantage they'd get from doing their own manufacturing. The scale of their server CPU business is more than enough that your point ought to be evident there, yet the data I can find doesn't support it.


Intel sees server the way nvidia sees gaming, just release anything at any price just to keep a foot in the door, people will still buy it.
That logic doesn't apply to anyone but Nvidia, right now. Nvidia is in a class of its own, with respect to their AI training products.
 
  • Like
Reactions: King_V
2024 revenue:
$18.6b - DCAI + NEX
$30.3b - CCG
So that's not close enough to twice as much for you or what?!
Your point was about product pricing and the advantage they'd get from doing their own manufacturing. The scale of their server CPU business is more than enough that your point ought to be evident there, yet the data I can find doesn't support it.
Wow they are already making servers with 18A?!?!?!
Because I kinda remember that my point was about 18A.
It was also about flooding which is not going to happen with CPUs in servers since servers have moved on to ARM and GPU since forever ago.