We seldom write about individual companies unless they are close to our banking/finance wheelhouse and have some sort of larger significance, such as are a focus of systemic risk or illustrate larger trends.
We are hoisting a discussion from comments yesterday on the thesis that Intel is exhibiting many of the same pathologies as Boeing is and has a similarly Herculean task in fixing its production problems and salvaging its competitive position. Even worse, Intel lacks the advantage of being in a cozy duopoly in its main line of business.
Here is a quick summary of the Intel mess. That stock price plunge puts it at its lowest level in 50 years:
Intel is cooked
– 15,000+ to be laid off
– $INTC down 29% today
– Dividend suspended
– Foundry business hemorrhaging billions
– Broken Raptor Lake chips unfixable
– ARM taking over laptop market
– ARM taking over data center market
– Totally missed the AI hype train pic.twitter.com/QmqtaiwpM2— Fireship (@fireship_dev) August 2, 2024
A more detailed overview from Vox:
Intel’s bad week really is more of a bad quarter: It started back in April, when the company revealed during an investor presentation that its chip manufacturing unit had, through a series of poor decisions, sustained $7 billion in losses in 2023, on top of a 31 percent decrease in revenue from 2022. Cost-cutting and other measures will save the company $10 billion in 2025, according to CEO Pat Gelsinger….
This is not the first time the company has had to implement cost-cutting measures — Intel did mass layoffs back in October 2022, after a brief, Covid-powered bump in the company’s fortunes.
“In February ’22, they put out revenue targets that — I mean, I use the phrase outlandish, they were ridiculously high,” Stacy Rasgon, senior analyst at Bernstein Research, told Vox. “They were sizing the company and sizing the investments to that COVID level of revenue,” based on the need for technology that allowed people to work from home or for kids to attend school remotely — a business that collapsed nearly as quickly as it arose.
But the current CEO, Pat Gelsinger, inherited a business that was coming off a decade of stumbles when he started in 2021. “He came into a situation that they were dire straits; they had no competitive product to really bring to market,” while Jensen Huang’s Nvidia dominated the curve on AI tech, Daniel Newman, CEO of the Futurum Group, told Vox.
Intel’s other recent big bet has been its foundry business — three facilities in the US and three overseas to manufacture semiconductor chips, with other facilities in Asia and Latin America for testing and assembly. But that’s gotten a bumpy start; for instance, Intel declined to invest in cost-effective extreme ultraviolet machines for its manufacturing facilities, then had to outsource 30 percent of the manufacturing to a rival company, TSMC.
In comments yesterday, readers painted Intel’s plight in darker colors.
Reader Keith started by linking to Timothy Prickett Morgan’s sober take. Representative snippets:
…when Pat Gelsinger, the prodigal and presumed savior of the company that he loved first and loves best, tells Wall Street that this is the most significant resurrection of Intel since it had to exit the memory business for the first time – we’re not talking about 3D XPoint and flash here, people, but DRAM back in 1985 – he ain’t kidding. This might qualify as a miracle when it is all said and done.
In fact, Gelsinger is seeking divine help:
hello I'm new to the stock market is it good when the intel ceo starts praying https://t.co/E4Tb5nzgTr
— DatNoFact ↗ (@datnofact.bsky.social) (@datnofact) August 4, 2024
More from Murphy:
Sometimes, such harsh actions work. IBM, which had its self-described “near death experience” back in the mid-1990s, had tens of billions of dollars of writeoffs – just absolutely unheard of, unimaginable, inconceivable for the original blue chip stock – and laid off 200,000 of its 400,000 workforce as it pivoted to software and services and trimmed its independent systems fiefdoms to get its own costs in line with revenues. IBM eventually got back to 400,000 employees, but the way, and has subsequently sold off a lot of systems and services businesses to focus on being a hybrid platform provider, with Red Hat at the center of that strategy. This is IBM’s fifth rebirth in its Herman Hollerith built punch card machines to do the 1890 census in 1890, which is the true kernel of the company we know as Big Blue.
….This rebirth of Intel is more like a company reminding itself of what it learned to do in the mid-1980s: design good chips and make good chips, and be paranoid enough to survive. …
Intel has even rougher road ahead, but it is, in fact, getting its foundry act together. It is getting its chip designs together and it will be relying less on Taiwan Semiconductor Manufacturing Co for the chiplets in its most advanced CPUs and bring them home to its fabs with its 18A, 14A, and 10A processes. It remains to be seen if others will use its foundry for the 14A and 10A processes, but Intel is clearly getting the third party tooling together from Cadence, Synopsys, Siemens, and Ansys together so they can, and we think it will get some business from chip makers looking for alternative etching and packaging. We think it will have a profitable foundry operation and that it will make good chips that people want to buy as well. But we also think the X86 market is going to decline as Arm rises at the hyperscale and cloud builders, and Intel not only knows this, but accepted this years ago and that is why it must have an open foundry business. If you can’t beat Arm CPUs, you have to make Arm CPUs.
