Apple is America's semiconductor problem
semiconductor-digest.com
semiconductor-digest.com
Apple was not the first or the only company to look to East Asia for cheap manufacturing, elctronic equipments, chips, or cheap labor. The trend was started and accelerated by motor companies and hardware just followed suit.
Then, some other points are just...there. Apple is America's semiconductor problem because Apple is so big it purchases things in bulk and gets discounts - like every other big company in every industry.
TSMC sold 100% of the capacity to Apple for 3nm chips because no one else had the designs ready, Apple needed them in bulk, and the yield wasn't that high. It is problematic, but you need to mention these things when making a claim.
The problem is two fold - other companies are not really innovating and experimenting at semis with a scale at which Apple is. This is a market failure more than an Apple failure. Others were reliant on Qualcomm, broadcomm, samsung, and Intel, with slower pace of innovation. Then there is Asian countries with cheap economies of scale.
It's one thing to assume the most nefarious intent from the start and then look at everything from that lens. It mostly leads to paranoia, not insight.
Exactly, and if Matt Stoller was honest he would bring up that point.
Most automotive chips use 28nm, 40nm, and 60nm nodes and China has been very competitive in this space now because of existing capacity built from OSAT and AMTP consolidation in China by Taiwanese and Malaysian/Singaporean players like UMC before the 2019 HK protests, MediaTek, and ASMPT from the 2000s-2010s.
It's analog chips, 28nm/40nm nodes and Packaging that have been getting a significant amount of CHIPS funding, as well as similar funding from the Japanese, Taiwanese, South Korean, Malaysian, and Indian governments recently.
From a NatSec perspective, it's those kinds of nodes that have a critical impact as most weapons systems can subsist on a Intel i7 and Nvidia Maxwell comparable CPU and GPU.
Think EWS, C-RAMs, Guidance Systems, etc.
> Isn't a lot of CHIPs Act money help prop up Intel who didn't succeed in mobile and is laying off thousands anyway
Those layoffs are staff who don't have experience with High NA EUV processes (aka staff who didn't skill up). I can safely say that.
The CHIPS funding is a mix of subsidizing High NA EUV nodes as well as domestic OSAT and AMTP/Packaging capacity.
1) Auto manufacturers incorrectly predicted that the pandemic would result in much lower sales, and reduced their component orders accordingly. When auto sales stayed stronger than expected, they were forced to buy those parts on the open market - reducing the number available to other customers.
2) A couple of major semiconductor manufacturers had incidents at their facilities (most notably, a fire at a Renesas fab) which limited production.
There was no “chip” shortage. There was an assembly labor shortage in China, a real shortage, cause by the disruption of the pandemic.
Where are you getting that idea? There was absolutely a chip shortage! A lot of electronic components which were readily available before 2020 became unavailable or prohibitively expensive, and stayed that way for as long as a year or two afterwards.
Arguably it's a Microsoft failure more than anything.
If they had delivered a version of Windows for ARM capable of seamlessly emulating x86 code with no compromises and released it a decade ago the world would likely be different.
Even now ARM is niche.
x86 already was on its way towards such a goal.
And it took the threat of the PC industry moving to ARM to light a fire under Intel.
they definitely were asking for longer battery life and cooling systems less noisy than a fighter jet, but the supply side of the market was a desert filled with minefields and duds until the M1.
So there's a catch 22 here. ARM was niche on desktop because no-one made the first move. That is, until Apple launched the M1, which is exactly the failure of imagination of Microsoft that GP is talking about. Though a quick google shows they did indeed investigate this avenue in 2016 for Windows 10, though they obviously didn't commit to it:
> On December 7, 2016, Microsoft announced that, as part of a partnership with Qualcomm, it planned to introduce support for running Win32 software on ARM architecture with a 32-bit x86 processor emulator, in 2017. Terry Myerson stated that this move would enable the production of Qualcomm Snapdragon-based Windows devices with cellular connectivity and improved power efficiency over Intel-compatible devices, and still capable of running the majority of existing Windows software (unlike the previous Windows RT, which was restricted to Windows Store apps). Microsoft is initially targeting this project towards laptops.
https://en.wikipedia.org/wiki/Windows_10#Features_in_develop...
This is hardly possible on most ARM chips because x86 has a much stronger memory model than ARM. Also concerning "with no compromises": common x86 implementations have a very fast implementation of the SIMD instruction sets (SSE..., AVX/AVX2, perhaps AVX-512) that is much slower to emulate on ARM because their SIMD instruction set is different. The only reason why people don't realize this is that a lot of common software makes no intense use of these SIMD instructions. Then there are the subtle parts that (as far as I am aware) the ARM FPU handles multiplication of denormalized floating-point numbers slightly different than x86 (both implementations are allowed by the standard) etc.
Why? My impression is that due to Apples permissive licence, they are able to make more changes. If we just look at the mess that ARM is making of Qualcomm situation, where we finally have a performing alternative chip.
Arm hasn't granted anyone else the ability to make significantly modified chips for multiple platforms, so Nuvia were fine while it was all just research.
> Arm's claim against Qualcomm and Nuvia is about protecting the Arm ecosystem and partners who rely on our IP and innovative designs, and therefore enforcing Qualcomm's contractual obligation to destroy and stop using the Nuvia designs that were derived from Arm technology
Its rather telling that Qualcomm are allowed to make mobile chips, but not these ones because they want to "protect the ARM ecosystem".
RISC V can't come soon enough.
> Arm hasn't granted anyone else the ability to make significantly modified chips for multiple platforms,
> so Nuvia were fine while it was all just research.
Qualcomm acquired Nuvia with the clear plan of using their IP for other use-cases ("powering flagship smartphones, next-generation laptops, and digital cockpits, as well as Advanced Driver Assistance Systems, extended reality and infrastructure networking solutions"). Apart from the fact that Nuvia had no license for these markets, the acquisition voided the license they had and the contract states that without license all the IP developed under the license has to be destroyed.
There is no dispute on this, Qualcomm confirmed this in court, but they argue that they have ANOTHER license they want to transfer the IP to, and that the terms of the ARM-Nuvia Agreement are offending and "a threat to the industry"
> Its rather telling that Qualcomm are allowed to make mobile chips, but not these ones because they want to "protect the ARM ecosystem".
