Trump’s was able to leverage attack on Intel CEO Lip-Bu Tan, based on the false idea that the Malaysian/Singaporean was a Chinese operative, into a controversial 9.9% stake by converting CHIPS grants into equity precisely because the once-dominant semiconductor company is in a world of hurt. As we’ll explain, this agreement does nothing to improve Intel’s wobbly fundamentals and at the margin, weakens it financially. This deal is best thought of as yet another dominance display by Trump, attempting to demonstrate his acumen by acting as a chiseler who retrades deals to squeeze more from weak parties.
But this arrangement is not socialism, as many decry. The bare facts of this agreement are not even being properly reported. The Associated Press stated that the US is getting non-voting shares. That is not correct. From Intel’s site:
Under terms of the agreement, the United States government will make an $8.9 billion investment in Intel common stock…
The government’s equity stake will be funded by the remaining $5.7 billion in grants previously awarded, but not yet paid, to Intel under the U.S. CHIPS and Science Act and $3.2 billion awarded to the company as part of the Secure Enclave program. Intel will continue to deliver on its Secure Enclave obligations and reaffirmed its commitment to delivering trusted and secure semiconductors to the U.S. Department of Defense. The $8.9 billion investment is in addition to the $2.2 billion in CHIPS grants Intel has received to date, making for a total investment of $11.1 billion.
Sports fans, this is a loss for Intel. It either had gotten or was due to get the cash.1 It has instead had to issue more common stock, which dilutes current shareholders and at the margin would make it more costly to raise new equity. And even though (as we’ll soon explain), Intel’s problems are operational, not financial, and can’t be solved by money, any more than throwing more funds at Project Ukraine can magic new weapons into existence.
And Intel was downgraded by Fitch on August 4 to from BBB+ to BBB with a negative outlook. S&P had already downgraded Intel to BBB in December. Moody’s still has Intel a notch higher, at BBB1, just downgraded from A3 on August 8. Regardless, BBB is the last major ratings grade above junk level. So it might behoove Intel to sell shares at some point before its rating condition becomes more perilous.
Much of the breathless coverage gives the impression that the Federal government got the sort of extra rights that a large shareholder by virtue of a one-time investment might extract, such as board seats or veto rights. Nope. Again from Intel:
The government’s investment in Intel will be a passive ownership, with no Board representation or other governance or information rights. The government also agrees to vote with the Company’s Board of Directors on matters requiring shareholder approval, with limited exceptions.
The government will receive a five-year warrant, at $20 per share for an additional five percent of Intel common shares, exercisable only if Intel ceases to own at least 51% of the foundry business.
Intel also stopped paying dividends over a year ago.
So the government got no more influence over Intel than it already possessed due to Trump’s very active use of bullying and name calling. Keep in mind that those alone are the reason Nvidia and AMD capituled to the demand that they agree to pay a probably unconstitutional 15% “export tax” on sales to China.2 The business press also attributes Trump demands that companies eat tariffs as a big reason why major retailers have passed through only a small portion of those new costs to customers.2
It’s not a secret that Intel’s woes are deep seated. Like Boeing, its decline is the result of Intel’s once great engineering being subordinated to rule by MBA. Bill Lozaonick, who has been a long critic of stock buybacks, has written extensively about Intel. Consider how far Intel’s decline had progressed as of his and Matt Hopkins’ 2021 article How Intel Financialized and Lost Leadership in Semiconductor Fabrication:
For a half-century after the invention of the integrated circuit at Texas Instruments (TI) and Fairchild Semiconductor in the late 1950s, the United States was a leader in global semiconductor fabrication….
Meanwhile, from the early 1980s, “fabless” semiconductor companies – firms that designed, but did not fabricate chips themselves – proliferated, creating products for varied segments of the memory and logic markets…
Responding to the opportunity to manufacture chips for fabless firms, in 1987 Morris Chang, a Taiwanese native with electrical engineering degrees from MIT and Stanford and 25 years of work experience at TI, launched Taiwan Semiconductor Manufacturing Corporation (TSMC) as the world’s first “pure play” foundry…
Larger, however…. as a semiconductor fabricator is Samsung Electronics Corporation (SEC), the flagship company of the Korean Samsung chaebol, with estimated 2020 foundry revenues of $14.7b..As an IDM [integrated device manufacturer], SEC was second globally to Intel, which had $77.9b. in 2020 semiconductor revenues. Among the fabless firms competing with Intel at the high end of the processor market are U.S.-based Nvidia (2020 revenues: $16.7b.) and AMD ($9.8b.)….
