> Through our voluntary early retirement and separation offerings, we are more than halfway to our workforce reduction target of approximately 15,000 by the end of the year. We still have difficult decisions to make and will notify impacted employees in the middle of October.
Gonna be the finest engine you've seen since the industrial revolution. Grease those gears, guys, we're shedding a head count of no less than 15,000 to keep this baby going. And that's just this year.
That's not easy. Apparently the amount of employees they have, it's difficult to innovate. I'm not surprised. Innovation is easier in smaller groups of people.
There is a little bit of "when a measure becomes a target, it ceases to be a good measure" with respect to chasing financial performance.
While you can't escape thinking about financial metrics, the goal should be something like creating great products, building a competitive barrier etc. Financials can act as a constraint rather than a goal.
A concrete example is Costco.
Even here, Gelsinger puts it last, which is sort of reads like a constraint. Seems fair.
Yup, sure. I'd argue one of the factors involved in the long term problem is when the company starts trying hard to make money as opposed to serving customers well (again, financial metrics must be a constraint). It's not the only factor (incentives get whacky, bureaucracy is difficult), but it's a factor which isn't appreciated as much as the other two.
It'd be nice if they could give me a compelling reason to upgrade my computer more than once or twice a decade, other than 'Our new AI computers have keyboards that go to 11'.
Man this is sad. My read is that they’re throwing in the towel and they’re going to milk their x86 duopoly and government subsidies through a steady, managed decline into irrelevance.
IBM may not be the world leader any more in a number of ways, but that doesn't mean they're not relevant to all of their customers. So maybe the x86 will just limp along, but it's still doing pretty well. By the end of the 80s, the RISC guys were going to dethrone the CISC monsters. It's taken them decades to get this far.
> A key priority for Intel Foundry is to increase our capital efficiency. <...> Now that we have completed our transition to EUV, it’s time to shift from a period of accelerated investment to a more normalized cadence of node development and a more flexible and efficient capital plan.
That is, they no longer see a way and will no longer try to turn money into real innovation, at least as far as chip manufacturing is concerned.
"no longer try to turn money into real innovation"
because they can't. they've failed at every turn which is why they are in this predicament. They let a company with less resources, fewer engineers, no fab trounce them. It's like Intel hasn't done a thing for years and years when they were far ahead of the game. But CEO gets to walk away with MILLIONS while losing the company BILLIONS. If you ever wanted to look up the meaning of loser, look no further. And Pat is left to fix the mess. What's he got to work with? A bunch of losers.
I disagree. Intel seems to be as good as either one at chip design, but they are hamstrung by behind behind in manufacturing. They haven't really struggled with performance, its performance per watt and cost that are their problems. Both of those are mostly fab issues.
Lunar Lake, built on TSMC, looks really competitive. Both with Apple, and with AMD (though AMD has had constant struggles with shipping laptop chips to the point they aren't even seriously considered for large scale enterprise deployment).
On the product design side, they have far more employees than their competitors, and they still produce a lower quality product. When you translate that to finances, it means their gross margin is low (lots of employees means high cost).
On the foundry/fab side: Same story. The amount of revenue the fab generates per foundry employee is tiny compared to TSMC - and it's not because Taiwan is cheaper.
Intel was behind TSMC. When Pat became CEO, he set his vision of "5 nodes in 4 years" to get ahead of TSMC. Next year will be the 4th year, and so far that plan is panning out. But doing 5 nodes in 4 years is unprecedented, and the only way they achieved it so quickly is by throwing all kinds of resources at it (e.g. lower than acceptable yields, with Intel eating up the costs of the low yields). This is what he means by "accelerated investment".
So what he is saying in the quote above is: Now that we are near the end of the 4 year period, we need to focus on improving the efficiency of the fab to be competitive with TSMC's prices.
After a few nails in its coffin, a Chinese co. will acquire it, and then the West will wonder how they could let such foundational tech be lost to a competitor nation.
And, if this, in fact, is the strategy, perhaps the government should decline to provide the subsidies and instead reallocate the money to younger hungrier companies. The subsidies are supposed to be to maintain long-term American competitiveness, but to me it's fairly clear that Intel is not going to provide that even with the government money.
They have 125,000 employees and 13B in revenue. TSMC, Nvidia, and AMD combined (fab, GPU, CPU) have around 88,000 employees on 710B in revenue. Intel may honestly have far more headcount than they need.
Intel also have or had a wide variety of side businesses - compute sticks and NUCs, NICs (WiFi, Ethernet, optics), datacenter SSDs, and probably other things I'm forgetting.