Reader Pearl Rangefinder saw this view as altogether too optimistic:
I think he really, really under-counts just how screwed Intel is. Their entire business model is completely fucked because when Intel was at the top, they basically spent a decade with their thumbs up their asses burning $100 BILLION in share buybacks while their competitors caught up to them and left them in the dust. They are now left with basically zero competitive products in the core parts of their business – their GPUs are way behind Nvidia and AMD, their x86 processors are uncompetitive (and the ones that are suffer from crippling defects which they are so far refusing to recall or even stop sales of ), their datacentre business is a dead man walking, and the chip fabrication wing of the company is uncompetitve with TSMC, the Taiwanese chip fabbing giant that Intel used to be ten years ahead of technologically.
The worst is still to come. Intel’s fabs are so far behind that they have had to slowly switch to using TSMC to manufacture their latest processors. They already use TSMC’s fabs for the GPU part of their processors on the ‘Meteor Lake’ generation, and from rumors and news sources Intel’s upcoming processor families (‘Arrow Lake’ and ‘Lunar Lake’) will be entirely outsourced to TSMC. Intel’s entire business model was predicated on them leveraging their unrivaled in-house manufacturing expertise with their unrivaled processor design expertise, and having enough scale with sales to feed their chip fabrication plants. Scale is king in the fab business, so every dollar they send to TSMC is a dollar not being spent on their own fabrication plants. A vicious downward cycle.
Intel is like the Boeing of the tech world right now, probably even worse TBH. They need to pull multiple miracles in the various parts of their business to save themselves. Nothing over the past decade would suggest that their management is even remotely up to the task.
$100 billion is share repurchases over the last 10 years. Since 1990, the total is $152 billion.
What you describe sounds like voluntarily chosen self-Boeingization by the Intel leadership.
Am I wrong to understand it in that way?
Not at all, the similarities between the two are quite apt I think. The biggest similarity being how both companies incinerated massive amounts of cash in idiotic share buybacks, with management seemingly more focused on financial engineering than looking after their actual businesses. For Intel, it seems even more idiotic when one considers how capital intensive advanced chip foundries are nowadays (at least $10+ billion dollars; TSMC’s 2nm fab is supposedly going to cost them over $30 billion dollars to build!). On the chip design side, it takes years to get processors from the drawing board to production, so you need a lot of money to keep things afloat while you are designing the next generation of products. That also means if you have a less than stellar design, you could be bleeding money for years waiting for the successor product – this specifically is what happened to AMD when it launched its ‘Bulldozer’ micro-architecture in 2011, it nearly sank the company as it could never perform as well as Intel’s products did in those days. AMD started designing the successor ‘Zen’ architecture in 2012, and took four years to get Ryzen/Zen-1 out the door by 2016.
I don’t want to suggest that only share buybacks are the cause of all of Intel’s problems (and Boeing’s for that matter), because they definitely aren’t, but goddamn it, there is a good reason why stock buybacks used to be illegal. Only a sick society lets these finance parasites play these stock manipulation games at the firm’s expense, or survival.
Intel was a bloody cash cow too, their net profit margin could hit 30%+ in a good year, with revenues pushing $70 billion. That is absolutely bonkers for a hardware manufacturing company. Imagine screwing that up?
A savvy friend, before stock buybacks became pervasive, argued, “Why should I invest in a company if management isn’t investing in it?” But the stock repurchases had the effect of making it easier to engage in the slow-motion liquidation we first described in a 2005 article in the Conference Board Review, The Incredible Shrinking Corporation. It had become evident even then that American companies in aggregate were net savers, which was abnormal and a bad harbinger:
Companies typically invest in times like these, when profits are high and interest rates low. Yet a recent JP Morgan report notes that, since 2002, American companies have incurred an average net financial surplus of 1.7 per-
cent of GDP, which contrasts with an average deficit of 1.2 percent of GDP for the preceding forty years. While firms in aggregate have occasionally run a surplus, “. . . the recent level of saving by corporates is unprecedented. . . . It is important to stress that the present situation is in some sense unnatural. A more normal situation would be for the global corporate sector—in both the G6 and emerging economies—to be borrowing, and for households in the G6 economies to be saving more, ahead of the deterioration in demographics.”
This article elaborated on how short-termism, specifically the fixation on quarterly earnings, had become deeply entrenched, and was leading to all manner of dysfunctional behavior. Critically, managements were taking the view that it was lower risk and had much faster payoff to simply squeeze current businesses as hard as they could. Investing in growth not only took longer to produce results and was more hazardous, but nearly all “investments” hit the income statement near term (such as marketing plans, adding to office and/or factory space) which made them even more to be shunned. Shrinkage or low growth in existing business lines was often masked by acquisitions.
Nevertheless, changes in degree are changes in kind. The level of corporate negligence at work for Boeing and Intel needs a name. Any suggestions?
72 comments
Comments are closed.
imho it’s plain, old fashioned hubris.
I found myself on the receiving end of this kind of hubris at one of Intel’s foundry competitors. Although way down the corporate food chain. At the time they simply seemed to big to care.
Thanks for the hoist into this article.
It’s more than hubris, it’s greed.
Intel has under-invested in research, in capital expenses, and has made numerous errors, most notably with their 10nm node. Money could have solved at least some of these.
There’s also the matter that Intel has suffered under many years of bad management, stock buybacks, and now it’s all starting to bite back. In an industry like chips, it’s super capital intensive and once a firm loses ground, it’s very hard to gain back.