The most advanced chips are produced for smartphones as “system-on-a-chip” (SOC). The leading designers of SOCs are Apple (USA), Qualcomm (USA), SEC (South Korea), and MediaTek (Taiwan). HiSilicon, a wholly-owned subsidiary of Huawei Technologies (China), was also a SOC leader until U.S. trade sanctions, implemented from August 2020, terminated its access to TSMC’s fab output.[4] TSMC and SEC use 7nm and 5nm process technology to manufacture many of these mobile processors along with the most advanced AMD computer and Nvidia gaming SOCs. According to TSMC, compared with 7nm, 5nm offers 15% faster speed, 30% less power consumption, and 1.8 times the logic density…
Intel still leads the global semiconductor industry in total revenues. But, as an IDM, Intel manufactures almost all its CPUs at 14nm, and its 10nm capacity has been stuck, with limited output, since 2018. Meanwhile, Apple is abandoning Intel processors for its Mac computers, turning instead to TSMC to fabricate Apple’s own designs. Intel itself already contracts with TSMC and UMC to produce 15-20% of its non-CPU chips. Moreover, later this year, TSMC will commence production of intel’s Core i3 processors, inside advanced laptops, at 5nm.
So stop a second and note the importance of that last paragraph. The bulk of Intel’s output was of much older, workhorse 14nm chips; it wasn’t even much of a player at 10nm chips and was having to outsource the manufacture of its 5mm chip designs.
Lazonick and Hopkins soon observe:
Why has Intel fallen behind TSMC and SEC in semiconductor fabrication, and why is it unlikely to catch up? The problem is that Intel is engaged in two types of competition, one with companies like TSMC and SEC in cutting-edge fabrication technology and the other within Intel itself between innovation and financialization. The Asian companies have governance structures that vaccinate them from an economic virus known as “maximizing shareholder value” (MSV). Intel caught the virus over two decades ago.
The authors documented the extent of Intel’s payouts, both via dividends and buybacks, but pointed out that is it not money per se but financialization that was driving Intel’s decline:
The foundation of financial commitment is retained earnings. In the case of Intel, as shown in Table 1 above, in recent years the company has made substantial allocations to P&E and R&D, even as it has distributed almost all its profits to shareholders.[20] But Intel has been able to tap other cash flows to make, simultaneously, large-scale productive investments and shareholder payouts. For the decade, 2011-2020, these other cash flows included depreciation charges of $87b., long-term debt issues of $45b., and stock sales (mainly to employees in stock-based compensation plans) of $12b.
Given the availability of these sources of funds, the vast sums that Intel has wasted on buybacks have not thus far imposed a cash constraint on its investments in semiconductor fabrication. Rather, it has been a deficiency in organizational learning—the essence of the innovation process—that has hampered Intel’s implementation of process technology. The generation of high levels of productivity from P&E and R&D expenditures requires, as a second social condition of innovative enterprise, organizational integration, working in combination with financial commitment. Organizational integration mobilizes the skills and efforts of large numbers of people in a hierarchical and functional division of labor into the collective and cumulative learning processes required to transform technologies to generate a higher-quality product and, then, access markets to attain economies of scale.
The root of Intel’s failure in organizational integration lies in the financialized character of a third social condition of innovative enterprise, strategic control. Accepting stock yield as the measure of enterprise performance, in recent years Intel’s senior executives who exercise strategic control have lacked both the incentive and, increasingly we would argue, the ability, to implement innovative investment strategies through organizational integration.
Matt Stoller, in an important August post, brought the story of Intel’s fall up to date:
The answer is that Trump is a more authoritarian version of the same finance oriented leadership style we’ve had since Jimmy Carter…
If Trump were different than his predecessors, he’d try to save Intel, and be disdainful of Nvidia and Apple. So far, he isn’t. He cut deals with Nvidia CEO Jensen Huang, who he sees as a winner. Same with Apple CEO Tim Cook. He humiliated Intel’s CEO Lip-Bu Tan, precisely because Intel is doing the dirty physical work that Wall Street hates….