ARM has designed new architectures (Blackhawk, Cortex-X) which achieve comparable performance to Nuvia's IP, but Qualcomm's assumption is that they can apply Nuvia's IP on top of their existing architecture without the need of licensing any new ARM design.
That’s a marketing tactic that even works on HN I’m afraid. Create an enemy and then tell a story about that nemesis.
Our human storytelling brains need to hear tales of good and evil.
If you want to get an article like that ranking that’s one way to do it. Unfortunately.
I guess you mean that the story is harmfully reductive or demands further critical examination, but I just want to defend telling oneself stories as a thinking tool.
I wanted to say that good vs evil are hardly absolutes. No one at a large corporate comes to work saying: today I will be evil again.
Instead they tell themselves they’re helping investors getting a good return, or that they’re “smashing Google for copying iOS”.
There is a great quote by Shakespeare saying: “There is nothing either good or bad, but thinking makes it so.”
I’d say storytelling makes it so!
The many parts of the agreements were clearly net negative outcomes for markets and users. Basic forms of collusion and cartel making were common.
- Screw everyone if I get more money and power
- Self-aggrandazing evil
- Hurt people for fun evil
- Comic-book style take over the world evil
- get them before they get you/ we have to do it because our competitors will.
ie fear is the driver, 'survival' is the goal.
I'd argue that the use of motives to define evil is often used as a - when they do it it's bad/ but when I do it it's fine - excuse.
For that reason I'm not really interested in the 'evil' part of this conversation - rather more what drives perceived bad outcomes.
So all I'm saying is fear is a big driver of many of those bad outcomes. That's not to justify it - but rather understand it as a first step to fixing it.
Note I'm also not saying whether the fear is rational or not.
I’d encourage you to have a conversation with Theil or Larry at some point and really hear in person how detached to the point of ‘evil’ they can get — and wrapped up in reshaping society for their own egotistical reasons.
Do you have a source for this?
Right now there are three. Intel, Apple, and AMD. In 2022, there was one.
Samsung manufactured for Qualcomm who were in testing phase.
It isn't Apple who is the problem here, it is Intel who refused to manufacture ARM chips for Apple. So Apple didn't have much choice as there aren't much fabs around to choose from. Now Intel, sore looser, is competing for CHIPS money against TSMC/Apple (and Intel will definitely get and waste a lot of those CHIPS money thus making the problem even worse). The article, which looks like a hit-piece, is written by a think-tank funded by Omidyar who is pushing for CHIPS and seems to be connected with Intel a bit here and a bit there. Did somebody say "submarine"?
Notable also is the omission of Nvida - which I also believe mostly uses TSMC.
ARM allowed companies to focus on design, and not having to worry about manufacturing. Nvidia is mostly dependent on TSMC, but truth is they can move to SK Hynix and Samsung without a drop in quality. (Basically whoever has the capacity to build the chips at their nm level).
I laud the authorities for stopping Nvidia/ARM acquisition a couple of years ago. More than any other acquisition, this would have skewed things up massively.
The intent is pretty clear and uncontroversial: profit by all means, like any other for-profit entity. Real problems arise when the company becomes (nearly-)monopolistic and their actions start to harm whole industries, like in this case. In other news, monopolies are bad for you.
If intel/amd/etc can't build their own fabs, and expect we-the-people to pay for it, then we-the-people should jump to socialism and nationalise the lot.
If we have to pay for it, we should own it...
I feel like I'm getting trolled by the author who is citing sales and profits instead of units produced as a metric.
Moreover, give iPhones are significantly more expensive than comparable Android phones, Apple’s ability to claw market share back from Android phones over the last decade is the opposite of how a monopolist might operate by flooding a market with cheap alternatives. It speaks to a consumer perception that iPhones have a larger value.
[1] https://gs.statcounter.com/os-market-share/mobile/worldwide [2] https://www.statista.com/statistics/266572/market-share-held...
Google would not be doing it if it wasn’t for the advertising business. It’s pretty much the definition of a loss-leader to get people into the ecosystem for things that google can monetize. So android as a product is implicitly, foreverially tiedup with marketing and spyware, because google isn’t onboard otherwise. Same as gmail or search was a loss-leader to get people onboard for advertising too. Google only cares about these things insofar as they might stop being a funnel into their money-makers.
SOC vendors can’t make a run providing 7 years of driver/firmware support for a product they sell once at bleedingly thin margins. Or at least, they really don’t wanna.
OEMs can’t make a run providing 7 years of support for someone else’s software, especially when they also have to do a lot of the driver work themselves thanks to IHVs abdicating their job.
Consumers get stuck with a product that loses support in 2 years or whatever, and may even have landmines involved with unlocking it to continue support (Sony wipes the camera firmware if you unlock the bootloader for example). They face a completely unnecessary hardware and software treadmill due to all these other factors. Supporting your own phone is not a reasonable expectations for Joe Sixpack either.
The idea is supposed to be the “linux model” but honestly linux has the same problems, it is reliant on the same unpaid labor around driver work to make up for the inability of vendors to track the ecosystem and provide the long-tail of support. In cases where the vendor can’t open source it, the functionality simply ends up broken, and driver support, kernel versioning, and DKMS is a constant battle for end users. Just like with custom roms for android.
Android simply has too little margin split among too many disinterested parties to ever really work. And fixing it would involve either increasing the size of the pie (margin), which consumers in this segment hate to an unfathomable degree (android users = cheapskates is a reliable first-order approximation, borne out by the app store revenue too).
But that's the free-market system working as intended, right? Literally every penny has been squeezed out of margins, software costs pushed onto free labor in the open-source community, and ad revenue used to contra-fund and push end-user prices even lower. Android is the finest solution the free-market can deliver, that's how the system is supposed to work, and it’s delivered an excellent product for the needs of the customer - it's just you're not the customer, you're the product.
The alternative is vertical consolidation and bringing more things under the same roof, raising the price, and targeting the consumer needs instead of the ad revenue needs. Basically the apple model. But that can never be a viable path in a GPL world. And it will still probably involve paying more - phone costs are currently subsidized by all these indirect costs like ad money and vendors cutting corners on support. There is more work that will need to be done, and that contra-revenue from advertising revenue needs to be backed out of the purchase price, so at the end of the day consumers will simply have to pay somewhat more (hopefully not apple prices of course). But again, people are cheapskates, android users doubly so.