Personally I think it should be nationalized – Intel is too strategically important to not be bailed out.
Eat today, starve tomorrow. As a CEO why worry about tomorrow when you will have left the company and your bank account will be full.
And the day after tomorrow you’ll be dead anyway, so why care?
Very true…but ultimately, any company that goes down the DEI path, no matter how much pressure it’s under to do so, will be destroyed by the competitors who are hiring the best people it can get…But the executives figure they are getting Pokemon points for their next executive job, so they go along with this lethal fad…
Detroit Syndrome?
Carly Fiorna (HP), Bob Nardelli (Home Depot), Mr, Gelsinger says hold my beer. Intel joining Boeing in the proverbial Dumspter Fire category? These companies have to view their very existence is at stake, and turn it around lest they leave themselves no alternative.
Beneficial largesse from the federal government Chips Act notwithstanding, Intel looks like the professional sports league franchise holding onto their aging stars. Aka, not facing facts on the ground as they exist. Intel reminds me now of a circa mid 2000s GM, a behemoth of an American manufacturer that has not kept up the changes in the industry.
The post cites an article giving IBM as an example of a firm that was able to “resuscitate” after going through a near-death experience in the 1990s.
Well, IBM is yet another example of a corporation that, despite of maintaining several pockets of world-class competence, has been on the whole drifting into some kind of “me-too” provider of software and cloud services, helped — you guessed it — by massive share buyback programmes (hello Mrs Rometty!) These were reduced to amounts around USD 300M / year only after she left as a CEO.
The “woman power” CEOs have been a pretty reliable indicator to avoid the stock in a company for quite a while….Carly Fiorino wrecked two companies that were otherwise quite viable…
Don’t forget Cisco Systems! Darling of the 1990s!
https://www.nakedcapitalism.com/2023/03/losing-out-in-critical-technologies-cisco-systems-and-financialization.html
Slowly eating itself to death over decades…like Pizza the Hut in the movie, Spaceballs! :)
What if the current ( Age of Legal Stock Buybacks) type of high managers ( CEOs and stuff) are not motivated by the sort of ego-subsuming identification with the Corporate Institution that at least in part motivated managers in the Good Old Days before stock buybackkery?
What if they are the moral equivalent of mafia infiltrators viewing the corporations they infiltrate at the top as bust-out opportunities for their own personal selves and personal families? Would it make sense to refer to ” stock buybacks” as ” stock buybustouts” so we can think more clearly about whose purposes these stock buybustouts are really meant to serve?
If we can’t re-outlaw stock buybacks, is there a single thing we can do about stock buybustout corporate management?
Let’s just call it by its old fashioned name — looting.
I read that reductions in US federal corporate income tax rate in the neoliberal era are another factor relevant in these riches-to-rags stories about American firms. From 1953 to ’86 the top rate was between 52% and 46%. It’s now 21% and some pay a lot less than that by exploiting international accounting. When a business had half its profit taxed away there was a powerful incentive to minimize profit by spending surplus income on R&D, building or upgrading factories, etc. Or so I once read.
Intel has disappointed the last two years’ summer outlooks!
This year’s disappoint is coincident with the wheels coming off the US [questionable data and US’ rudderless, wrt elected officers, government egging on dangerous, unnecessary wars] more than just the economy, with out the “slowly” part.
I sold off my [small] Intel position in January!
Can the US weather this?
Intel like Boeing, Strategic necessities, will be recapitalized by the good ole taxpayer. Socialism!
Overall, a successful scheme to make money. Short term thinking, but skills have been acquired.
I suggest a plant under the Rockies… the err ‘weather’ may get inclement, if the full plan is implemented?
I used to call Texas Instruments a hedge fund that makes chips for a hobby. They spent enough cash buying back their stock that they could have bought Intel before it crashed. But that would have created an even larger hedge fund.
Auction rate securities, anyone?
They are another zombie company — why can’t Wall Street analysts read, analyze and understand financial statements. Like Boeing, Intel is an asset based company — not crappy, bug and security risk laden software. You can only hollow out your physical assets so far in the name of goosing your share price.
But, it strikes me that share price above all else brings additional issues; putting aside the arbitrage of share compensation not hitting the P&L/CF statement, but being deductible for taxes. So many of these business models work when the stock price goes up in attracting human talent. What happens when your stock price sinks (besides potentially creating huge moral problems for employees who paid taxes on stock that vested at a higher price)?
I cannot see Intel pulling off multiple miracles simply because the same management team that is supposed to rescue Intel are the same one that steered it onto the rocks. Does the US government/Pentagon need Intel’s offerings? Will there be a talk of a bailout? But from this post, it does not seem to be a problem that you can make go away by simply throwing money at it. Certainly both parties won’t even talk about nationalization as a solution. So does it go into bankruptcy with the various parts sold to the highest bidder? There would be a freakout if the Chinese turned up for that auction.
If the globalization gambit fails, there’s no guarantee that TSMC will be available to the US chip industry. In that case, Intel has the best fabs the US has. AMD & Nvidia would be in deep doo doo. ASML, the Dutch company that makes the machines to make chips, is still mostly selling to China despite US sanctions. If the dollar collapses, China will be the main semiconductor manufacturer, and Europe will have no choice but to realign. ARM is British but Japanese owned. It’s not hard to license or replicate. The US would be foolish to let Intel collapse.