The best way to understand how Trump is running a more authoritarian version of the same playbook is to look at how he treats the three most important semiconductor firms in America. The first is artificial intelligence firm Nvidia, which makes a monopoly software platform called CUDA, and outsources the production of physical chips to run its software to Taiwan Semiconductor. Nvidia is the most valuable company in the world by stock price. The second is Apple, which is the biggest buyer of semiconductors in the world and an important designer of chips. It too offshores its production of chips and electronics, and it is the third most valuable company in the world by stock price. Finally there is Intel, which is the only major American fabricator of high-end semiconductors, producing physical product in the United States. Its stock price has gotten killed since 2020, and now the company itself is dying…
But in 2005, the company pulled a Boeing. The board hired its first non-technical CEO, Paul Otellini, who moved the company away from investing in continued technological leadership and focused the company on anti-competitive practices, like using rebates to stop customers from buying from smaller rival AMD….
The clearest strategic misstep was missing out on supplying the chip for the iPhone, which Apple released in 2007…
Apple soon decided that the iPhone required a new kind of logic chip, one that could balance performance with the efficiency needed for a small battery-run device. To make acquire such a chip, Apple bet on restructuring the entire chip industry, designing chips in-house but having them produced by a partner whose sole focus was running foundries, Taiwan Semiconductor. So-called fabless chipmaking at scale wasn’t industry standard; one expression among chipmakers had been ‘real men have fabs.’ Also, nothing on the scale of the iPhone had ever existed.
Apple and TSMC eventually became a Wintel-style partnership…
Intel had always been an engineering-driven firm that took big risks when industry dynamics shifted, but under Otellini, it financialized and ossified. And so it missed multiple key turns in the industry – mobile, process leadership, and later on, the turn to a different kind of chip with AI. Former Intel principal engineer Francois Pidnoel said that the company’s culture resulted in “no innovation, no aggressive road-maps, and no people driven because they are are discouraged because the MBA’s are the only ones rising.” The Chair of the Board today is Frank Yeary, a private equity guy, and the board even has a VP from Boeing on it. The mirage of Intel’s strength has long since dissipated, but the company’s board members have not been held accountable.
Stoller also argues that policymakers played a big role in Intel’s decline. The company was the target of anti-trust cases over its use of rebates, but Chicago-school free market loving judges ruled for Intel.
And he posits that the fabless model, which now posed a threat to Intel, is the direct result of fabless customers commanding monopoly rent. In other sectors, like memory chips and analogue chips, vertical integration rules.
Stoller contends that government financial support cannot solve Intel’s woes (even if the money was ideally directed) because it could never be enough:
But the CHIPS Act, while it did spur some domestic production, did not fix the fundamental problem in chip markets. Government money is very limited compared to the actual demand from big customers. For instance, Apple, the largest chip buyer in the world, “has spent $458 billion since 2018 on dividends and share buybacks, an amount equal to nearly ten CHIPS Acts.” And it has no incentive to see Intel succeed, because it benefits from TSMC’s fabrication monopoly.
The comments section of the Financial Times story on the Trump-Intel agreement was overwhelmingly skeptical of the notion that Intel could right its ship, much the less that the US share acquisition would be of any benefit. Some examples:
Confabulator
The company has been going in the wrong direction since at least 10 years ago, and even longer if we include the debacle of missing out on the smartphone market. With AI, that’s 2 major innovations they missed out on. And now they’re increasingly less competitive in their traditional market segments too. Plus, their manufacturing has fallen behind the likes of TSMC whom they used to lead by a generation.Eclair
The question is « Does Intel have it anymore or are they completely lost »? They seem to be destined to manufacture the commodity chips but nothing too complex. This investment is at best shaky it seems. It certainly confirms Intel is in trouble.Strix technica
Poach TSMC’s talent? It’s worth a try, but their knowledge of TSMC’s processes alone (unless they got lucky and the individuals helped solve for TSMC specific problems similar to those that node 18A has) probably wouldn’t help much except insofar as they’re employed by TSMC because they’re smart people who know what they’re doing.Earlier this month, three TSMC current/ex-employees were arrested on allegations of IP theft. I forget where I read this now, but it was said that this stolen information was of limited use to the recipient, though how anyone could know that (especially in so short a time) I don’t know.
