Intel Discloses $7 Billion Operating Loss For Chip-Making Unit (reuters.com) 82
Intel on Tuesday disclosed $7 billion in operating losses for its foundry business in 2023, "a steeper loss than the $5.2 billion in operating losses the year before," reports Reuters. "The unit had revenue of $18.9 billion for 2023, down 31% from $27.49 billion the year before." From the report: Intel shares were down 4.3% after the documents were filed with the U.S. Securities and Exchange Commission (SEC). During a presentation for investors, Chief Executive Pat Gelsinger said that 2024 would be the year of worst operating losses for the company's chipmaking business and that it expects to break even on an operating basis by about 2027. Gelsinger said the foundry business was weighed down by bad decisions, including one years ago against using extreme ultraviolet (EUV) machines from Dutch firm ASML. While those machines can cost more than $150 million, they are more cost-effective than earlier chip making tools.
Partially as a result of the missteps, Intel has outsourced about 30% of the total number of wafers to external contract manufacturers such as TSMC, Gelsinger said. It aims to bring that number down to roughly 20%. Intel has now switched over to using EUV tools, which will cover more and more production needs as older machines are phased out. "In the post EUV era, we see that we're very competitive now on price, performance (and) back to leadership," Gelsinger said. "And in the pre-EUV era we carried a lot of costs and (were) uncompetitive." Editor's note: This story has been corrected to change the 2022 revenue figure for Intel Foundry to $27.49 billion, as reflected in the source article. We apologize for the math error.
Partially as a result of the missteps, Intel has outsourced about 30% of the total number of wafers to external contract manufacturers such as TSMC, Gelsinger said. It aims to bring that number down to roughly 20%. Intel has now switched over to using EUV tools, which will cover more and more production needs as older machines are phased out. "In the post EUV era, we see that we're very competitive now on price, performance (and) back to leadership," Gelsinger said. "And in the pre-EUV era we carried a lot of costs and (were) uncompetitive." Editor's note: This story has been corrected to change the 2022 revenue figure for Intel Foundry to $27.49 billion, as reflected in the source article. We apologize for the math error.
So (Score:5, Interesting)
My understanding is that when Gelsinger came on board, one of the first things he did was get in touch with ASML and order a whole batch of their latest gen machines. Which take a while to deliver, and even longer to configure for Intel's designs.
Add to this the fact that Intel isn't really competitive with AMD in CPU department and only getting started with discrete GPUs way behind Nvidia and AMD... It makes sense that it will take years to catch up, and until then Intel has to invest a lot of money into its fabs and chip designs to become competitive (again in CPUs, and first time ever in discrete GPUs). But if Intel is successful, they may actually have the advantage of having the newest ASML machines, which should be more efficient than ones TSMC is using without burdens of having to operate legacy EUV systems in parallel with newest ones.
We'll see how it goes, but at the very least Intel got off the "minimizing the costs" train and back onto "we need to try to be industry leaders, and that means investing a lot".
Re: (Score:2)
There are multiple tiers for cpu and gpu. Intel still holds the top of,the x86 cpu market. They were never ever ever ever competitive in gpu. Their gpus are suitable for light office work though in machines without discrete gpus.
So what do you mean by competitive? Raw speed? Performance/watt? Performance/price? Total units shipped? Something else?
Re: (Score:3)
We are sitting in this weird spot that business-class desktop machines that are a decade old can still perform adequately to meet the needs of the user short of things like gaming.
I suspect that the biggest issues are in enterprise computing on virtual machine platforms where both CPU density (and associated heat) and GPU density have become core elements.
Re: (Score:2)
Short of things like AAA gaming. I can play Minecraft just fine on a 10-year-old Dell laptop with a 4-core Intel something or other (not running Windows, obviously). Heck I still have an Acer Aspire One from 2008 which is fine for email and office docs.