Re ARM: it will be interesting to see where RISC-V goes. There is a lot of interest in China.
If there is a war over Taiwan, hasn’t the US said it will destroy TSMC. A ridiculous strategy that will have more negative effects on the West than on China.
Still, that’s par for the course these days.
Intel had both chip design and chip process leadership which allowed them near monopoly position. They let process leadership slip banking on design strength. AMD zen shattered Intel design leadership. Zen catchup which started at low end now extends all the way upto Datacenter chips. Intel must get back process leadership else it’s days as very profitable company are numbered.
Well, given how much they’ve been focusing on “shareholder value” over the last decade, I think it’s fair to call this process Shareholder Assisted Suicide. Companies with suicidal thoughts should have the numbers of their favorite congressmen on speed dial. I see a bailout in Intel’s future or failing that a merger/takeover with/by AMD. Or maybe some Private Equity firms will take Intel private.
Riffing on this proposal, and inspired by the “shareholder value maximization” theory of the sole duty of corporate governance and management, perhaps one could call it “Shareholder-Oriented Suicide”. This term has the advantage that its acronym is an apposite double-entendre.
If some Private Equity firms took over Intel, it would be like giving it over to a financial chop shop.
I agree w/everything above….but does anyone know off hand…
from 2000 to 2020, Intel’s stock did not pass its dot-com 1.0 highs. Did they buyback stock or issue dividends during that period too? I don’t recall from the top of my head.
the big face-palm is that for decades Intel had access to cheap long-term debt for a potential turn-around that was squandered.
here is the answer… wow. https://ycharts.com/companies/INTC/stock_buyback
AMD had the “luxury” of having no excess cash. so it had to pivot to Ryzen at the same time and deliver product
Intel milked its x86 line just as Boeing is milking its 737, 787 line.
“From 2000 to 2020, Intel’s stock did not pass its dot-com 1.0 highs. Did they buyback stock or issue dividends during that period too?”
https://www.intc.com/stock-info/dividends-and-buybacks
June 24, 2024
Intel Corporation – Buybacks
We have an ongoing authorization (originally approved by our Board of Directors in 2005 and subsequently amended) to repurchase shares of our common stock in open market or negotiated transactions. As of June 29th, 2024, we were authorized to repurchase up to $110.0 billion, of which $7.24 billion remain available. We have repurchased 5.77 billion shares at a cost of $152.05 billion since the program began in 1990.
I’ll gladly dance on the grave of Intel. They’re near the top of the BDS list, so good to see them bleed.
I wonder how much they spent on preparing for the suspended chip factory in Israel.
According to this article,
It does not say how much of those 25 billion had already been spent. The article also says at least 150, but possibly up to 1500 employees will be laid off in the Zionist entity alone. Good riddance to them.
–‐-The level of corporate negligence at work for Boeing and Intel needs a name.
It’s the HBS-neutron bomb. (nothing person to HBS, Wharton, et al are just as bad)
It has been said that much of the Chips Act was to start the process of moving chip production out of Taiwan so as not to leave it to China when it eventually retakes the island and the US knows it can’t prevent that. Pure speculation on my part, but a merger between Intel and TSMC might be a good way to help speed that process along and provide a cover story for a major redirection of capital investments at TSMC and a reason for a covert bailout of Intel.
Doubtful, at least not that directly.
Intel, and its subdivisions are so large that TSMC or any other large contract manufacturer (STMicroelectronics, GlobalFoundries, etc) that could absorb the foundry business of Intel, or any chip manufacturer large enough to absorbe the chip business of Intel (AMD, NVIDIA, Samsung, etc), or any memory manufacturer large enough to take over the flash memory business of Intel (e.g. Micron, Hynix, etc), would immediately attract the scrutiny of the competition and monopoly authorities of the USA, EU, and China. It would take a lot of time to clear up the mess.
Subdividing the various business areas of Intel further to sell them off to various firms active in the same area (perhaps via auctions) would also take a lot of time.
Hence, I suspect that, should Intel collapse, a lot of its industrial capacity and know-how will simply end up as scrap.
There already is a term. In the Sopranos, Tony took over his high school friends’ sporting goods store to pay off a gambling debt. He then used it as a short-term bank, leveraging the store’s equity to pay themselves huge amounts of money, disregarding the actual function of the store (selling sports equipment) until finally it’s dead and the building is for sale. It’s called a “Bust-out”.
No way, that’s the private equity model. Way too short term for this example.
This is a slower, longer-term bleed over a decade or two.
Slow bust-out? Stealth bust-out? Slow-bust bleed-out?
Some word which would capture the Sopranoform culture, ethics and motivation of the stock buybacking management.
Intel’s downfall is not only due to financialization, but also due to hubris and arrogance of their engineering leadership. Their problems started around 2012 with repeated failures to successfully produce 10 nm chips, originally slated to launch in 2015, which did not fully launch until 2020 (some laptop chips launched in 2018, but in limited volumes).