Still, ex-TSMC employees — provided they’re not the alleged thieves, of course! — could well be of help to solving Intel’s node 18A problems.
TwistedElegance
It’s very sensible for government to own a stake in key industries. However Intel’s trouble comes from the fact its latest processors are terrible compared to AMD. AMD delivers more cores, more performance, and lower power requirements. 2 years ago I built a new PC based on Intel (I don’t really game on it.) 18 months later, I find out that there is a bug in the processor firmware that means its slowly destroying itself by requesting too much voltage from the motherboard! A bios update has supposedly fixed this, but who knows the damage done up to that point?I don’t think I’d buy Intel again.
So what do we make of this “deal”? At a minimum, it’s yet another example of Trump’s default of violating the Sun Tsu warning: “All tactics and no strategy is the noise before the defeat.”
It also demonstrates Trump’s bizarre fixation with trying to extract monetary wins when there’s nothing, or almost nothing, to be had. As one Financial Times reader put it:
Armagheddon
Pro: at least the taxpayer gets something for their money, 10% in a failing business is better than nothing.
Con: this is still a bad use of money. Intel is saying “we should stop, we can’t find customers for this product”. US Gov: “do it anyhow, we like the optics on this one”.Bonfire of taxpayer money.
And this scheme is yet another demonstration of Trump engaging in bad faith behavior, here of pressing Intel to give up stock to get previously Congressionally-approved grant funding. Members of the Trump Team seem either to defer to his extremely transactional posture or actually believe in it. They seem unable to grasp that many things that governments normally do yield payoffs that Wharton MBAs find hard to put in spreadsheets, such as social cohesiveness and trust in institutions, and soft power.
This episode reminds me of the bigger row between Trump and Zelensky over the “raw earths” deal. Ukraine had opened this door by suggesting a natural resources partnership to induce the US to continue to fund the war. As Fortune then explained:
US President Donald Trump said on Saturday he was trying to get money back for the billions of dollars sent to support Ukraine’s war against Russia….
Trump told delegates at the Conservative Political Action Conference (CPAC) near Washington: “I’m trying to get the money back, or secured.
“I want them to give us something for all of the money that we put up. We’re asking for rare earth and oil, anything we can get.
“We’re going to get our money back because it’s just not fair. And we will see, but I think we’re pretty close to a deal, and we better be close because that has been a horrible situation.”
As readers may recall, Zelensky dug in his heels and refused to sign many highly extractive US sets of terms, finally accepting something less unreasonable.4
The whole idea was quickly shown to be a big joke when, shortly after the signing, Russia secured territory that included the biggest lithium deposits in Ukraine….one of the prizes the US sought.
Admittedly, with Ukraine having gotten the US to wrangle hard over an empty bag, there is the risk we warned about at the time, that the mere execution of this pact created US interests in having Ukraine survive, holding as much territory as possible, as opposed to focusing on the bigger prizes of normalizing relations with Russia and cutting Project Ukraine losses.
This US stake in Intel analogously runs the risk of committing the US to the Mission Impossible of saving the ever more troubled Intel. But it will take a while for Intel’s unraveling to advance to a crisis point, and that could easily occur after Trump’s watch. So the question of whether Intel is too essential to fail or merely list badly will remain in play.
_____
1 One might try to argue that the conversion of the $2.2 billion CHIPS Act grant eliminated claw-back and profit-sharing provisions is a benefit to Intel. Perhaps. But profit sharing provisions pre-suppose profit, so having to pay that would not be a hardship.
The bigger arguable issue was that CHIPS Acts disbursements were held up by red tape. Recall that this bill had bi-partisan support and was an attempt at industrial policy, both to reverse the US’ loss of dominance in the semi-conductor business, and reduce supply chain/procurement vulnerability. However, as a new Wall Street Journal article makes crystal-clear, Intel’s CEO Tan ran to see Trump after Trump aggressively attacked his loyalty. This deal had nada to do with Intel wanting to get its hands on the outstanding CHIPS Act funds. Intel already had $22 roughly in cash and equivalents before the Trump and Softbank infusions.
