US revokes Intel, Qualcomm licenses to sell chips to Huawei
bloomberg.com
bloomberg.com
Huawei is already banned from selling, why would they kick the dead horse so to say.
If anyone has any ideas, genuinely please let me know! I'm curious.
Ptychographic X-ray laminography has come a long way and can image even the latest chips at the transistor level:
https://spectrum.ieee.org/xray-tech-lays-chip-secrets-bare
True, it requires a synchrotron (one of the few things that actually costs more than a chip fab), but China has those:
https://en.wikipedia.org/wiki/Shanghai_Synchrotron_Radiation...
China might not be able to make bleeding-edge 2nm chips, but it has the equipment needed to expose any backdoors in them. Publicly, worldwide, with proof. That is an existential threat to any hardware manufacturer -- getting caught doing that is a bankruptcy-level event. Until recently they didn't have to be afraid of getting caught.
Being able to image chips at the gate level or better is a critical lever in fixing the balance of power between hardware manufacturers and their customers. It's why Bunnie Huang's work on IRIS is so important:
Things like the Intel ME are what make this arrangement work. The fact that the ME exists is public knowledge; there is a huge team of engineers at Intel who are responsible for designing it. The flaw that lets you tickle it in just the right way can be added by and known to only one or two people.
Just this year a spectacular example of one of these bugdoors was revealed in Apple's mobile chips, where an EL1 (kernel) attacker could tickle an undocumented L2 cache debugging mechanism to elevate to hypervisor/bootloader (EL2, EL3) level, and completely bypass Pointer Authentication:
https://media.ccc.de/v/37c3-11859-operation_triangulation_wh...
This hardware bugdoor (CVE-2023-38606) was discovered only because it was used against security researchers who were good enough to capture a trace of exactly what it was doing. Otherwise we would never know. We got very lucky here; there's no telling how many more of these things go undiscovered.
Intel ME ? Qualcomm blobs ?
So all this is doing is harming Intel and Qualcomm, while incentivizing Huawei to invest even more in domestic chips.
https://www.asiafinancial.com/china-providing-90-of-chips-us...
What sanctions ? You do realize that US jurisdiction is in US and some vasal states.
"Russia had been forced to repeatedly establish new supply chains for acquiring chips. As sanctions disrupted existing smuggling routes, it had set up new ones, likening these outcomes to a game of “whack-a-mole”, that was frustrating for both Western governments as well as “supply-chain managers at Russian missile and drone factories”.
The biggest failure in enforcing the controls is not that Russia continues to have some success in smuggling — that’s not a surprise — but that China continues shamelessly to sell Russia so much via normal trade routes."
Short term, it's a minor annoyance for China.
Yet the primary effect it will have is for China to become less dependent on foreign chips faster.
If companies can defend themselves against espionage and/or takeovers there can be a moat.
https://www.washingtonpost.com/news/worldviews/wp/2017/01/18...
Which BTW USA _still_ DOESN'T produce - not can't (distinction important) just like PRC didn't want (again, not can't) to produce tips until 2017.
Prior to 2017, only Switzerland, Japan made ballpoint tips.
Reason PRC ended up making ballpoint tips was premiere used tips as proxy for scaling up precision manufacturing of tungsten carbide for advanced munitions. Which they did in about a year, because PRC focusing state capacity pretty good at demonlishing moats. In this case they got state metallurgical factory (TISCO) with revenue with $15 billion to divert (waste) their massive resources to make some ballpoint pen tips, which had a global addressable market of 20M, aka rounding error for PRC state owed metal industry. Meme/rumor is TISCO made one batch of ballpoint pen tip metal and that's enough to supply domestic tip maker for decades. The _real_ story behind the pen tips is building up machines to make advanced munitions at scale, few years later, i.e. now - we have stories of PRC cruise missile gigafactory that can churni out components for 1000 missiles per day. A few days is sum US+partner production per year. For reference JP acquiring 400 tomahawks for 1.7B.