I don't know why people got so allergic to the idea of paying for their operating system, the baseline assumption now seems to be that it needs to be free, and if that's the case you will never be free, only stuck in a choice between advertising-mongers and exploiting unpaid labor. And that can either be in hardware costs, or in actual recurring support costs, but either way, someone needs to be paid to sit down and make sure the bluetooth and sound drivers work.
You see the same problem in software too - open-source projects get commercial entities tapping their value without providing contributions back, or existing via patronage to the needs and goals of the commercial entity. Without an incentive by the actual developers to provide end-user value, and with permissive licensing, you end up with a constant struggle for financial homeostasis. Firefox/Mozilla, for example.
People complain about android but they still are not willing to pay a little more to opt-out of these problems. The revealed preference is for purchase price above all else, and people still think in the yardstick of Apple being "too expensive" rather than Android being "too cheap". The yardstick is still the artificially-cheap advertising-subsidized Android product.
These are fundamentally problems of not enough margins to support all the players in this ecosystem, which leads to them looking for places to find the revenue to make it work. Pay a little more and these problems go away.
If you want to stop being the product, get used to opening up your wallet. That transactionality is a good thing - you can’t really demand boundaries when you’re living on someone else’s dime. It’s Google’s house and they’re letting you crash for free. But this is the very deepest core of the problem - people will do anything except just pay a little more.
You will never stop being the product if you can’t bring yourself to be a customer.
(Yes, I pay for kagi, how did you know!?)
This doesn't negate any of your points
It’s about “code revenue” vs “services revenue”. If I can’t sell you a copy of Office 97, well, obviously that pushes things to the Office 365 model. And the unfortunate reality is that a huge amount of important code is now GPL, which means a huge amount of the world’s codebase is now un-monetizable in that sense. This has fueled a massive push for “alternative revenue”, and the whole google model has been that the "alternative source" is advertising dollars.
Moreover, GPL itself has conditioned people that the right price for software is “free”. And obviously software is not free to produce, nor is the ecosystem of phones conductive to the Linux “hardware bazaar” model in the same way as 1990s era IBM-compatible PCs. [0]
Well, if the only valid price for consumer-facing software is “free”, and a huge amount of the software in the world is now un-monetizeable in the sense of being able to sell a copy of the software to fund your R&D effort… obviously that just pours gas on the fire of the tivo’ization and SaaS revenue models. That is a correct business-development response to the changing market conditions.
GPL has essentially forced the collapse of the traditional “fee for source” or “fee for binary” model. I think it probably would have been out-competed by service revenue eventually anyway, given that consumers obviously prefer “free” to paying money, but basically this is an accelerationism thing whereby GPL has more or less collapsed the entire proprietary-software market (to the extent that many people now view hardware that doesn’t have open drives as being somehow fundamentally bad or illegitimate), driving everyone into the arms of the SaaS providers. It has accelerationism’d us right into tivo’ization.
Big “my neighbor’s cats keep getting eaten by coyotes and he says that he just goes to the shelter and gets another cat and I said it sounds like he’s just feeding shelter cats to coyotes and his daughter started crying” energy. Like GPL literally has been so successful that people think proprietary software is illegitimate… except for the “free” ones backed by advertising dollars.
GPL has been feeding consumers to the coyotes. It doesn't mean they meant to do it, but, functionally by killing fee-for-software models and by conditioning everyone that the only valid price is "free", that's what it's kinda done. Proprietary software still exists... you just don't pay with money anymore.
I am of course human too, I groan at the thought of paying $350/year for a personal jetbrains license or paying $1000 more for an apple laptop than a comparable commodity one... but software costs money to develop and you make your choices about what is worthwhile and which relationships you want to accept enshittification on. My point is just that transactionality (ie, you pay money and receive a service) is actually a good thing, because those services are much less likely to enshittify if it's going to come at the expense of actual paying customers rather than just some livestock waiting to be sheared. It changes the nature of the relationship.
It doesn't mean there aren't bad companies that do enshittification anyway, but if the expectation is pay money => receive good, then obviously you can reasonably hold expectations about the nature of what you're going to receive, vs GPL and proprietary models fostering the "it's free, why are you complaining" mindset. GPL and SaaS models are alike in that respect and GPL has both directly reinforced that mindset, and also pushed far more companies to SaaS far more quickly with its own success.
[0] (Wireless+modems have specific regulatory requirements that are inherently incompatible with open-licensing, it is illegal to release a wireless chipset which is capable of violating local band-regulation requirements, which functionally imposes the requirement that devices listen to the local band to detect weather radar upon bootup to see if they're allowed to use the band in a given locality. This is part of why modern routers are so slow to boot up! And to prevent tampering/replacement, this effectively must be put in a closed blob to pass certification. Which is why projects like OpenWRT are struggling with Wifi 6E/6GHz and other newer hardware - when I checked maybe a year ago, there were zero 6E devices supported. And unfortunately, it is not possible to make a very good phone without wireless connectivity. It literally is one of the most challenging domains to try and go open-license on because the FCC ain't playin'.)
https://news.ycombinator.com/item?id=34134905
--
Outlawing this "contra-revenue"/ad-supported moel is low-key one of the biggest things the EU could tackle though. Big problem, wide problem.
Granted they're walking down that path with GDPR dismantling a lot of the worst of the surveillance capitalism, and with forcing Android to have actual reasonable OS support lifespans... they are gradually forcing these externalities/defects to be priced in properly. But I think it is explicitly worth stating as an end-goal: transactionality and fee-for-service or fee-for-product is a good thing, because then companies are competing to service customers and not competing for more livestock to be the product.
If you want to stop enshittification, get rid of contra-revenue models. The thing you are paying for, should be profitable to produce at the price you sell it. If you have engineering effort that's shared with other products: fine, amortize/attribute it however makes sense. But if you are selling a widget that costs $10 to produce for $9, on the expectation of $3 of revenue: no, that should be illegal.
If nothing else it's a competition issue, you can't have "honest" firms ever break this situation if you have other companies "dumping" and selling below-actual-cost. It is really no different from any other kind of dumping.
Otherwise, the advertising-supported companies will always be able to out-compete the "honest" companies, so it will always be a race to the bottom.