Main issue with the process was that Intel insisted on producing large monolithic chips, where everything, CPU, memory cache, I/O is produced by the same process. Error rate in chip manufacturing is proportional to surface area, each silicon wafer will have a certain number of defects. If you are making only 20 large chips out of a wafer and there is 20 defects per wafer, you may only get 10 good chips (50% yield). If you are making 100 smaller chips, you’ll get 85 or so good chips (85% yield).
AMD, Apple and mobile chip manufacturers all switched to a chiplet design, where only the CPU cores (maybe 30% of the chip area) are produced with the latest process in a 2- or 4-core chiplet, which is then connected to the main die containing memory cache and i/O produced by an older process (typically 12- or 14 nm). Chips produced this way are not a fast as monolithic chips, but the yields are much, much higher, and the cost is kept in check by producing only the CPU chiplet with the latest process.
Intel refused to adopt the chiplet design for the longest time because they wanted to have the highest performing chips no matter what. They thought they are better than everyone and took pride in the fact that theirs is the only monolithic chip on the market. Didn’t work out well for them.
It’s a trade-off. Pads (which connect to the outside world) are placed on the outside edge of the chip and are big because you need to connect them to the silicon. Thus the number of pads is determined by the circumference of your die. The more pads, the more data you can ship in and out in parallel. The fewer, the more data you have to send serially, which adds complexity (caching/serializing data) and latency. The alternative is to make a monolithic design which is simpler, but suffers worse yields. AFAIK, the largest chips currently shipping are 46200mm^2 composed of 7×12 dies of 550mm^2. Intel’s chips are around 150 mm^2 and they also use 12 inch wafers. It would be surprising if their yield were only 50% if cerebras is shipping (which they are).
AMD pioneered the recent chiplet solution, now adopted by Apple and Nvidia, but before that P6 was essentially a chiplet solution. As I said, it’s a matter of tradeoffs.
For sure, it is an absolute pile of problems they have to solve along with the financialization. Anecdotally, I’ve only known one person who worked for Intel, and he referred to working at Intel as an “old fashioned corporate hellhole”. I read that they will be cutting the free fruit and beverages to their employees now (I mean, they can’t afford free drinks for their employees??), so ‘corporate hellhole’ sounds about right. Intel also really, really loves the H1B program, a big user (abuser?). Always angling for more cut-rate foreign labor.
Perhaps you recall when Intel hired on Jim Keller a few years back? Keller is the closest thing to chip engineer superstar, he’s a genius lead architect for chip designs at AMD, DEC, Apple etc. and sort of works like a ‘hired gun’ for these companies. Intel signed him up and he barely lasted two years there before he quit for “personal reasons” , and it was quite a shock at the time for industry watchers. The rumours were that Intel’s organization was just that dysfunctional. Before Keller, Intel lost their long-time chief engineer François Piednoël for possibly the same reasons. He blamed Intel’s crappy Skylake QA for losing Apple’s business in the Macbooks.
Intel’s hubris goes a lot further back. In 2001 Intel shipped Itanium after ten years of development work, which was anointed as the new 64-bit architecture for all future Intel processors. Just one major problem: It was NOT backwards-compatible with the IA-32 processors. Compare this to the previous progression: Intel started with the 8080. The 8086/8088 was not strictly backwards-compatible to the 8080, but code could be easily ported to the 8086 by an instruction translator. The 80286 was backwards-compatible with the 8086. The 32-bit 80386 was backwards-compatible with the 80286. This continued with the 80486 and various flavors of “Pentium” branded processors.
Itanium had a completely new and complex RISC-like instruction set. At the time it shipped, the available C compilers were buggy and produced seriously slow code. It took another five years to ship a workable compiler, and only a masochist would try to write in assembly language on this instruction set.
Itanium was an utter flop. Unclear how much it cost Intel….easily tens of billions, and could even be as high as $100 Billion. It also saved AMD by taking Intel out of the game for the better part of 15 years.
AMD quietly went off and developed the AMD-64 instruction set for their line of (backwards-compatible!) 64-bit processors. Intel then developed their own version of a 64-bit processor (called IA-64) that was backwards-compatible with the older 32-bit processors, but it was awkward and significantly less capable than AMD’s version. The market spoke loudly in favor of AMD’s version, and the ultimate humiliation for Intel: They ended up using their old cross-licensing deal (that gave AMD the rights to the IA-32 instruction set) to license the AMD-64 instruction set and that has been the basis for all 64-bit Intel processors ever since.
In my opinion, the arrogance that is bringing Intel down today started 35 years ago. This is why I’m not wildly optimistic about Gelsinger’s ability to turn the company around: His original time at Intel was in the middle of this period of maximum arrogance. If he’s bringing that old arrogance back to Intel, it won’t go well this time either.
Not only is ARM taking share from Intel – but there is a spectre haunting both ARM and Intel. The spectre of RISC-V:
https://www.semianalysis.com/p/ventana-risc-v-cpus-beating-next
“Last week we attended the RISC-V Summit in San Jose. Traction in the RISC-V ecosystem is accelerating faster and faster. The embedded world has shipped 10’s billions of RISC-V cores collectively over the last few years. Qualcomm has silently shifted control and security cores that are not exposed to the user since the 2020 S865 chip. This has enabled them to ship a total of 650 million RISC-V cores! Andes and Codasip are each at over 2 billion RISC-V cores shipped, Western Digital ships over a billion RISC-V cores a year, and even Apple is converting some non-user-facing functions to RISC-V! Google also announced that Android works on RISC-V, and tier 1 vendors will have RISC-V-based phones sooner than most expect.”