Their strategy is to focus on 18A by H1 of next year, which will achieve parity with TSMC on process again. Then they figure the fab business will work itself out since everyone buying chips wants an alternative to TSMC that is competitive. Even if 18A slipped to H2 of next year that is when TSMC 2nm is expected to enter volume manufacturing so they would still achieve parity.
18A was already rolling down the tracks. Either it'll be great or it won't. But if it is, Intel needs to be poised and willing to make the investments to capitalize on it.
Instead what we're getting with this message is "we want to save 10 billion and everything that isn't 18A go-to-production is on the chopping block".
Maybe instead of cutting to hit arbitrary financial performance numbers that they set for themselves in the hopes of impressing investors, they could explain to the investors why 18A is such a game changer and how Intel will capitalize on it to regain their crown.
> Instead what we're getting with this message is "we want to save 10 billion and everything that isn't 18A go-to-production is on the chopping block
This comment (https://news.ycombinator.com/item?id=41588241) explains the financial predicament pretty well with regards to saving money. Intel is far more "wasteful" than their competitors, and they cannot survive that way.
Intel fabs is set for a loss of $8B this year. I believe next year the forecasted loss is over $6B. They need to both get more customers to increase the revenue, and save money. The former alone will not get them to profitability.
> Maybe instead of cutting to hit arbitrary financial performance numbers that they set for themselves in the hopes of impressing investors, they could explain to the investors why 18A is such a game changer and how Intel will capitalize on it to regain their crown.
They do that every time they talk to the press. Wall Street understands it already. There's just skepticism that they will get as many customers as they need.
Ok. Let me ask you this: You have a company that is less profitable than its competitors. It has worse products. It has a significantly higher number of employees.
Would you bet on them if they insist on continuing to be wasteful? Even if they make a product that competes with NVIDIA once day, you know their gross margins are still going to be pathetic with that much fat. Why would you buy their shares? Why would you want to work for them knowing that by being so fat they have less money to pay you and their stock price won't accelerate much?
This isn't a case of Intel not investing in technology. They're not scaling back on research. They're realizing that they were very inefficient for almost two decades because they could afford to be. They no longer can afford such waste.
Is it just me (non-US) that is grossed out by this dehumanizing talk which is somehow en vogue?
Your "waste" is the salary of tens of thousands of smart engineers that put significant chunks of their lifetime into the company to enable their family (probably) a comfortable middle-class lifestyle.
They made Intel no. 1 with outrageous profit margins that fueled many-billion stock buybacks.
The company is troubled because their executives are legendary for their bad decisions.
They missed out or gave up too early on almost all major opportunities:
- Mobile computing
- GPGPU (twice! Rumours are their GPU team was cut after the crypto craze)
- Chiplets for economic manycore CPUs
- AI (remember Optane? The nonsensical AVX-512 strategy?)
- ARM
- Efficiency (higher IPC + lower clocks)
Seriously, how about fixing their awful strategic decision making instead of blaming employees? Do you think a company with out-of-touch leadership will somehow start winning with fewer employees?
> Your "waste" is the salary of tens of thousands of smart engineers that put significant chunks of their lifetime into the company to enable their family (probably) a comfortable middle-class lifestyle.
The only layoffs Intel mentioned specifically are in sales and marketing. Does this change how you feel?
edit: As a reference, last quarter Intel spent ~$4.4B on R&D and ~$1.6B on SGA (sales, general, and administrative).
No, it doesn't. I did not comment because of Intel specifically but because many comments I have read about similar mass layoffs tend to dehumanize ordinary workers for being "useless bloat" while completely ignoring that the root cause of the problems is moronic decisions by higher-ups.
Mass layoffs that unexpectedly upend the lives of thousands of people are basically guaranteed to cause a couple deaths by suicide, ruined families or homelessness so at least have some basic human decency when talking about those that are affected most.
> Your "waste" is the salary of tens of thousands of smart engineers that put significant chunks of their lifetime into the company to enable their family (probably) a comfortable middle-class lifestyle.
Likely, many/most of those engineers didn't play much of a role in Intel's success.
Trust me, I know many current and former employees there. While they certainly have good SW people, they also have absolute trash (e.g. cannot write a function to compute a factorial, cannot handle branches in version control, cannot do anything on the command line - including a basic understanding of directories and how to navigate them).
As I said, when a company is very profitable, it's easy to hire subpar engineers. And they did.
When I say Intel is fat, I'm not exaggerating. They have a lot of poorly capable engineers who are not contributing much.
That's one category. Then another category is hiring a lot of capable engineers for a zillion side projects - none of which are profitable. Ideally you'd move them to existing products that will be profitable, but there's no need for them - those products are already well staffed. What do you want Intel to do? Keep paying them for useless work (which is what they've been doing all along)?