--
anyway, as far as hardware startups specifically: those exist and they just haven't gotten any traction. Fairphone, pinephone, lightphone, etc. These inevitably end up getting very little traction among enthusiasts etc. I don't think the existence of these solves any real problem with the android ecosystem.
Also, ecosystem is a major factor in OS adoption. Like you could go run MenuetOS right now, so why don't you? Ecosystem and network effect and available codebase etc. So these hardware startups have to work within the existing OS ecosystems etc - almost all of these startups are android and the ones that aren't, frankly are doomed. Consumers don't want featurephone or phone-specific app stores in 2024, especially for some niche product with no actual apps released for it. There is very little point to making a custom OS from vxworks or whatever, that's not where the problem lays here.
Similarly: Framework and System76 are attempting to thread that needle of supported hardware+OS configurations on "generic linux". This exists, just not to massive success or fanfare or changes in how the world works. And they're not re-engineering the OS. Even System76/PopOS is based on ubuntu iirc. Jumping to MenuetOS or other completely-new-OS things is a whole other can of worms.
Despite Framework/System76 existing, the world still runs on Windows 11 ad-supported installs or free Linux labor. So I don't know why it would be different for phone hardware+android.
things like that really need to cost money, because they cost money to deliver. and "I get it with my ISP" isn't a satisfying answer either.
SMS 2FA is another similar problem. It's basically an identifier token at this point and you have very little choice over it etc. I actually have been denied for credit cards twice (I think, microcenter told me why it usually fails for them and it makes sense) because my phone number is in my parents' name... (I pay them for a line on a family plan as an additional line). It's obviously not the same problem but it's getting at the problem of these "bundled" services having become de-facto identity providers.
Cause it's not adobe making photoshop free, right? It's bundled search, mail, identity provider, as a free service (you don't pay with money).
But there are many many smaller places it happens etc. I don't quite know how to draw the line of "this is a feature that's bundled" vs "this is actually a separate service" but like, in the large picture there are places where it's not really questionable, right?
"Free as in freedom" is obviously on the clear side of the line. I don't quite know how to draw it for everything, because again, you can see ways where "free MS windows" could be exploitative even if it were free. I guess that falls into the "actual damages" sort of regulatory scenario.
I haven't exhaustively developed the concept but yeah, I mean, "dumping" seems obviously problematic in a market-fairness sense.
Google and Samsung tried and failed. Google with their flagship Pixel line, which just doesn't sell well, so they moved downmarket. Samsung with their Galaxy line, but they're always playing catchup/copycat to Apple.
The sad fact is that 1) Google sucks at hardware, 2) Samsung sucks at taste.
Apple really has a magical combination of (arguably) good taste, and operating chops to deliver the hardware at scale.
People really paint Apple and Android ecosystems in auras that weren't true already a decade ago.
Agreed that Samsung is just playing copycat to Apple. They used to make great, innovative phones without being a copycat, like back in the S4/S5 days, but then they just decided to copy everything Apple did, like eliminating the headphone jack.
Not sure what kind of competition is needed to browse the internet, watch YouTube, listen music and occasionally shoot few pics.
On top of that, I have been locked out of my apple account for 2 months when my iPad with the authenticator suddenly died, good thing I had a desktop to use as I could not login in my MBP. Apple just kept sending me emails of them contacting me in two weeks to reset my password and kept not doing it.
And I won't even mention their "privacy" charade. If you forget your MBP password you can reset it in recovery mode and thus access any MBP.
Sure, all your apps get logged off, but you can still open any file.
I hate their products and wish I didn't have to deal with them, but I have to test apps and websites on their hardware.
My experience switching from iPhone to Android is I lost pretty much nothing and gained a lot, particularly by going to a folding phone (which Apple obviously doesn't offer yet). Android is much easier to work with in most ways - iOS is so set in its ways it can be very difficult to do basic things like transfer files from the phone to a computer.
Apple's processors may be faster, but this is becoming an increasingly less useful distinction on phones. The only thing on phones that continues to require increasingly faster processors is stuff like games, and fair enough, but I honestly just don't play them on my phone personally.
This misperception is key to understanding the market, which is really two markets: Android has cheaper phones and dominates there but in the mid to high-end market the situation is reversed because the equivalent Android devices aren’t cheaper and because Qualcomm/Samsung lagged so far behind on CPU performance you’re getting something which performs like 1-2 iPhone generations back in most apps.
Breaking out of that dynamic is hard because the Android manufacturers have to share more of their hardware revenue with less service revenue to compensate, so they don’t really have much room to lower prices since they’re already underperforming at the same price points.
What could change a lot would be regulators forcing App Store competition or limiting revenue sharing across units. Apple and Google both benefit from that at the expense of the pure hardware vendors, but I’m not sure how effective e.g. the EU App Store regulations will prove in practice.
A bit off topic, but I personally find Google, and by extension, the Android ecosystem, to be an underhanded business model. I don’t feel bad it’s ending poorly for them. It’s especially rich Google ripped off Apple in order to get Android launched [1].
[1] https://www.mactrast.com/2013/12/inside-story-android-ripped...
Apple makes 25% of its profit from laundered ad money. Like you need to stop trying to understand a sophisticated duopoly ecosystem from a fanboy blog.
Even if you're the most anti-Google person in the world, you have to admit it's a bit ironic that Android can be built with all Google services disabled whereas iOS, due to it's wonderfully-designed Google Search integration, cannot.
That won't stop a hefty share of HN posters from repeating that same whinge.
"Global Smartphone Shipments (Millions)"
https://www.counterpointresearch.com/insights/global-smartph...
That is actually terrifying, because logic dictates they make up profit in other ways. I can imagine how an ad company or nation state can go about it.
The vivisection of the PC division, the semiconductors, iOS, iPhone, and iPad into separate companies would be deeply satisfying.
Shareholders might actually applaud the results in the end.
Because the article is talking about semiconductor manufacturing, and the total numbers of semiconductors purchased.
Profits and monopolies are a different article.
Not sure if this is good or bad, but it's certainly unusual, and it's part of the reason that this "monopoly" question is sometimes confusing--Apple makes by far the most profit despite also selling the highest-end product, and their unit sales, despite being somewhat less than Android, are still in the same ballpark.