Intel is f-ed.
These Companies are victims of a MES (Management Enrichment Syndrome).
Worked at Intel 12 years. Management heavy company. Noted that majority of managers were people who were the least technically competent and thus resulted in very poor decision making. Job was outsourced 4x while there, first to India, second to Poland, third to Korea, fourth to Taiwan. Greed run rampant. First the US exported it’s manufacturing, then it’s engineering. What do we have left? As long as a company is profitable it’s priority should be to take care of it’s employees. The stock market is only concerned with increasing profits year over year which is never sustainable and should not be the basis for any economy.
I would propose, given that similarity across many segments of the US, history will name it “The American Disease” in a hat-tip to the “Dutch Disease” or the curse of too large a natural resource/commodity in an economy.
For example, the Intel disease is similar to the US Military, where cash has not been funneled into functional weapons (investment) and capacity but into the connected’s pockets.
The American Disease: Short term-ism and looting the public wealth.
Also would suggest a (minor) change to the masthead:
Fearless commentary on finance, economics, politics, health, and power
This is just my personal opinion. Intel works great in my Linux PCs. Compared to Nvidia whose drivers don’t work. Worse there are two of them, the free version and the proprietary version. The Nvidia H-100 chip which is the hot chip for AI costs between US$20,000 to US$40,000 A clever person will figure out how to use a less expensive AI chipset and then what for Nvidia? Sure Intel has problems, but I’m rooting for them
General observation: standalone graphic cards (whether NVIDIA/GeForce, AMD/Radeon, or whatever) are with a surprising frequency the source of very annoying problems: buggy or incompatible drivers, risk of overheating, reliability of the graphics processor — whichever operating system you are using (Windows, linux, MacOS).
If you do not absolutely need the acceleration for graphics-intensive jobs (e.g. video editing, gaming) or for workloads that benefit from a graphics processor (e.g. bitcoin mining…) then forget about those Geforce/Radeon/etc chips/cards and go with an integrated graphics processor. Intel used to have them and they were good enough, AMD used not to have them and required you to have an additional graphics card.
AMD has offered “APUs” (Accelerated Processing Units, as described in https://en.wikipedia.org/wiki/AMD_APU) that combine CPU and GPU functionality since 2011. I’m currently running an AMD 5700G, which has a Zen 3 core with embedded AMD Vega graphics.
And like Fred, I’m running Linux, where Intel and AMD graphics have long had good support. [NVidia? Not so much.]
What you meant to say is “Linux works great on my Intel PC”.
This says nothing about Intel’s strengths or competence, they had nothing to do with Linux, they were just the default platform everyone needed to work well on.
Intel is the number one contributor to the Linux kernel, Nvidia is about 20th. A lot of PCs come with Nvidia as standard, not just high end video cards. I did mis-state that Nvidia cards don’t work. It’s more that users can’t manage them, especially during updates or when they have video problems
Very true Fred, Intel in fact did a lot for the open source community and were very early supporters. The Open Source Technology Center did a lot of good work. Dirk Hohndel ran it for Intel for a number of years.
I’ve found the smoothest experiences running Linux is on old Intel based Thinkpads, the things run basically perfectly on whatever flavor of Linux you prefer. Soooooo easy. And no bloody Broadcom WiFi drivers to deal with :)
Think you should use anything but NViDIA, since they are hostile to Linux.
Possibly part of the trend, another example of a company that might be heading the wrong way, Sonos:
https://www.forbes.com/sites/jonathanreichental/2024/05/09/4-lessons-from-sonos-app-upgrade-fiasco/
&
https://www.express.co.uk/life-style/science-technology/1929642/Sonos-app-issues-apology-fix-speakers
Less serious failure than Boeing, less costly than Intel but might be part of a similar trend.
The Sonos app update has been nothing short of a fiasco. Not sure what they were thinking, but functionality has been borked.
Enshittification underway?
As Lambert might say, it’s a damn shame.
I was an observer at several meetings of an American public technology corporation. The board members took pride in not actually understanding anything about the technology of the business. The only technologists allowed to speak were those who twisted the story into support for the financial and sales view of the rest of the board. What they were really doing was relying on the massive efforts of their very good technology teams to accommodate bad decisions, but that has its limits. People operate a business by the levers they understand. In this case, after those levers stopped working, the corporation was acquired.
I have always thought in my naive way that capitalism works by having a firm generate profits that are then rolled back into the firm, creating a virtuous cycle and providing investment capital for the firm that does not need to be borrowed. Firms that generate more profits have more money to invest in themselves, giving them a tremendous advantage and allowing them to do more speculative innovation and take more risks in their business area, increasing the health and vigor of the overall economy.
Obviously, using the firm’s profits to buy back “used” stock on the open market changes this whole dynamic. It allows existing stockholders to loot the firm’s wealth, while depriving the firm itself of benefits from its innovation and work. Businesses that would have been benefited from additional self-investment instead become stagnant.