You're not wrong about the stock buybacks and the bad executives. But you're not proposing a solution either. They've raised compensation. They've stopped buybacks. They've cut and then eliminated the dividend.
The reality is the company is bleeding money, and competitors are squeezing them. If they don't lay them off, they face extinction.
Even the 15,000 number is low. Expect more layoffs in future years. Unless they can come up with another cash cow that their competitors cannot, they are not a viable company at this size. The only reason they were so wasteful in the past was their lead in Moore's Law. Now that Moore's Law is mostly dead, other companies have caught up (and temporarily surpassed them). Unless a revolutionary breakthrough occurs in semiconductor fabrication, which only Intel can capitalize on, the good (wasteful) days are gone forever.
I think you're not realizing the reality of the situation. With the exception of GPUs and the AI craze, for everything else you list: Even if they had become number 1 in those categories they would still be poorly profitable and fat. Just look at the competitors in those spaces - none are as inefficient or overstaffed as Intel. Even Nvidia, with their insane profits, remains relatively lean.
All the things you mention aren't why Intel is suffering so much. It's due to the fundamental economics of running a fab, along with the stalling of Moore's Law.
(And BTW, the sales and marketing team got cut 35% - and I bet if they reveal the numbers, you'll find that the cuts from the fab will be less than the 15% target. So it's not just engineers).
Intel is a dog. It needs to be burnt and rebuilt from the ashes at this point. Too much legacy bloat in the org. Would be nice if my local police force went through this sort of thing every once in a while.
I have a lot of hope for Intel getting back in the x86 ring with real contendors. Lunar Lake is looking incredible (the MSI Claw with it looks like a stunning system), Arrow Lake ought to be solid.
I am a little curious to see where Intel goes with data-center chips. They have been expensive and hot, and the many-small-core offerings at least finds efficiency again. Otherwise it's less clear to me what coming up has promise, and gee, it sure seems like Nvidia and AMD both are super focused on that massive data center market.
One thing that was super interesting in this message was what Amazon want's Intel's 18A for. It's not a CPU, they want it for AI fabric? Interesting seeing the switches be the highest demand. Switch chips are normally quite big, yes? Given how much likelier defects are as size increases, that's going to be a hard test - where-as AMD for example has lots of small CCD's it can stack on a interposer. But also Intel has some fantastic advanced packaging that maybe makes them an ideal partner here - maybe EIMB bridges to PHY or on-package optics stuff, what's grown up from integrated Omni-Path (although not Omni-Path itself, that got sold off already). https://www.intel.com/content/www/us/en/newsroom/news/intel-...
Altera has been known but its so sad. They bought them less than a decade ago for $13B and did nothing with it. They really got themselves into a great position for Heterogeneous computing and then just did nothing.
It's depressing. This should be an age of new hardware, with near-memory computing and flexible architecture. Instead it's GPUs for the foreseeable future, Nvidia has a near monopoly, AMD seems uninterested (despite also owning Xilinx), and Intel's throwing in the towel.
I wanted to prototype a systolic array coupled with a DRAM slice to hold the weight matrix, but there's not even DRAM blocks distributed on any of the FPGAs I saw. I guess Xilinx has HBM but still not quite what I wanted.
They tried a lot of things with Altera and failed. The real problem is the price they paid (it was more than $13B). At the time they bought it, Altera had net revenues of $0.5B. They simply paid too much to ever recoup the expense.
> Our AI investments—including continued leadership of the AI PC category, our strong position with AI in data center
I had a lot of hope that Pat Gelsinger being an engineer would lead Intel to a revival. But this is total delusion. Intel isn't even a remote player in AI.
If they can't admit the dire situation that Intel is in, having missed the AI boat almost entirely and even managed to fall behind Apple somehow, they aren't going to find a way back.
They have nothing to offer over Nvidia for AI. They have nothing to offer over TSMC when it comes to their fab aside from being a US based alternative (and taking billions from taxpayers). x86 has nothing new to offer; their insane moves with AVX have fragmented the platform terribly. It's not even easy to ship high performance x86 code these days.
Looks like all this is, is an announcement that they're going to fire a lot of people soon to make their financials look good while the ship continues to sink.
I don't understand why executives don't seem to understand the basics of effective communication. It's not like they don't have access to staff and expertise here.
When you have difficult news to share, get it out first. Be direct and authentic. Say you're sorry, that you messed up (hint: if you are the CEO, every success is partially yours and every failure is partially yours).
Okay, this means reducing Innovation to a bare minimum I guess. It is baffling to me, how this giant company manages to suck at everything they touch. They managed to be unrelevant in every trend over the past decade.