Because it's Matt Stoller.
I've commented (with citations) enough on HN about Stoller's biases and issues around analysis [0][1]
He's upped his content publishing the past few months now that he is competing with Oren Cass to change mindshare and build a prominent profile, but Cass is a Romney protege so he's limited.
If Trump becomes president, Stoller is most likely going to be nominated as an FTC commissioner because of his close relationship to JD Vance [2]
[0] - https://news.ycombinator.com/item?id=40624532
[1] - https://news.ycombinator.com/item?id=38147499
[2] - https://nationalpost.com/opinion/matt-stoller-whos-afraid-of...
His most damning quote was from 2023 that “fundamentally, my view is that authoritarianism is coming from the private sector,” not from those who’d attempt to seize control of democratic institutions by force. [0]
> Trump will "most likely" fire Lina Khan and hire Matt Stoller
He doesn't need to fire Lina Khan (though he most likely would) because Melissa Holyoak's term is expiring in 2025
[0] - https://www.politico.com/news/magazine/2023/04/21/matt-stoll...
But I find an article that doesn't even address any other companies a little hard to swallow. Are other companies simply not able to "reshore" chip fabs?
I have no idea what some of the tidbits thrown in about the apple watch has anything to do with the topic. Let alone some of the heavy handed lines in the article that simply aren't supported as far as I can tell.
Some of the lines in this article seem borderline hysterical...
Intel has fabs in the US, doesn't seem to me after a couple readings that the article isn't about CPU fabs, its about the semiconductor supply chain.
> Some of the tidbits thrown in about the apple watch has anything to do with the topic
They seem part and parcel with ex. the App Store mentions, antitrust, and monopoly references.
> Let alone some of the heavy handed lines in the article that simply aren't supported as far as I can tell.
Such as?
Basically, it's mostly: Apple is lying with their marketing, here's what proves they really don't care about moving production to US, but only about price and their profit margins.
I do like the US - don’t get me wrong - but Apple is in the business of making devices, not making investments the government should be at least facilitating.
Be mindful that such tariffs wouldn't hit Apple alone.
Because Apple's dominant position as a monopsonist means it has price setting power and depresses the profits in the chip industry, making all but the cheapest locations effectively unsustainable. Monopsonists essentially absorb the profits of their suppliers. This is for example why nationalized health services are cheap.
In your example, not only has the government huge bargaining power, but the law itself limits other customers, and this there is no demand outside of the government. In contrast, there are plenty of customers for semiconductors. Demand is still here.
So yes, Apple might reduce TSMC’s margin, but the real problem is not that. It’s that there is nobody else operating on the same scale. The fact that there is no alternative for cutting edge nodes is a sign of a market failure.
Besides, this “it’s Apple fault and they depress margins” sounds disingenuous. Apple famously invested quite a lot in TSMC’s fabs in exchange for this access to the latest nodes. And TSMC also benefits from stable, predictable demand. The real problem is the spectacular failure of Intel and the complete lack of effectiveness of the handful of governments that could do something. Things seem to be changing, even if it is slow, but we should never have been in that situation in the first place.
there aren't that many, that's the point. Apple is responsible for most of its suppliers revenue as the article points out, otherwise it wouldn't have price setting power to begin with.
We aren't in a supply shortage, with the exception of a bunch of AI related chips (the only thing the media talks about, hence the skewed perception) we've bin in a supply glut. Semiconductor revenue declined by 11% last year. AI chips aren't the only chips in the world, DRAM makers like Micron, an American manufacturer are the kinds of companies that are effectively at the mercy of Apple because of how outsized of a portion of the market they are.
There’s at least Qualcomm, AMD, and nVidia. All of these are huge. And then there are the smaller niche actors (who tend not to use the cutting edge, but help with RoI by forming the long tail). Demand is clearly there.
> Apple is responsible for most of its suppliers revenue as the article points out, otherwise it wouldn't have price setting power to begin with.
Apple has price setting power because they helped funding the fabs and the process. I agree that the result is not great for the industry overall, but they did not show up one day and kick the other customers out. Again, the main issue is the limited bandwidth because everyone depends on effectively a single supplier.
> we've bin in a supply glut.
Not on the leading nodes, though, right? The ones over which AMD, Apple, and nVidia are fighting. Or I missed that, in which case I would love to have more information (I mean solid, with numbers). That said, if that’s the case it does not help the story’s argument. Oversupply leads to lower margins, but one customer taking a large chunk of production restricts supply, which has the opposite effect.
> AI chips aren't the only chips in the world, DRAM makers like Micron, an American manufacturer are the kinds of companies that are effectively at the mercy of Apple because of how outsized of a portion of the market they are.
DRAM and CPUs tend not to use the same processes, so the only way they are in direct competition is because investment is limited but that point is undermined by Apple financing the fabs. So what is the actual problem here? IIRC Micron is investing quite a lot in new fabs in the US, with the help of the government, which is exactly what needs to happen.
The more I think about it, the less convincing the whole thing is. The story is a collection of rants because the author does not like Apple, but it is full of contradictions and non sequitur. There are some good points beneath, but they are well hidden and not directly relevant.
This is why we need right-to-repair legislation with teeth.
How is this relevant to the chips TSMC manufactures? You can't exactly hand solder a 24nm feature.
I'm not sure there's a fundamental understanding of how large businesses are changing their approach to supply chain in the name of competition with this statement.
The choice of these manufacturers is either take the volume order from Apple at lower margins and with stipulations or deal with market volatility. It's the stipulations aspect of what Apple is doing that is anti-competitive and generally horrible for consumers. Do you happen to work for Apple? Seems to be a very Apple-apologetic post for no good reason.
> Also, I can guarantee you that there’s a good amount of vendors that would rather not be forced to offer their chips that they provide to Apple on the general market.
OK. I'll take you up on that guarantee. Feel free to reply with reputable sources on the matter.
I don't, I just work in the industry. These aren't an either-or decision on Apple or not Apple products. IP from developments on Apple-exclusive stuff is used elsewhere. Look at TI's USB-C port controllers as an example used in Macbooks. They have a custom one for Apple, but you can buy their general market parts online just fine. It's not like Apple is preventing TI from selling USB-C controllers.
> OK. I'll take you up on that guarantee. Feel free to reply with reputable sources on the matter.