Re: (Score:2)
Re: (Score:3)
Intel is second to AMD in all features of the top of the x64 CPU market. Maybe there's some exotic x86 market that I don't know about and that I can't find on search engines that really need the 32 bit instruction set where intel is better. There no such issue in x64 market, where AMD has been a king for several generations.
AMD has much higher core counts, much less energy hungry cores for the performance and much faster interconnect for HEDT and server processors. That's the top of x64 market. Intel's pres
Re: (Score:2)
Servers: ok.
Desktop: Intel rules here at the top, especially in gaming if your metric is pure power (fps at super hi res). I've been shopping and researching for a gaming pc for 6 months. Yes the amd is great for the price. It is not the fastest. It is almost as fast as 4080 in most games and the 4090bkows it away. If rumors hold, the 5000 series will get a big speed boost but amd's next gen will not. By this time next year Amd won't be an option for serious gamers.
Intel gpu: what gpu are you talking
Re: (Score:3)
In reality, intel is dead at the top. They actually admitted to this more or less, and its very visible in gaming benchmarks where they have their most expensive consumer chips routinely losing to AMD chips that cost half but have 3d cache, while in HEDT world AMD's better chiplet+interconnect means that they can just have more cores at the same price, and these cores lose way less performance when having to perform tasks on multiple chiplets.
Intel has simply nothing to offer in terms of 3D cache for forese
Re: (Score:2)
Ok so you're talking price/performance in most of these cases. That's true. For various gaming home use, higher end Intel typically costs more than 50% extra over the 3xd without providing 50% boost and the 4090 is 100% more than a 4080 for a (roughly) 50% boost.
Re: (Score:2)
Not quite. I'm also talking about absolute performance, and reality is that 7800x3D is by far the best gaming CPU right now. There are some games where 14900 and 7950x3D get ahead of it, but for most modern games it will be the best CPU on average. 3d cache is just that excellent, and caught intel completely by surprise. This is why there are games where 5800x3D is ahead of 14900 when coupled with 4090. Not as many as for 7800x3D, but still quite a few.
Re: (Score:2)
Where are you getting your game benchmarks from?
I use Tom's, Anand, and occasionally a few random ones off Google if I don't find what I'm looking for. Your data doesn't match my data. Also, I'm only looking at the games in highest res mode. We're not in sync on something here.
Re: (Score:2)
If I say "hardware Jesus", will you get the reference?
Re: (Score:2)
Expanding on my previous post, today I accidentally came across Intel's own promotional materials for their absolutely fastest gaming CPU ever: 14900KS, that's the one that boosts just a little bit faster than 14900K.
In their own marketing charts (not independent testing, but marketing aiming to sell their own CPU) they concede that your 7800x3D still beats it in some games.
https://imgur.com/a/Vuu5rcT [imgur.com]
Re: (Score:2)
I can't quickly find the game benchmark I'm looking for where they do hires gaming but in the low res benchmarks yes the x3d comes out on top or close enough in most benchmarks.
The 7900 costs more than the i14 but the 7800 costs less and does well enough in most cases.
Anyway I do have a 7800 in my Newegg cart because it's definitely the high end price/performance winner by a lot but if I can find the hi res benches I'll post links later.
Re: (Score:2)
Ok, never mind, I finally found a set of benchmarks for what I was looking for. I was wrong. I don't know what I was remembering but it wasn't what we've been talking about. Humble pie for me.
https://www.techpowerup.com/re... [techpowerup.com]
In almost all cases, as you said, Amd wins, the actual fps are close enough as to be irrelevant in the real world and the 7800 costs a lot less.
Re: (Score:2)
That's an easy humble pie to eat considering you made the right choice, and just got carried a bit by the marketing after the fact :D
But my personal point of hilarity in the intel's own chart is that they put 7950x3d as the baseline to which everything is compared. Because they knew that if they put 7800x3D as baseline, their "super binned absolutely best we can do at the time" CPU is getting trashed by... an older chip that is one price category down.