Woah. Who are they arming themselves up to fight with that many missiles? Trisolarians?!
Broader TLDR is telegraphed/public US strategy last few years for the next decade+ to deter PRC is to preposition more hardware, in distributed manner, make more survivable (agile basing, marine MLRS) etc. PRC retort is saying, they're going to make so much advanced munitions, that they can hit essentially any hardware US+co can possibly preposition in theatre that has to sit still for more than a couple hours, with 100,000s to spare to degrade other strategic targets. It's basically conventional MAD for any US partners in region, who are mostly import dependant islands, significantly more so than PRC who is still a continental sized land power with associated resources land connections. PRC hinting they can degrade their critical/energy infra, lock down their ports from importing energy/calories, basically turn them into cuba/gaza, indefinitely, from purely PRC domestic inputs. So think twice about allowing US to operate from their territory in 1st island chain in TW scenario.
They're probably onto something since this development is almost 6 month old and non of US/western strategic writing is openly discussing it, because there simply isn't a viable counter, short of assuming China can't make 100,000s of a thing, or Made in China things must be duds. As many are relearning, having industrial capacity to make magnitude more commodity fires > wunderwaffles. West loves to mention about US ship building prowess during WW2, and how US simply outproduced adversaries. Now we have stats like PRC having 230x advantage in military ship building. Just military, PRC peacetime shipbuilding just passed 160 million dead weight tons, which is more than ENTIRE US ship building program in WW2. Extrapolate that to cruise missiles. IMO the biggest strategic development in region in a long time.
This contrasts with the American inability to ramp up artillery shell production recently. Not enough money in artillery shells so the defense industry doesn't have capacity, they've been focusing on more profitable weapons.
Does any MBA care if they sell 100 x thing for $1000/thing or they sell 1000 x thing for $100/thing?
We're it to come to this, a full scale war with China, what actual benefit does leveling SK, Japan, Taiwan, and the Philippines do strategically?
Japan has a top 10 military these days, it's not like they will be able to occupy these areas real easily.
So we park our aircraft carriers a little further away - or use any number of military bases we already have in Asia in other regions.
We have declassified planes from the cold war era that can take off in Kansas, bomb anywhere in the world and land back in Kansas. Planes that fly so high that by the time the bombs hit the ground the planes are hundreds of miles away.
Until China can control the skies, this is all just kinda moot
I don't think it's enormous, I think it's probably a trivial amount relative to PRC productive capabilities / acquisition effectiveness. Benefit is PRC can potentially kick US out of East Asia, outside hedgemans don't abdicate unless their presence is demonstrated to be untenable. Rationale = PRC dragging US art5 tier partners in the region into broader war and showing everyone the US incapable of fulfilling security commitments in the PRC backyard even if the US wanted to. Incapable is key, only way to deter others in region from hedging with US secuirty if TW off the table is to demonstrate US simply cannot. Incidental benefits are also huge. PRC competes with SK/JP/TW on many fronts, especially in region, breaking their knees if they are actively involved in TW scenario = PRC has an opportunity to fill holes. Samsung phones can be replaced by Huawei around the world. JP infra investments can turn into PRC contracts. Killing 90% of high end semi that overwhelming benefits west = PRC close semi gap. Least of which = resolving ongoing maritime disputes. No occupation needed, they're islands, they can be turned into Cuba from the PRC mainland. There's only top2 mil and everyone else, JP doesn't have enough offense to really matter, realistically no one in the theatre is projected to except PRC. Meanwhile PRC is multiple times bigger = they have more targets/more resilient, and more labour/construction overcapacity to recover. Doesn't mean this is what PRC wants, but they're positioning MIC in a manner that they can / won't say no to escalation.