You won't find 'published sources' on this. You'd need to have a fundamental understanding of semiconductor industry. Support for general market customers is HARD. They tend to be particularly needy, especially compared to the return on investment. If your organization is designed around helping large customers (not just Apple, but any large customer) then you just do not have the appropriate business structure in terms of sales, distribution, etc, to manage that. Companies like TI/NXP/ST have that infrastructure, but there's plenty of smaller companies that just do not really support small customers.
I get that repair shops 'just want to buy the chip', but that's really not what making parts available on the market means. It means fielding questions, publishing datasheets, having FAE/sales support, having distribution channels, etc, which all adds to cost for very little ROI. Repair shops are basically a rounding error in terms of buying power compared to a Samsung, Apple, etc.
This is where your argument breaks down at the fundamental level because this is flat out incorrect. Organizations like Digikey, and many others, do exactly this. They've been around a long time and they cater to low volume buyers at reasonable margins.
Unfortunately your argument seems rooted in missing fundamentals and conveniently are ignoring entire market segments to bolster your argument that Apple isn't in the wrong here. They are. And what they're doing is anti-competitive.
They also don't produce any documentation, nor produce any application boards, collateral, drivers, etc.
Market segments that a company (a semiconductor company) wants to service all have different needs. General market/distribution is a very different business from custom/large customer business. Just like it's very different from industrial. Or automotive.
The title tends to give this impression, to be honest.
I urge you to look at the earnings of some of the most popular chip manufacturers and then look at the earnings of apple (also ignore nvidia as it is an obvious outlier). Creating a modern chip fab takes A LOT of money and is simply impossible when some of the most popular semi companies struggle to even make a small profit. TSMC's fabless business model allows many chip companies to share the capital investment required, but apple just straight up has the money to make the investments but choose not too. Also it would be quite scary if apple is allowed that type of vertical integration.
Scary in what way?
Why would that be? It’d free up TSMC’s top shelf capacity to other manufacturers.
Certain kinds of vertical integration can last, but when a key part is a moving target (fab capabilities) or commodified then they are inherently time limited assuming any kind of competition (and there is plenty of competition in the PC and phone markets still).
That'd be a risk, but they could also pull ahead if both TSMC and Intel drop the ball on one generation or another.
Apple also did that because it is a _business_, and has not always been the behemoth it is today. They went overseas because it made more sense, not because they had some anti-American goal of crippling US mfg capabilities.
The AELP is, ideologically, a anti-big-co institute, which is admirable but is also why this piece engages in post hoc ergo propter hoc rather than discussing meaningful policy proposals:
> The American Economic Liberties Project launched in February 2020 to help translate the intellectual victories of the anti-monopoly movement into momentum towards concrete, wide-ranging policy changes that begin to address today’s crisis of concentrated economic power
Now you're on the hook for showing us a company that cares more about people than profits... and I mean actually cares where it counts, in budget and headcount and not just in gauzy HR propaganda.
I think Intel probably deserves a mention in the article but that line of question feels like the wrong angle of approach for it.
My only point is that the story of Intel’s disorganized retreat is at least as relevant as Apple’s march forward
Re mobile:
1. Timelines. Apple only transitioned to Intel in 2006 on their computer side, the first iPhone was announced in 2006. As a company they had little overall Intel experience. There was no "sticking with Intel" because they'd just switched to Intel concurrent with the development and release of the iPhone
2. iPod. Apple's prior mobile devices were already using ARM processors. That's a lot of experience and existing supply chain connections to take advantage of.
3. Mobile Intel chips sucked. They drew too much power and were not suitable at that time for a mobile device like a phone (already more power hungry with the wireless components compared to something like an iPod).
https://www.theverge.com/2013/5/16/4337954/intel-could-have-...
Also, when Apple switched to ARM for their desktops the core motivation may have been to bring everything in house. But they gained a momentous advantage in the market because the chips they made were so much better than Intel’s. Again, explaining what happened to Intel to make them lag the cutting edge should imo be the key question, not why Apple made a rational business decision (albeit possibly an unethical one. Shareholder capitalism has entered the conversation.).
https://www.canalys.com/newsroom/global-smartphone-market-q1...
If not, then the logic is strenuous at best, you just viscerally hate Apple and want to find reasons to blame it as opposed to doing anything constructive for the American Semiconductor Industry. Something constructive would involve fixing the work ethic of American Semiconductor Factory employees which seems to be heavily lacking compared to their Taiwanese counterparts, a problem that we don’t see with say the AI workforce or any upper order white collar workforce.
How do you pinpoint this to be the problem? How are you sure that it's not some narrative used by corporate to excuse outsourcing or putting pressure on wages?
I think an industry where technical knowledge was outsourced for cost cutting reasons for decades now is probably having issues because only the managers are left, and they're not productive on their own. Blaming it on the factory worker seems uncalled for.
My anecdotal experience aside, there have been quite a few news articles about TSMC’s struggle in Arizona that highlights the difference in work ethic: https://restofworld.org/2024/tsmc-arizona-expansion/
I'm not convinced this is because of the workers still. What you have in Shenzhen is that you have everything in one place to build that PCB. Everything your factory doesn't have, you can get from a neighboring company. The supply chains in the US have been starved for years, so if you want to do the same from the US, you have to wait for parts to be shipped from China. Making quick iteration impossible.
Shipping only the finished PCB is of course faster than if the development of the PCB needs multiple shipments of parts.
I don't have a solution for this, but the US outsourced so much manufacturing, that it can't manufacture efficiently anymore.
If your app shows ads, apple customers are more valuable.
If your app sells any kind of product apple users buy more.
Market power is not determined by the number of users but by control of revenue.
When a buyer has outsized market power it's called a monopsony.
I guess that's the concept gp is reaching for.
The "smell" of a monopsony is the ability to squeeze suppliers to the point of having the power to drive them out of business if they don't meet your terms.
Regardless of what you call it, any market vertical that you can define with less than 5 genuine competitors invariably winds up being anti-consumer and needs to get broken up.
Apple is the focus here, but Google is no better and neither is Microsoft. Search, messaging, email, video--all of these are easily definable market verticals.
All of these need to get broken up until we have at least a half-dozen competing companies in those spaces.