That's why I saw that chart and thought of you. Have som
Re: (Score:2)
I don't know Anand's story but way back when Tom left, I thought he sold it and moved on. Did they fire him? Nice guy, I exchanged a few emails with him when he was still the guy.
During my search for the One True Benchmark I did find stuff about the 7900 vs 7800 design. Interesting that AMD crippled the 7900 that way yet charges so much more for it. I do wonder what a proper 7900 with full 3d cache would bench at. Maybe they'll do a surprise release update this Fall.
Re: (Score:2)
Technically he sold it and left. But sites like Tom's Hardware are very much tied to the company culture set by the founder. So founder is kept on staff in leadership role to continue.
But as written media continues the decline in the wake of rapidly worsening ad revenues for it, a lot of the "old greats" started running into walls, where their fairly expensive testing regimes and fairly expensive hardware nerd staff was no longer profitable because earnings side went down. So a lot of staff "reshuffling" (r
Re: (Score:2)
I think media quality in general has gone down a lot in the last few decades. In college I read several newspapers every day cover to cover (except the sports section which bored the fuck out of me) but today I find most of it very templated copy paste bullshit and difficult to read. The online versions are no better, but does save trees and transport.
As far as Tom goes, iirc this was a side hobby for him at first. He was some sort of medical doctor I think. I always assumed he sold and just went back t
Re: (Score:2)
Gamers nexus, very nice site. Thanks for pointing me at it.
Re: (Score:2)
Intel is second to AMD in all features of the top of the x64 CPU market....In consumer desktop space, intel is basically dead
AMD cant really compete performance wise. They were the fastest a couple of gens back but thats over now. Their epyc cores won out on multi-core performance
https://www.cpubenchmark.net/s... [cpubenchmark.net]
The only thing they can show for it is marginally higher instruction per clock, which is largely irrelevant in HEDT
you have to be baiting me here - performance largely irrelevant in the High End Desktop sector?
and its hardly insight to say AMD are better performance per watt etc because thats been AMD's whole strategy for the past 25 years
Re: (Score:2)
1. Citing single thread as the main relevant thing in 2024, when almost everything is parallelized and single hardware server is commonly split into multiple virtual machines.
2. Citing cpubenchmark.net.
3. Claiming that AMD's whole strategy for 25 years was performance per watt.
You could try to be more anti-reality if you tried. But you'd have to push really, really hard into flat earth territory.
Re: (Score:2)
1. games, ya fuckwit.
2. edgelord
3. didnt say that
Re: (Score:2)
You:
>and its hardly insight to say AMD are better performance per watt etc because thats been AMD's whole strategy for the past 25 years
Me:
>3. Claiming that AMD's whole strategy for 25 years was performance per watt.
you:
>3. didnt say that
I rest my case.
Re: (Score:2)
AMD are better performance per watt etc
what does etc mean
Re: (Score:2)
1. Citing single thread as the main relevant thing in 2024, when almost everything is parallelized and single hardware server is commonly split into multiple virtual machines.
and dont forget, you actually typed that and submitted it
Re: (Score:2)
> Intel still holds the top of the x86 cpu market.
LOL. Found the User Benchmark shill. /s
AMD: HEDT called. They want their Intel propaganda back.
Intel hasn't been top of the x86 market since the 50% price gouging/drop of the i9-10980XE from the previous i9-9900K when Ryzen and Threadripper started eating Intel for lunch in 2017. Even shill site User Benchmark is blind to how faster Ryzen is when they changed their multithreading scoring algorithm because AMD's Ryzen 3000 was making Intel look incompeten
What does "competitive" mean? (Score:2)
Add to this the fact that Intel isn't really competitive with AMD in CPU department and only getting started with discrete GPUs way behind Nvidia and AMD
What does "competitive" mean? If we're talking stock price appreciation, then Nvidia and AMD are leaving Intel in the dust. If we're talking benchmarks or reviews or Slashdot fandom, Intel is once again behind. If we're talking market share, then AMD has never been competitive with Intel, in any CPU segment. It's not even close. In fact, if we're talking market share and not benchmarks for GPUs, AMD hasn't been close to competitive with Nvidia for a decade. In terms of GPU market share, AMD is much cl
Re: (Score:2)
Market share is meaningless when your margins are in the toilet.