Wikipedia gives further detail about the CHIPS Act:
The CHIPS and Science Act is a U.S. federal statute enacted by the 117th United States Congress and signed into law by President Joe Biden on August 9, 2022. The act authorizes roughly $280 billion in new funding to boost domestic research and manufacturing of semiconductors in the United States, for which it appropriates $52.7 billion. The act includes $39 billion in subsidies for chip manufacturing on U.S. soil along with 25% investment tax credits for costs of manufacturing equipment, and $13 billion for semiconductor research and workforce training, with the dual aim of strengthening American supply chain resilience and countering China.: 1 It also invests $174 billion in the overall ecosystem of public sector research in science and technology, advancing human spaceflight, quantum computing, materials science, biotechnology, experimental physics, research security, social and ethical considerations, workforce development and diversity, equity, and inclusion efforts at NASA, NSF, DOE, EDA, and NIST.
The act does not have an official short title as a whole but is divided into three divisions with their own short titles: Division A is the CHIPS Act of 2022 (where CHIPS stands for the former “Creating Helpful Incentives to Produce Semiconductors” for America Act); Division B is the Research and Development, Competition, and Innovation Act; and Division C is the Supreme Court Security Funding Act of 2022.
By March 2024, analysts estimated that the act incentivized between 25 and 50 separate potential projects, with total projected investments of $160–200 billion and 25,000–45,000 new jobs. However, these projects are faced with delays in receiving grants due to bureaucratic hurdles, shortages of skilled workers, and congressional funding deals that have limited or cut research provisions of the Act by tens of billions of dollars.
The clawback provisions sound as if they were mere eyewash. Again from Wikipedia:
In October 2022, Senators Elizabeth Warren and Tammy Baldwin and Representatives Sean Casten, Jamaal Bowman, Pramila Jayapal and Bill Foster sent a letter to Secretary Raimondo urging her to detail how the Commerce Department would enforce the law’s provisions preventing companies from using CHIPS Act money directly on stock buybacks (they noted the law does not prevent recipients from using the money to free up their own funds for stock buybacks), as well as whether the department would claw back misused funds and resolve conflicts of interest.
2 In the Intel case, there’s a similar question of the legality of Trump running roughshod over the CHIPS Act, which was intended to be generous to semiconductor players and boost their US operations and competitiveness generally, by the “grants become equity” extraction.
3 Some analysts also contend that tariff costs are turning out to be lower than originally estimated.
4 Details are sketchy. However from PBS:
The deal covers minerals, including rare earth elements, but also other valuable resources, including oil and natural gas, according to the text released by Ukraine’s government.
It does not include resources that are already a source of revenue for the Ukrainian state. In other words, any profits under the deal are dependent on the success of new investments. Ukrainian officials have also noted that it does not refer to any debt obligations for Kyiv, meaning profits from the fund will likely not go toward paying the U.S. back for its previous support.
Officials have also emphasized that the agreement ensures full ownership of the resources remains with Ukraine, and the state will determine what can be extracted and where.
It does not mention any explicit security guarantees to deter future Russian aggression that Ukraine has long insisted on.
The text of the deal lists 55 minerals but says more can be agreed to…
The agreement establishes a reconstruction investment fund, and both the U.S. and Ukraine will have an equal say in its management, according to Svyrydenko.
The fund will be supported by the U.S. government through the U.S. International Development Finance Corporation agency, which Ukraine hopes will attract investment and technology from American and European countries.
Ukraine is expected to contribute 50 percent of all future profits from government-owned natural resources into the fund. The United States will also contribute in the form of direct funds and equipment, including badly needed air defense systems and other military aid.
Contributions to the fund will be reinvested in projects related to mining, oil and gas as well as infrastructure.
I have to admit that it is hard reading of a once great organization being augured into the ground by nothing more than MBAs and financialization. Some top bosses make the big bucks but the entire organization starts to sputter and maybe even die. Funny how Trump bought into them. It was not that many years ago that Republicans would have called that “communism” or something. Now that this is all facts on the ground, will it serve as a blueprint for Boeing when it starts to go wobbly down the track?
I’m assuming Boeing is partnered with a short-line rail operator. Maybe someone running a 10 mile set of track. Train should be in very soon!