This isn't limited to tech. Grocery store chains, industrial suppliers, financial companies, etc. all need to get smacked with the same breakup stick--especially if we want not just to not only stomp out anti-consumer behavior but also to have genuine supply chain resilience.
You seem to be forgetting, we had your “at least a half-dozen competing companies" situation in the past. It was called the late 2000s and the early 2010s. And the reason it disappeared is not the traditional “everybody merges until only two or three are left standing”, it’s because iOS and Android were so much better than the competition that every other phone manufacturer starved to death or switched to Android.
By the time a new generation of smartphones with sufficiently-equivalent OSes had arrived — in the form of webOS and Microsoft Phone — iOS and Android were so established the newcomers couldn’t successfully compete.
Even today, while there are only two major OSes, there are still numerous successfully competing manufacturers: Samsung, Google, Apple, OnePlus, BLU, Lenovo/Motorola, Huwaei, Xiaomi, Vivo, Oppo, etc. They all make mobile phones, with varying levels of market share across the world.
I would argue that it was because Apple and Google could cross-subsidize the market from their other buisnesses.
Google viewed the ISPs as an existential threat and saw both mobile and fiber as a way to disintermediate them. Apple saw phones as an existential threat to iPod which was a huge chunk of their revenue.
Google didn't release Android until 2008. As of 2009 Nokia still had 33% market share and even Samsung(13%) and RIM (15%) beat Apple (11%).
> iOS and Android were so established the newcomers couldn’t successfully compete.
That is practically the definition of monopoly, you know?
A monopoly using anti-competitive tactics attracts regulatory review, but anti-competitive tactics are forbidden for both small and large companies.
> integrate so much better
Integrate with what?
with other Apple products of course!
Congrats on being judgmental without even trying to understand anyone else’s perspective I guess.
I'm not sure what the author is suggesting here. They licensed technology from Imagine Technologies, so therefore they must continue licensing their technology?
Also: Imagine Technologies is a chip designer, not a chip fabricator. It's stuck in at the end of a paragraph about Apple's suppliers, but unless I'm mistaken—which I absolutely could be!—Imagine Technologies wasn't one. Also it's a UK company, so I'm not sure how Apple ending their contract with Imagine Technologies hurts the US's semiconductor problem.
I think you're overreading it. I don't think the author is arguing that Apple must stay a customer. Rather they are making a point about how much power Apple has a company. They clearly have the power to make or break other companies. That of course doesn't make the rest of the article correct, but the point about Apple being very powerful does make sense to me.
Maybe, but to me it reads more like "I'm writing a piece complaining about Apple, so I'll bring up every complaint about Apple I can think of, even if they're not really related to the article topic itself".
It's not just about semiconductors, but the entire supply chain.
https://www.theatlantic.com/business/archive/2012/01/why-the...
https://www.nytimes.com/2019/01/28/technology/iphones-apple-...
The Imagination Technology example in particular is poor since they were a supplier of intellectual property, not physical goods. So Apple’s usage of them as a supplier had limited bearing on the firm’s other customers or lack thereof. Similarly, the ban of Apple Watches over IP theft has nothing to do with the supply chain issues that the authors attempted to outline.
The thesis is also all over the place. The title indicates the problem is semiconductors, but then lots of the arguments have to do with suppliers outside of semiconductors. If Apple’s reliance on TSMC is the big problem, then why isn’t even mentioned that TSMC is the only competitive fab on cutting edge nodes?
Lastly, the overarching theme is severely misguided. All of these supply chain dealings are as old as semiconductors themselves. When the industry moved towards targeting consumer electronics in the 70s, semiconductor manufacturing moved out of the US long before Apple was relevant. And if Apple weren’t a powerhouse of consumer electronics today, other companies would be doing the same things. It’s just part of the nature of trying to manufacture and sell mass market electronics.
With as much expertise as the author has. I didn't read solutions or ways forward. I suppose the main objective is awareness?
Then it goes on to state that Apple stole technology when that is not actually a fact -- it's part of an ongoing legal battle that hasn't been proven (innocent until proven guilty in America right? Right?).
So, it just comes across as a weird hit-piece with some pretty shaky arguments.
Except if you're a company so large that the government has no choice but to bail you out in a crisis.
The Soviet situation is top down planning based on what "should" be produced. Actual consumer demand isn't important. The Soviet union would decide some years that industrial equipment was the priority, then years later consumer goods.
Apple is bottoms-up (through customer and market data), with the leadership approving recommendations from the bottom based on what the consumer wants.
What would that be exactly? Seems like an easy thing to say without specifics, but things like environmental rules and minimum wage laws exist for reasons. If those are obstacles, I'm not sure they're worth discarding.
How do you "create an environment" without "imposing more rules"? If you don't really do anything, status quo remains. I guess you could be referring to rules that need to be abolished, so I wonder which ones are really stopping competitors to Apple from springing up that we should do away with?
If Apple had not been there, are we saying manufacturing would have stayed onshore - steel foundaries, car factories and chip manufacturers- I mean China and India account for over 3 billion people, who want roads, houses, cars and washing machines. Where else are the factories going to be - the weight of demand means even if US owned the factories they would still be in Guangdong.
So yeah, Apple probably misuses its market power, but it’s hard to imagine it did anything but tip the playing field further down the way it was already pointing.
So if we tax or punish Apple, do we really think it will all flip back?
My personal take is that energy and computing are the things that need to be in shored - no matter the cost differential. Industrialised counties, hopefully, will become energy independent (lots of solar, lots of insulation, lots fewer transport) and computing is going to be similar - focus on that verifiable secure designs
"With this position, Apple uses its outsized buying power to squeeze the margins of its suppliers such as Foxconn, leading to poor pay and terrible working conditions in Chinese factories."
So Foxconn is either choosing to not do business with Apple's competitors who would pay more (??), or the competitors can't place orders of the magnitude that Apple can and Foxconn workforce would just be smaller with a lot less people being employed.
The intellectual dishonesty in the argumentation is so bad that I'm pretty sure I'll skip future Matt Stoller articles.
But there's a hint of dishonesty in your retort too: obviously, there are competitors who would pay more (and they are paying more elsewhere, since their margins are so much smaller than Apple's), but for much smaller quantities of production. A supplier like Foxconn can't afford to lose a customer like Apple because they were already selling their services cheap with low margins. Apple is basically squeezing them ever more — instead of growing business needs leading to improved conditions for employees, Foxconn is forced to provide more for less because Apple can demand that "or else".