Re: (Score:2)
Market share is meaningless when your margins are in the toilet.
Mmmmmmmm... They said that about Amazon for quite a few years before its margins improved and it surfaced triumphantly with a near-monopoly.
Jeff Bezos didn't get nearly $200 billion from a loss-making business.
Re: (Score:2)
Bezos didn't get nearly 200 billion. His property is valued at nearly 200 billion. Those are completely different things.
Re: (Score:2)
Amazon was busy killing competition. Intel is bleeding money to maintain market share they once commanded with 40-60% margins. Big difference.
The datacentre market has been particularly bad for Intel as of late. They practically have to give their chips away to larger customers just to stay in the server room, and even that strategy is failing them.
Re: (Score:2)
Industry leaders that are conducting basic research should be developing their own manufacturing equipment to produce chips using those revolutionary processes/scales that are supposed to be a major component of being industry leaders.
That they're buying machines for their flagship products indicates that they're no longer the leaders they used to be. It would be one thing if they ended up buying machines for expanding their legacy products offerings simply because their own in-house machines are worn out
Re: (Score:3)
What are you talking about? ASML has dominated the semiconductor field for years. TSMC, Samsung, and Intel all depend on them.
Re:So (Score:4, Insightful)
That they're buying machines for their flagship products indicates that they're no longer the leaders they used to be.
Not necessarily. Every manufacturing process needs specialized machines to produce parts and some companies offer unique (and expensive) best-of-breed machines. For example, if a company is doing metal CNC work they may have a very expensive Tormach milling machine. However, a company might see there is a cost saving by purchasing a larger, faster, more precise 5-axis milling machine [lagunatools.com] if they can justify the volume and precision upgrade.
Same with Intel. They can probably produce chips in-house down to about 15nm (a guess) but for processes smaller than 7nm you have to invest in an extreme ultraviolet (EUV) machine from the Dutch firm ASML. Intel probably decided to do their RND work on their slightly older lithography machines and leave the final masking and wafer production to TSMC.
Now that Intel has decided to move more of the high precision masking and wafer production back to their own factories, they will have to invest in the EUV machines that only ASML produces. Intel's expertise is making computational units. ASML's expertise is making photolithography mask machines. Tormach's expertise is making CAM machines. Just because a manufacturer buys a unique, precision machine from a different company does not mean that manufacturer is not a leader in their area of expertise.
(Wibbly Wobbly, Timey Wimey - that sentence got away from me...)
Re: (Score:3)
Industry leaders that are conducting basic research should be developing their own manufacturing equipment to produce chips using those revolutionary processes/scales that are supposed to be a major component of being industry leaders.
That they're buying machines for their flagship products indicates that they're no longer the leaders they used to be.
What? You clearly have no clue how industry works.
ASML is the company doing the research on revolutionary production methods and scales -they make the equipment for ALL the companies that produce chips. Nobody (successfully) rolls their own...
Intel designs chips. They are buying equipment from ASML to produce the chips they design.
It would be far outside Intel's area of expertise to try to design machinery to produce chips.
Re: (Score:2)
Where do you see evidence for any of that? TSMC has multiple fabs with more coming online every year. There's no indication that they're struggling with newer EUV equipment. Intel isn't moving to anything truly revolutionary anytime soon anyway since they scrapped their plans to utilize high NA EUV for 20a and 18a.