There’s a bit more to it than MBAs and MSV. In the case of Boeing and GE etc the engineering wasn’t quite so challenging. Making planes is complex and hard for sure but I’m not aware of any singular investment in their business quite like SOTA fabs. Getting to 7 and then 5 nm is spectacularly intensive in both capital and know-how. It looks like Intel has missed that boat, can’t catch up now and will be relegated. MBA/MSV-think played its part, I guess but given the cost of 5 nm fabs I wonder how many of them the global market can support.
Maybe Intel has some internal copies of a book that might provide some guidance.
It is titled “Only the Paranoid Survive” and was published in 1999.
There may be some corporate memory of the author, Andy Grove.
He was Intel CEO.
I do believe Andy was the last of intel engineer CEO. It was downhill from there.
you are forgetting Craig. https://en.wikipedia.org/wiki/Craig_Barrett_(chief_executive)
Good wrap up. Bum deal for Intel, indeed.
It might be worth adding that “Nvidia is the most valuable company in the world by stock price” largely due to ongoing high demand for their GPUs being used for crypto mining and now AI apps.
If the AI bubble really deflates, tech stonks will take a serious hit, and Nvidia could be hit the hardest. We are hearing now that there are worries about the AI bubble, though there is also so much money being thrown at this tech that it has a TBTF quality about it. Another threat, which this article discusses some, is of course the rise of China. There could be some unwelcome surprises coming, as the Chinese semiconductor industry grows.
In any case, I would expect destabilizing developments on those fronts before Intel’s problems become more acute.
Some of us have long thought of latter day Intel as being joined at the hip to Microsoft with new products from the former provoking new products from the latter and vice versa. Both companies presumably make most of their money from the business machines market and forced sale of average performance computers to consumers at places like Walmart.
So Intel’s lack of interest in smartphones–so very much an Apple thing–could parallel Microsoft’s lack of interest in the same. Meanwhile the problem of course is that the Chinese have undercut the PC market hugely although MS leverages IP law to still get their cut regardless.
Therefore to bad management add a changing computer scene where Intel’s high priced processors were less competitive given that smartphone “systems on a chip” now cost very little indeed. Intel rose on the back of a fast changing hardware market that has now matured.
I think “mm” is being used where “nm” is intended in paragraph beginning “So stop a second and note…”
Thanks, fixing.
Thanks for this article. I largely agree with the analysis.
One point, however: If T wants to use this as leverage to increase onshore US manufacturing there is a glaring problem, the same problem data center buildouts are facing.
The US electrical generating capacity and grid is creaking and becoming more and more expensive. If T really wants to re-onshore US manufacturing the electrical grid will need upgrading and expansion of both generating capacity – by fossil fuels or water/dams or renewable or all – and the transmission lines will need to be upgraded. Good luck.
I’ve heard it quipped that data centers with their enormous electric power requirements aren’t being built for AI; they are being built to bring in modern nuclear power plants for electricity generation without the 20-year permitting process. Maybe.
It’s interesting that the biggest driving force against nuclear power, starting in the 1970s, is the coal industry.. Old King Coal was behind a lot of anti-nuke actions, including heavy Congressional lobbying. Make of that what you will. / ;)
Again. In the end, it is always a failure of management. Whether at the CEO level, or the Presidential level.
As an aside, they are running out of light spectrum.
This is going to push all new GPU and CPU off shore. Intel, AMD CPU designs are at least thirty years of date. The only changes in the last thirty years has been the optimizing of the CPUs to run crap code. Nvidia is around twenty years out of date. China is currently designing an AI specific processing unit. The RISC V open design will most likely come out in China first. I like a lot of the RISC V design. It’s flexible and addresses a lot of current CPU (x86) design crap the everyone hates. I’m surprised Trump’s advisors have had him try to kill it. Try designing state of the art military hardware use stuff that’s 5-10 years out of date. You would not believe the paper work required to something not on the approved parts list. That’s the only thing that can make the design work. Kinda explains why another F-35 fell out of sky a couple weeks ago. By the way I have designed CPUs and GPUs. Things must be really bad in the defense industry. I had the strangest email from Locheed Martin in my whole career (other a recruiter from Google yelling at me because I had no desire to work for Google). The email basically said, Locheed Martin has a things you could work on. If you’re interested reply to this email and someone will call you immediately. No resume or application needed. In summery there is nothing that the government can do with their current mentality to stop the tech decline of US.
There’s already RISC V architecture chips available, such as on the Raspberry Pi RP2350 board for example.