I am not saying this is true, but this line of thinking is still obvious just from the short quote you posted.
> (economics) a market in which goods or services are offered by several sellers but there is only one buyer
well that is my new word of the day I guess, not a bad one at that
I've been involved in a few exercises to decide what type of machine to use as a standard developer spec device. MacBooks generally turned out to be our best option when taking all requirements into account.
Linux-based offerings are getting steadily better, which is a good thing, but levels of support are still a bit of a problem.
In the 2 other companies, I started work with my personal Linux computer but they forced me to switch to a company-supplied Macbook later. It's not about costs.
Even non-profits often have to abide by the same compliance requirements, often insisted on by their funders.
Issuing a standard computer with a standard managed configuration just removes a ton of headaches for IT.
I'm sorry what? Poor pay and working conditions in the semiconductor hubs of China has been a problem as long as and before Apple made the iPhone. Apple certainly isn't HELPING that situation, but none of them are. All these fabs have made the news at different times for everything from suicide prevention nets to using slave labor to make Apple products, sure, along with every other major tech OEM in the business. I'd be shocked if you could find ANY large manufacturer of these things that hasn't been embroiled in one scandal or another over shitty conditions in their factories.
> Even for the American suppliers who have managed to stay in business, things are hard. With Apple accounting for most or all of the revenue of many of its suppliers—by buying most of their output and blocking its competitors from using similar components—suppliers “dare not put a foot wrong” by speaking against Apple, or even mentioning it by name.
Then how do you know if they refuse to talk about it? This feels like a piece banking on outrage clicks about Apple.
> In 2017, when Apple announced it was moving away from using UK-based Imagination Technologies for graphics processors, the company lost two thirds of its value overnight. Apple’s monopsony power means component suppliers have few buyers.
Yeah, Intel lost a bunch on the stock market too when Apple announced they were rolling their own silicon. Losing Apple as a client is absolutely going to suck for any supplier and a stock dip makes perfect sense in that situation.
Like, none of this is strictly wrong but it's just describing the highly centralized nature of this industry. None of this is unique to Apple. I'm sure Samsung would have no issues at all swinging suppliers by the tail if they were so inclined to do everything outlined here, and I'm sure they have too.
Edit: TIL about the word monopsony.
It's not unique to Apple, but Apple is far better at it and far more focused on it than anybody. It's Tim Cook's particular strength, why he took over after Jobs died and has been a major focus of the company.
When the dominant player comes to town specifically BECAUSE of these terrible conditions and low prices, and use their market power to push prices even lower, they are absolutely part of the problem.
I didn't read it as "Apple is the problem because they are evil", but as "Apple is the problem because they reached a critical scale in this industry".
No need to rush to defend Apple for being accused to focus on profit, no need to start 'Whataboutism' how other companies have similar goals. Someone describing how Apple took actions in the interest of maximizing its profit is NOT slander or bashing, it is FINE that they do that.
The article talks about Apple’s sheer size as a buyer and how their actions contributed (and partially even caused) the current state of this supply-industry, and how important it is to understand this and evaluate careful steps to counter this trend.
It also describes how Apple's narrative of being a "good corporate citizen" for forcing its suppliers to setup small-volume production in US is only half of the story, and WHY that is. And this is important to state as well, because Apple is and will continue to focus on maximizing its own profit, which is NOT making them the enemy, but also not an ally.
Now they pay the price
That's what happen when your government consists of clueless warmongers
"Oh no, Samsung is developing their own homemade industry, sabotage! quick!"
Yeah so here it is. The moment you start talking command and control economies you need to GTFO the room.
One of these citations isn't even about chips, but he kind of implies it is. Not to mention, he said that apple stole technology. That has not been proven, the court case hasn't happened yet. Citing sources out of context, and making untrue claims about what the source says, makes the article less compelling in my opinion.
So prices are so low, they can only be made using boarderline slave labor.
It’s 100% about borderline slave labor pay. I used to think it was because chip manufacturing is very nasty but I think it’s both pay and subpar working conditions.
This is the part I don't understand... they have placed orders with the new fabs that aren't even built yet. How are they killing American manufacturing if there is no American manufacturing for them to buy from?
And I keep hearing that there are hundreds of thousands of people working on iPhone assembly. I can't even imagine how long it would take to replicate such a workforce in the US. Wages aside, this seems to be rather exacting, meticulous work with limited paths of advancement. Where are these laborers going to come from?
As Steve Jobs is reported to have told Barack Obama in 2010: "These jobs are not coming back".
As Obama put it, “Those jobs aren’t coming back”.
While they make a gross margin of 46.3% overall, that’s mostly buoyed by the high margin on services, not on the products.
I think their products are genuinely competitive. It's not just marketing.
As far as I’m concerned nobody’s even trying to compete with Apple. There are other personal computing products but none of them address the same market as Apple at all. The tiny set of possible competitors for their space all seem to have decided not to even try.
[edit] and, to be clear, I’d very much prefer that they had competition in their segment. Even if I stuck with them, it’d at least put competitive pressure on them to improve and/or cut prices.
Integrated ecosystem is something a number of players have tried and failed at, as they simply don't have the resources to build it up and sustain it until they are competitive.
Apple has started small too: tiny little iPods with iTunes + digital content sales really put them back on the map, before they scored a hit with iPhone which was an iPod touch with a modem. They have been riding the iPhone wave since, gobbling up money with iPhone and iPad manufacture and app sales, which allowed them to expand to SoC production for laptops too, which is what got them back on the PC (as in "personal computer") map too.
I mean, that's not totally wrong, but by the release dates completely backwards. The iPod Touch is an iPhone without the modem.
But sure, it is a simplistic look: for the purposes of the commentary here, I did not want to pretend this is the full analysis, even if I could do one (and I probably couldn't).
I wasn't that impressed with Android long ago, but decided to give it another look last year when I helped a family member buy a top of the line phone.
Android is still not comparable.
Of course, it all depends on what the end user is looking for, but at least for overall quality of Apple phones is higher both in hardware and software.
Also, 200 years ago was the development of colonial empires, which is pretty much the opposite of what you say.