It's really sad to see how far Intel has sunk, back in the good old days they'd lay off 10,000 at the drop of a hat. These days they're so rubbish it's taking them months to deliver those juicy juicy firings. Nice to see they're still doing well with their arbitrary movement of thousands of employees between TLAs though. Man that CCG is looking juicy right now.
I wonder: if my sentiment from https://news.ycombinator.com/item?id=41587866 is correct and Intel is essentially giving up on innovation in chip fabrication, what, if anything, does this mean for AMD? I'm quite ignorant on how AMD works, e.g. do they even have a chip fabrication plants?
If they do, can they compete with Intel's for US govt grants, or has that ship sailed now that Intel got a grant?
If not, is there room for meaningful innovation in x86 chip design?
How about TSMC? Do they now have a monopoly on state-of-the-art chip fabrication?
Ignore HN. Almost every commenter on here is well outside of their depth.
Intel's been putting a significant amount of money into adopting High NA EUV in 18A, which is bleeding edge enough for a decade.
Furthermore, even though TSMC and Fabrication have become hot words on HN recently, it has been a financial dead end for almost 15 years now. This is why TSMC, UMC, and Samsung cornered the chip fabrication market - the margins are very very low and the upfront costs are very very high.
There are 3 core processes to bringing a chip to market
- Design: high margins, low capex, this is where American players like Nvidia, AMD, Intel, Qualcomm, etc really shine as well as their Israeli and Indian offices. This is also basically software at this point, as almost all design is done via EDA/ECAD
- Fabrication: Fabricating the aforementioned designed chips. low margins, high capex, commodified due to Taiwanese and South Korean industrial policy along with recent advances in China for 28nm processes, recently prioritized in the US due to COVID era chip shocks
- Packaging and Testing: packaging fabricated chips. low margins, high capex, commodified due to Malaysian, Thai, Chinese, South Korean, Japanese industrial policy, recently prioritized in the US due to COVID era chips shock.
Before the mid-2010s, American companies like AMD, Intel, and IBM Microelectronics used to do all 3, but the low margins of Fabrication and Packaging caused players like IBM Microelectronics and AMD to almost go bankrupt, forcing both of them to divest their fabrication and packaging capacity to a spinoff called GlobalFoundaries. Intel kept their packaging and fabrication capacity, but it has a negative impact on Intel's R&D capacity in the 2010s.
Now with CHIPS Act money, Intel is modernizing it's fabrication and packaging offerings, and splitting them off into "Intel Foundary Services" while keeping design under "Intel".
There are a lot more intricacies as well (bleeding edge nodes vs lagging nodes, analog chips, memory chips, the entire EDA tooling chain, etc) with entirely distinct market dynamics that deserve a paper unto themselves, but that's a long story and I really really hate HN.
If what you're saying is true, it's a bit sad that all these large-capex industries that are so important for society to have expertise and progress in are not considered promising or competitive.
Conversely, the money that was redeployed into higher margin industries had a net benefit for innovation.
AMD EPYC would not have existed if AMD was stuck in the fabrication and packaging space because they wouldn't have had the money to spend building a dedicated design team, and Nvida's breakthroughs in GPUs would have never happened if Nvidia had to also fabricate and package.
A company cannot have every single subsegment of an industry - it's unrealistic and a bad use of capital.
And high capex low margins industries are bad for workers from a salary perspective. This is why CS became so popular in comparison to EE/CE - the starting salaries are easily 2x as a SWE compared to working as a Fab Engineer.
And it's not a simple question of raising salaries. Becuase margins are very low (within the low double digits at most), a foreign competitor like TSMC, Samsung, and UMC can win by paying peanuts - which is what they did. Engineering salaries at those companies in Taiwan and SK are around $40-70k with 996 hours de facto enforced (and why Samsung employees globally are on strike right now)
They had, but divested[1] as they couldn't keep up. GloFo stopped chasing the cutting edge, but does have 12-14 nm capability thanks to deals with ST and Samsung.
> is there room for meaningful innovation in x86 chip design?
I think there is plenty of room for meaningful innovation. x86 won't ever hold the crown for performance per watt, but could keep building on just raw total performance and extremely wide vectors. Why stop at AVX10? Let's move to supporting 1024/2048 bits or even, eventually, get to full page size vector instructions at 4096 bytes!
Cool, new strategy?!
> Through our voluntary early retirement and separation offerings, we are more than halfway to our workforce reduction target of approximately 15,000 by the end of the year. We still have difficult decisions to make and will notify impacted employees in the middle of October.
Oh right.