It’s easy to see why management teams prefer stock buybacks. It’s easy, not requiring any difficult mastery of the firms business area, or even hard work, and there are kudos aplenty to be had for managers who enrich themselves and their friends with the company’s profits instead of reinvesting it on something that won’t bear fruit for years. For management, one or two years of big stock buybacks may set them up for life, regardless of what happens to them or their firm subsequently.
Seems like an easy remedy for much of this would be to reverse the legislation making stock buybacks legal. I don’t know if this is being discussed, but it sure seems like a no-brainer.
“design good chips”?! Intel?
My understanding is that Intel got rich the same way Microsoft did; fraud. Intel chips were as good hardware as Microsoft was software. IBM first approached Motorola, which really did make good chips. Motorola said they would need more time to ramp up production. Intel said no problem, and then took as long as Motorola would have.
the Intel 8000-series was as mediocre as the 6800 series was excellent and trapped the pc industry with mediocre technology for decades, as Intel used it’s market power to crush technologically better products.
AMD was superior and cheaper in every category for several years, including server chips, and still got crushed.
Intel created “new” chips by increasing the clock speed on their existing chips and selling them as new for about a decade.
Intel did get very adept at designing chips when they absorbed the microprocessor division from the legendary Digital Equipment Corporation (designer of the equally legendary Alpha line of processors, and before that the VAXen minicomputers). This was after DEC was sold to nouveau riche “IBM compatible PC” maker Compaq, which didn’t really know what to do with DEC’s advanced technologies and platforms. It just seemed like a good thing to use their newfound wealth on.
Compaq split up DEC after purchasing it and sold off parts separately and I can’t remember who got exactly what, but Intel definitely did get the CPU design and semiconductor fabrication parts. (Compaq itself was later bought by HP, which began the demise of what was another great company, but that’s a different story.)
DEC were at the time they were bought up recognized as the undisputed leader in chip design (and revered for many other of its accomplishments, like the VMS operating system, which by the way became the ancestor of Windows NT, and from there all modern iterations of Windows). Throughout the 1990s, an Alpha processor was almost always the fastest processor on the market. (This was when there were competing processor architectures, as opposed to just x86 with two makers, Intel and AMD, and then recently some more variety again with the ARM architecture escaping the mobile sector into new markets.)
Ah, memories. The first computer I used in the lab was a VAX minicomputer, when I was (very) young. Seemed like everything was DEC at the time. One VAX was programmed to run a microscope and another to run a spectrophotometer. They were called “online” instruments and a senior colleague built a small company out of the idea that still exists. A bit clunky on the front end (octal anyone?) but astonishing results were to be had. The combination of a near-perfect optical bench and optics with digitization was unbeatable. Now the optical bench is lousy in spectrophotometers, with digital processing making up for the poor physics. Microscopy has been less affected by scrimping on the physics while digital analysis has led to astonishing results. The Compucolor was also a thing of beauty!
” as the 6800 series was excellent”
I suspect you meant the 68000 and other 680xx variants.
The company I worked for at the time (circa 1980) was a big fan of Motorola devices
I remember one engineer hazarding that if Motorola had gotten the 68000 out a year earlier, that Intel would have not happened.
If I remember correctly, the first wave of Unix workstations largely relied upon the Motorola 68XXX series — before SPARC (from Sun Microsystems), MIPS (from SGI), and other risc-based processors took over.
There also was National Semiconductor in that space — and Texas Instruments later in the mobile phone market.
Taking into account the processors designed and manufactured by the likes of IBM, DEC, HP, there seems to have been a greater diversity 30-40 years ago than nowadays. And as we know, ecosystems exhibiting a greater diversity are supposedly more resilient…
68xxx running UNIX on the HP series 300 was the case.
A lot of good stuff ran on a humble 68000 with a massive 16 megabyte of addressing range.
Motorola did a 68332 for the auto industry for engine control.
I like SocialJimObjects suggestion of Shareholder Assisted Suicide.
The SOS variation is clever, but doesn’t have the Physician Assisted Suicide association and bite.
Unfortunately I suspected that Boeing was just the canary in the coal mine with regard to “maximizing profit by burning down the real wealth of the corporation”. (The real wealth of a corporation like Boeing is the knowledge and the means to manufacture a physical product.) Intel has obviously done the same thing as Boeing, and I suspect we will find that there are many other major corporate giants in America that are in similar shape. This is not an error, a mistake, or an anomaly, this is, by the regulatory laws and economic philosophy, the proper way to manage these corporations. To fundamentally change this behavior, those laws, and philosophy must be changed. I don’t see that happening so I expect more of the real industrial wealth of America to get flushed.
I think we will find (as we have been finding in Ukraine) that there are real consequences for the wrecking of America’s industry.
nice –
Boeing maximized profit by burning down the real wealth of the corporation….the knowledge and the means to manufacture a physical product.
I used to work for one of the mentioned companies, and was doing some work with Intel too, when the first mass layoffs hit…
After some decades in tech, I started using the shorthand “Viking Funeral”. The corporate body is mostly dead already, so the corporate body gets loaded onto a financial barge along with every combustible item the company owns, then is shoved out to sea and ignited from a distance by the executive team, of course, with flaming arrows, because the show has to distract everyone from what’s being done. It is a brilliant and spectacular sight to behold, but eventually the funeral party disbands and the survivors all go home, and the fire is extinguished with a final burp from the sea.