RISC-V is intended to be an open architecture design mainly to replace/work-around more closed architecture designs that use load-store conventions on the chip, like ARM or AVR. ARM designs already accomplish a lot and have certain performance advantages over x86 based architectures that motivated Apple to base their Apple Silicon chips on ARM cores, as well as Amazon, Microsoft, and Google to design their own ARM core based server products for their cloud services.
The drawback of ARM is that because it’s closed source, manufacturers have to pay big $$ up-front to license a design, as well as royalties for each chip using an ARM-core. ARM can also always decide to “take their ball home” and stop licensing designs to you. This is the main reason China has been investing heavily in RISC-V designs, because it’s an open architecture, they can’t get locked out or be accused of violating treaties/piracy if they continue to manufacture them if RISC-V International becomes defunct or privatizes itself.
Speaking of mobile and chips, Apple released its C1 modem chip earlier this year. That chip has its origins in Apple’s 2019 acquisition of Intel’s modem business. My gut instinct tells me fixing Intel’s chip manufacturing will take longer than that.
Any reason why shouldn’t the government get board seats? Maybe a seat for someone appointed by the National Academies, another for Department of Commerce (NIST, NOAA), and another for US Trade Representative to act as a direct line to the President. DoD has obvious interests here, but that may turn off potential foreign customers.
Interesting — I remember reading this some years ago:
Apple’s failure to develop its own modem detailed in new report
Turned out very differently than the Apple Silicon chips, which seemed to turn out swimmingly.
I always thought it was going to take four to five years for Apple to get a final modem product to market. In my experience working in this field, the typical System-On-a-Chip product cycle is about that long, with Apple’s acquisition being a soft reset of development. The failure is business analysts and management not understanding, to quote the Verge article, “So yeah, modems are hard”.
Apple is also rumored to finally come out with its own WiFi/Bluetooth chips, cutting out Broadcom (just like they’re cutting out Qualcomm).
Apple’s silicon success turned out swimmingly because most of the design work was done by the engineering team from PASemi, a company founded by legendary chip designer Daniel Dobberpuhl and which employed famous chip design superstar like Jim Keller and a whole bunch of former D.E.C. engineer alumni. PASemi had actually designed a better POWER cpu than IBM had managed to do and was pitching it to Apple before Apple decided to throw in the towel entirely on the POWER architecture and switch to Intel’s x86. Apple was so impressed with PASemi that they ended up acquiring the whole company and getting them to work on the A4 chip, which was Apple’s first completely in-house designed cpu, and which set Apple on the path to surpassing Intel entirely and dumping them for their own designs.
State ownership of productive assets is a hallmark of Communism. There will be pliant board members with plausible deniability.
Yves has said this 100 times if she’s said it twice… Trump has only accelerated America’s decline.
Trump and his administration keep saying they want to MAGA, but it’s really only about the money.
Nvidia & Apple can be the administration’s darlings while they ride the market bubble and their monopoly revenues, but Intel is kicked to the curb because it’s run out of road at the end of it’s financialization instead of engineering.
He’s like some master of ceremony for the sinking Titanic or a crazy partying rabble rouser from the Great Gatsby. As long as the money and power keep flowing who cares what the water level is at. (Someone’s got a life boat waiting for him right ;o )
Yve’s point on Intel’s problems being largely “operational, not financial” and that it can’t be solved by money is spot on. Intel was the nearest equivalent to a money printing press as you can find in it’s heyday, and even with their ridiculous stock buyback figures over the past ten years, they still outspent AMD + TSMC + NVIDIA in R&D spending:
Intel R&D spending
AMD R&D spending
TSMC R&D spending
Nvidia R&D spending
Intel just can’t execute, they don’t have the leadership or the company culture to do it. Everything they put their hands on seems to do worse than their competitors. Nobody in the tech space trusts Intel’s roadmap announcements worth a damn (Intel Foundry and ‘delay’ are synonymous) and hence they struggle to find any customers at all for their IDM 2.0 foundry push. And they absolutely NEED to bag a few big volume customers so they can help cover the costs of staying on top in the chip fab game. We are now in an era where a bleeding edge chip fabrication plant is pushing $20 billion :
When your factories can cost $20 billion, even $30 billion dollars to build just one, you need enormous scale to make sense financially. Chip fabrication is all about scale and staying on top technologically; this is partly why TSMC’s business model makes more sense, they can amortize the costs of such massive investments among many fabless chip design customers. Intel’s vertical integration worked amazingly well for them in the past because of their sheer scale (The Register’s famous moniker “Chipzilla” comes to mind), but they are approaching the point where they can’t shoulder that enormous burden on their own anymore, and certainly not if their products are consistently crappier than their competitors. Crappy chip design+behind the times fabs = big trouble ahead.