Gelsinger keeps making claims about catching up to and even surpassing TSMC, but from what I can see it appears that he's full of bullshit. Intel 4 is used in the rather disappointing Meteor La
Re: (Score:2)
You're misreading me. I never said that TSMC is "struggling". I said that they have to carry the legacy equipment. Engineer time isn't free, neither is technician time. Every design must be adapted to the litography hardware, and ultimately specific machine. Whereas intel is standardizing on the newer ASML machines, TSMC has to deal with multiple different tiers. That brings with it some costs because sunk costs are very real. But that is also why they have so much advantage of intel right now, when they ha
Re: (Score:2)
All Intel ever had was better manufacturing. Remember Intel came from memory, CPUs were a side-show for them for a long time. AMD, on the other hand, grew making signal processors. Since the ASML machines cannot give Intel an edge over the competition, and they never had anything else, really, than their manufacturing process (and a whole lot of scummy tactics and willfully let their customers get hit by a known security risk), I guess Intel is pretty much done for. Of course, giants die slowly.
Re: (Score:2)
And in real world, the other main advantage of intel has been software. There's a reason why you have to fuck around with AMD machines to make them work properly, why AMD machines still to this day have weird memory quirks and need multiple BIOS updates to work properly with some memory but not other, and why you will still run into other weird symptoms that get fixed with newest BIOS updates.
Whereas for intel, "it just works" is much closer to reality. This is because intel has a long standing dedicated te
Re: (Score:2)
I see you have been thoroughly indoctrinated. A One thing Intel actually does well.
Re: (Score:2)
Exceptional counter-argument. I especially liked your in depth explanation as to how.
Re:So, Taiwan quake to the rescue? (Score:2)
get it all out... all of the bad news (Score:2)
bundle it all up and get it done
refocus
execute
bundling in a large FPGA or a wicked fast AES accelerator (multi stream dedicated ) would help their DC business
smaller profit != loss (Score:1)
Re: (Score:1)
They have a loss. They have way lower revenues as well. No one is even talking about profits. Do you know the difference between profits and revenue?
Revenues down 31%? (Score:1)
I think they mean that revenues were down close to 70%.
FDIV (Score:4, Funny)
> "The unit had revenue of $18.9 billion for 2023, down 31% from $63.05 billion the year before."
I think I see why Intel is having money problems.
Re: (Score:3)
The journalist appears bad at math.
If my math holds, $18.9 is 31% of $63.05, not a drop or 31%.
Going from $63.05 to $18.9 is a drop of $44.15 and $44.14 is about 70% of the initial $63 To me that is a 70% loss in revenue.
I'm not a math major but fractions, right?
Re: (Score:2)
"If my math holds, $18.9 is 31% of $63.05, not a drop or 31%."
Down to 31% of its previous value is what they meant.
Re: (Score:2)
Re:FDIV (Score:5, Informative)
I can happily inform you that the Reuters article that was quoted in this summary is the source of the incorrect number. They have posted a statement mentioning the correction on their page.
Re: (Score:2)
Sounds like they're still using Pentium PCs in the accounting department with old copies of Excel as a cost saving measure.
Re:FDIV (Score:4, Informative)
Go AMD! (Score:1)
This is good news for sure. Intel took us to the cleaners by reselling us the same power hungry crap for years and now it is payback time.
Re:Go AMD! (Score:4, Insightful)
Payback? They're just a corporation. You can buy what they sell at the price they ask or not.
Re: (Score:2)
Business customers forced into an upgrade cycle stuck between an out of date OS and a CPU which didn't support TPM2 and a vendor who isn't AMD happy disagree with you.
There is a concept of a "captive market" in business. You would do well to look it up before you say that corporations' products are optional.
Re: (Score:2)
No one was forced to upgrade or use tpm.
It's been 20+ years since an office machine needed an upgrade to handle email business documents or surfing.
You don't need windows 11, tpm2, office 2024, 64 gigs, a 4090 gpu, and i14 to do email.