>park our aircraft carriers a little further away
Carriers are already parked much further away due to intermediate range PRC missiles, around 2IC in conflict. Spamming cruise missiles is about breaking the strategy of other US branches who has to operate in 1IC (airforce agile basing / marine mlrs / army missiles). Carriers were already being squeezed far enough that carrier aviation approaches irrelevance, ranges where they can't sortie to meaningfully deliver stand off weapons even with tanking. There's no other bases in the region that PRC can't reach, i.e. carrier aviation/escorts still need replenishment every few days, USN replenishment fleet anywhere in the region goes boom if they stay docked for more than 30m. Even further, other regions, we're really talking about Hawaii/3IC and beyond, and at that point the ability for USN to generate fires/sorties becomes negligible due to logistic constraints. For reference 5 carriers + regional air basing, operating unobstructed took 3 weeks to break Iraq, scale to PRC size, that’s 5 years of impossibly high tempo operation, assuming PRC can’t hit back at all.
>take off in Kansas
Everyone already knows what US long range strike options are, can extrapolate from B21 procurement what it will be. PRC answer to that is more missiles - conventional prompt global strike, which has entered their military strategy writing a few years ago, i.e. any carrier, sub, bomber that docks anywhere in the world, including CONUS goes boom in 60m. Also oil refineries, LNG plants, semi fabs, server farms, B21 plants etc etc.
"The US has long range strike options" argument misses the point. It's not about what the US always had/has/will have - ability to hit targets anywhere in the world - it's about how dynamic changes when US adversaries have what US has, when PRC can do to CONUS what US could always do to PRC mainland. And at what scale, JP still tried to fugu ballon CONUS during WW2 =/= PRC likely built an arsenal to hit every piece of strategic CONUS infra.
>control the skies
Controlling skies is a proxy for delivering effective fires. PRC betting advanced rocketry can skip the air control/supremacy layer with long range missiles designed to penetrate anywhere on earth via performance and volume. The thing with aviations/airframes is they can't be "classified" anymore, anything US is observable via space, because bombers still need wheels spin in the wild for testing. Versus PRC focusing on missiles, which can be manufactured/stored/launched from hardened facilities (look up PRC third front where they moved MIC into mountains), which IMO long term is much more resilient than bombers assembled/maintained in huge structures.
The reasoning here is that the US had 70 years to develop global logistics for carrier/bomber basing, because that was only technically feasible expeditionary hardware of the time. PRC couldn’t replicate that if she tried to (too much depends on uncertain geopolitics / alliances / access). But why bother if they can just skip all that and go straight for missiles that can hit broadside of a barn on the opposite side of earth. Arguably it would be cheaper/better, a lot of US force composition is arguably trapped in legacy inertia, i.e. Navy is not going to not carrier/sub, the air force is not going to not fighter/bomber - too much identity tied into platforms and money tied into procurement, and obligations tied into security commitments/alliances where US has to flex/visibly show presence with expensive hardware. Not to mention internal US defense drama killed earlier US efforts at prompt global strike, because why potentially cannibalize duties / undermine branch fiefdom. Meanwhile PRC/PLA modernization still new, not much sunk cost, is free to focus on rockets to break all those platforms at less cost. They don't need to FONAP, sail expensive platforms, fly bombers halfway around the world. That's a premium US maintaining hegemony has to pay for. Controlling skies in modern warfare = at the end of day is about delivering munitions to cripple war making capabilities. PLA started writing about PromptGlobalStrike in 2020 Science of Military Strategy, and acknowledged by the 2022 DoD report on PRC military. IMO we're looking at 2030 timeline where they get them in quantities western analysis would be forced to acknowledge implications. Incidentally at same time PRC also building up nuke program.
Arguably better than that US Ford thing, though slightly smaller. A 4th is already under construction, even larger, and will be nuclear powered.
All that effort just for showing 'W(h)e(e) too!' ?
Basically.
Would not discount prestige and desire to dick measure when every other large power has them. Technically on paper, carrier still good for projecting against non peer powers.