When even the CHIPS act creators are doubtful, you know you’re in trouble:
[Oct 2024] Intel CEO is “frustrated” with CHIPS Act payout progress — Intel has received $0 from the $8.5 billion that the US government promised
[Sept. 2024] US government to delay Intel CHIPS act funding until company hits milestones
Remember when Digital Equipment was in the Nifty Fifty.
Asianometry has some thoughts on what should be done to get Intel back on it’s feet:
Intel Should Second-Source Nvidia https://www.youtube.com/watch?v=5oOk_KXbw6c
I think it’s good that the elites on both ends of the uniparty have decided that it may be important to have a “functioning” Intel that actually makes state-of-the-art chips, but it’s pretty obvious that they see this as an Intel problem rather than a problem with the mega corporate incentive structure coded into our laws, regulatory agencies, Wall St imperatives, and neoliberal financial capitalism. Because it’s obviously NOT just an Intel problem, the exact same problems exists across the corporate sector in America, and without some serious changes to that incentive structure, we’ll just watch the Federal government (whether it’s Obama, Trump, or Biden) just figure out how to throw bucks at the problem and then brag about how much money they dedicated to kicking the can down the road just a bit more.
And our elites really wouldn’t even do that much, but they have realized that the rest of the world (i.e. BRICS) is not going to kowtow to them, but instead is going full speed ahead to go past them in terms of science, technology, and progress.
I cannot understand why Intel shareholders aren’t suing Intel’s board for agreeing to Trump’s obviously illegal blackmail scheme.
There’s a memorable sequence in the TV Series The Sopranos where Tony basically goes around and discusses a potential divorce scenario with multiple family law practices around New Jersey in order to prevent his wife from being able to use those lawyers because it would create a dual-client conflict of interest.
Trump managed at the start of his term to get numerous law firms to sign concessions offering large retainers/hours of legal work available to him/the government as a settlement for investigations into Equal Opportunity violations at those firms. Regardless of the merits of such investigations, the motivations behind them were more retaliatory in nature, focusing on firms that were either responsible for helping in lawsuits against Trump, or other Conservative/Republican individuals and organizations. Those firms are now basically unavailable for anyone who wanted to sue the Trump administration/US government. While not all firms settled and some fought back, I wouldn’t be surprised if other law firms are trying to avoid getting noticed by Trump and drawing his ire.
“This US stake in Intel analogously runs the risk of committing the US to the Mission Impossible of saving the ever more troubled Intel.”
At one point my eyes were blurry and I read that as:
“This US stake in Israel analogously runs the risk of committing the US to the Mission Impossible of saving the ever more troubled Israel.”
Hi Yves,
We all love you but
is actually not a Sun Tsu quote… according to this.
I am a native Chinese speaker and I did a little digging, it appears that this is a back translation in Chinese (from where I have no idea). To be fair, the original paragraph is near impossible (same goes for many text from The Chun-qiu Era) to translate and I do agree with the meaning of this translation. Lionel Giles put it in something like this.
Probably too much information here, great piece by the way ;-)
So I will have to say “Widely attributed to Sun Tsu” the same way “In theory, there is no difference between theory and practice. In practice, there is” is to Yogi Berra.
Lol. “Widely and mistakenly attributed to Sun Tsu”. Or try this: his action again goes against the advice from Sun Tsu, albeit a very creative translation.
US government direct ‘investment’ in any tech product is going to raise suspicions in the non-western countries around the world, that the product will be used for back-door surveillance – just as the US accuses China of doing with some of its tech products.
The result will be to accelerate the non-use of Intel products in favour of home-grown or alternatively sourced equivalents. China seems to be already well advanced in this area.
Nah. Countries around the world (should) already think that every US tech product is used for back-door surveillance by default. You can’t raise suspicions that are already raised. :)