One visible sign that a Viking Funeral is in the making is when the CFO or (worse) HR….sorry, what am I saying…People Officer….take over employee all hands and quarterly investor calls in part or completely from the CEO. There are many signs along the way, and there is incentive to stretch the funeral preparations out as long as humanly possible.
This topic is relevant to NC because corporate finance often plays the role of undertaker….
Might also be called Assisted Financial Suicide….but it is actually Murder on the Orient Express (all of leadership did it)
Intel suffers from the same disease as Boeing – hardly anybody on the Board understands the technical side of Intel’s business. Only CEO, Pat Gelsinger, went through wafer fab plants. Remaining 11 board members are bean counters or professors. Boeing? Few months ago when I checked, no one worked in plane assembly or design, no one was piloting a plane. How can such boards decide on investments in technologies and products when they do not understand core business of the company?
Intel had two bad CEOs that focused on financial engineering. First BK was way too focused on buying back stock and shiny things that didn’t help manufacturing transistors. Partially due to missing on mobile Intel went into drones, wearables, etc. and those all failed. He pursued things like a TV show, sponsoring the Olympics so he could run in the torch relay, etc. After he was fired due to a relationship with a subordinate Bob Swan was put into place. His background was in marketing and not engineering so he didn’t track the manufacturing issues. He also had a lot of compensation tied to getting Intel’s stock price to a certain price so he bought back stock, got his big payoff, and left. While Pat Gelsinger deserves some criticism, don’t blame him for the buybacks or losing the process technology lead.
Failiing in 2012 to deliver 10nm was a huge problem. Fab life is tough so a lot of people working in the fab only last one process node. When the delays happened lots of people left before 10nm was done and the knowledge was lost. There also wasn’t enough accountability and the CEOs just let the manufacturing side repeatedly fail to deliver. Intel’s products in 10nm would have looked a lot better if the process delivered on time.
Intel also has a problem with compensation. While its pay is high compared to average salary, it is far lower than the top tech companies. It has lost lots of talent and for the first time I wonder if it is actually going to recover.
Agree generally but one quibble would be that it was not the manufacturing side that failed to deliver, rather it was development failing to deliver a manufacturable process. Since 2014 manufacturing has been hell to paraphrase a certain zillionaire. Prior to that it was just common or garden horrible. The decisions taken in 2010 timeframe are what led to a lot of the current difficulties.
Wow, I’m happy I was able to contribute a teensy bit to your great site, Yves. Sorry about the swearing! lol.
More (bad) news for Intel today on the other obvious Intel/Boeing parallel, that being selling defective-by-design products. The Register: Another law firm piles on Intel for Raptor Lake CPU failures as complaints grow louder . Anyone that owns an Intel 13th or 14th gen processor should keep an eye out for developments on this front, as Intel will be releasing a software update mid-August that will hopefully stop its’ CPUs from frying themselves. Even then it might be too late depending on how much physical damage has already occurred on affected CPUs. Among tech enthusiasts this is battering Intel’s reputation big time, needless to say.
It remains to be seen if Pat Gelsinger’s leadership will be able to turn things around. He is an engineer with deep experience at Intel, that’s a definite plus I think, but Intel hasn’t been able to execute in years, the management problems run that deep. He’s been running Intel now for three years, and the design timelines on x86 chips are 3-5 years typically, so end of next year should really show if Gelsinger’s changes have paid off. He has gone ‘all-in’ on the manufacturing plan ( “I’ve bet the whole company on 18a” ).
Sort of goes back to their roots, deep in manufacturing. The late Andy Grove, one of Intel’s co-founders, would be happy at least. His 2010 article on the importance of high tech manufacturing jobs in America is still relevant today: Andy Grove: How America Can Create Jobs
Mikerw0
August 5, 2024 at 7:20 am
They are another zombie company — why can’t Wall Street analysts read, analyze and understand financial statements.
—————-
From “information-theoretic” perspective, there is a deadly paradox in the combination of (1) maximizing shareholder value, enforced by corporate riders (2) the strange process of producing share prices.
How is investment advise produced? From my limited understanding, there are “fundamentalist” who pay attention to financial statements and many other news indication competence of the management, technology trends etc, and “technicians” who analyze trading behavior of other participants in the stock market. Fundamentalist may correctly see long term dangers to the company value, but stock buybacks send positive signal to technicians, and their investment advise is superior for years until the company collapses. It is a vicious circle: if the market follows stupid advise, stupid advise is good, fundamental aspects of companies be damned.
Intelligent tax system would decrease the “efficacy” of technical advise, but the political system is highly resistant to advice that is not combined with fat political donations and other perks to the beneficiaries of revolving doors and other forms of legalized bribery.
I don’t know about Intel but I have a relative that works for Boeing. They think that in the past they were forced to do the DEI hiring thing. Most of the younger ones either can’t do the job or figured out they don’t have to. So management did their job and their subordinates job too. Less hassle. It snow balled from there. Supposedly there was some grade given to companies and if they not a low grade they were cut off from access to capital. Again controlled by BlackRock’s Fink. Anyone know anything about that? If true that is a financial attack on America in my opinion and it is way past time to start putting enemies of America in jail. And it ain’t gray haired grandmas during J6.