Re: (Score:2)
Or you can buy from a competitor now that there are other serious options in the market again, aka "voting with your wallet". And those votes can be payback for prior transgressions when Intel had no significant competitors and rested on their laurels. I fail to see how your statement refutes the OP - not buying products from a company can very much be a form of payback.
Re: (Score:2)
I make business decisions. I don't spend my money on payback for a corporation's previous "sins" like cashing in when they have the best in market products.
Re: (Score:2)
Everyone makes purchasing decisions, not just people "in business".
If that works well for you, then great. But the reputation of a corporate brand is important to many other people - it's why advertising is such a gargantuan industry - and it's why committing "sins" can have detrimental effects on a corporation for many years. Consumers don't easily fo
Re: (Score:2)
My purchasing decisions are based on what my current options are at the buying moment to get the highest quality/performance/whatever per dollar within my budget. That is a business-like decision even if not for a business end user.
I do not understand this concept of punishing a company that in the past cashed in by having superior products. What exactly are we punishing them for? What was the sin?
Good. (Score:1)
I like healthy competition and I don't want Intel to die off but they have been anti-competitive assholes from day one, so fuck them and fuck the people that invest in them.
I used to work at Intel in the early 2000s (Score:5, Insightful)
As an employee I always marveled at the vary large number of significant "bad decisions" a company could make and still stay in business. Six years into the gig, I left after learning how not to do many things. Those lessons have served me well. At my current job whenever we are debating the merits of one approach over another, I sometimes throw out the, "you know back at Intel..." and my coworkers immediately know I am about to tell them what not to do.
I have worked at two other very large businesses and know that not all large businesses are this bad at making decisions.
In some ways Intel reminds me of France right before World War II. Old guard. Thinks they know best. Haven't kept up with time. And very stubborn. One British commander said of the French leadership at the time, "There hardly seems to be any mistake the French did not make". Seems about right for Intel during the last 25+ years.
Re: (Score:3)
No surprise here. Look at Microsoft, Boeing, Credit Suisse and others. Over a certain size, it becomes almost impossible to fail unless you have a reasonable competitor and screw up your own core business really bad at the same time. Microsoft still does not have that competitor, even though they try very hard to screw everything up they do, including a full Azure/o365 compromise last year due to sheer incompetence. Boeing still has its military business (for now). Credit Suisse is effectively dead after a
Re: (Score:2)
By any measure MS continues to grow. By market cap, MS is the most valuable company in the world. They do have competition in all segments. On the cloud, they weren't even the first movers. They have competition from Amazon, Oracle, Google and others. Yet their cloud revenue is higher than Amazon and others are a distant 3rd. Their business suite revenue continues to grow despite free alternatives. They have competition from Apple and Google in their personal computing business (Chromebooks and Chrome OS, M
Re: (Score:2)
https://news.slashdot.org/stor... [slashdot.org]
At some time their crap will come crashing down. It already has started to crumble in ways they cannot fix.
Re: (Score:2)
https://news.slashdot.org/stor... [slashdot.org]
At some time their crap will come crashing down. It already has started to crumble in ways they cannot fix.
Yes nothing lasts for ever, if that's what you mean. No it has not already started to crumble by any measurable way.
Re: (Score:1)
Re: (Score:2)
In some ways Intel reminds me of France right before World War II. Old guard. Thinks they know best. Haven't kept up with time. And very stubborn. One British commander said of the French leadership at the time, "There hardly seems to be any mistake the French did not make". Seems about right for Intel during the last 25+ years.
Success can be crippling.
Wait a second... (Score:1)
The most stupidly over-priced CPU on the market isn't making money?
Who broke the Matrix?
Quick! Write them a check! (Score:2)
We must immediately get the United States Government to write them a check. An unprofitable business? TAXPAYER MONEY IS MEANT TO COVER LOSSES!
That is not a math error. (Score:2)