But looking at numbers, I would say relative to PRC ship building capabilities, they're not really investing in carriers. Adoption very slow: 2.5 carriers (with 60 J15s) in 15+ years with PRC's ship building is unserious. 2.5 (not 3) because Liaoning hull was rebuilt Varyag from UKR/USSR. For reference US was launching a carrier every year/other year post WW2, with comparable/larger displacement. Even considering PLAN starting from 0 for carrier ops, they're not waterhosing resources/bodies at problem, i.e. they were spamming submarines/subsurface fleet despite tech gap being even larger. IMO overall carrier investment pretty minimal/conservative, enough to build up institutions, training etc to test waters. Still slowly prototyping fc31/j35 and kj600 when already churning out 100+ J20s per year. Not saying they're not aiming for many more carriers groups eventually, but so far not urgently expanding when they could be laying down 2 per year. Nor are they zerg rushing big replenishment fleet to sustain operation outside theatre.
IMO, apart from glamour shots and training platform, I don't think anyone in PLAN seriously expects their carriers to survive vs US in full spectrum war. Or much of the surface fleet, but need grey hulls to navy stuff during peacetime, hence have many of them. I think gamble is making sure PLA has ability to ensure no USN surface ships survives either, which breaks US expeditionary model, meanwhile PRC likely will retain stupendous industrial capacity to rebuild fleet after war. And if they can sink USN with land based long range strikes, whose going to mess with PRC shipping in the interim.
https://upload.wikimedia.org/wikipedia/en/timeline/oj9bcx6vu...
Maybe having a few carriers gives them useful capability? Cruise missiles are pretty binary - you can't exactly fly the flag with them, and even shooting one past some target as a warning would be read as declaration of war. Having a mobile airfield you can park near someone else's water gives more flexibility in applying pressure without shooting.
No need to control skies and waters if you can land tens of thousands of missiles anywhere in the world.
That's a moot point if anti-air missiles shot down the plane before it even got to its target zone. We're not in the 80s anymore, radars are networked and missiles are much more capable.
However, it's entirely plausible is that it's a much trickier engineering problem than it sounds like, with a meaningful upfront cost (in both time and money) to solve, for a very small economic benefit at the end of the tunnel. And so there was no incentive for anybody to solve this in China because imports worked fine. Until they didn't - because a high ranking Chinese politician critiqued China's lack of a high quality domestic solution. And then, lo and behold, about a year later that domestic solution emerged.
It's not like there was some grand concerted effort to solve the 'pen problem.' It's just that trade was working perfectly fine, so there was no motivation to change. Give people a motivation to change or create, and they will. Like the old saying goes, 'necessity is the mother of all invention.' And right now these trade wars are giving the rest of the world every motivation to detach themselves from Western economies. It's not hard to see how this turns out.
I think an EUV machine litography might be harder to make if you need to invent it from scratch.
US still can't make high speed trains in 2024.
I assume at some point they'll make their own photolitography, but until then they seem to be getting what they need to produce their chips, one way or another.
That's not a given. US did to Japan something similar in the '80s, and they never recovered. Today, ASML is at the forefront of lithography, not Nikon and Canon.
Not quite. Japan was well underway with EUV lithography..
> Today, ASML is at the forefront of lithography
That's because it is so prohibitively expensive to develop. That, coupled with the fact that the machines are becoming prohibitively expensive to buy, indicates that it is probably a technological dead-end.
Which is why IMO PRC exploring SSMB /synchrotron EUV is so interesting.
IIRC EUV development picked plasma over synchrotron because plasma projected to be cheaper, even though technically synchrotron had more benefits. Queue many, many years of technical challenges and now commercialized EUV machines cost 200m, 400m for next high NA. Which is about the cost of multiple small or single medium size synchrotron facility in west. TLDR is it's amazing plasma EUV works, but it's also a failure in the sense that it is FAR less economical than originally envisioned. Economical dead end if PRC ever figures out cheaper alternative tools AND doesn't have to pay IP fees which they would have operating under western ecosystem.
This is not a solid advantage. Huawei has already proven a few years ago that they can design better chips than Qualcomm, which is when they have been immediately hit with sanctions, preventing them to manufacture their designs at TSMC.
So chip design is a solved problem for them. Replacing the chip design software tools from Synopsys, Cadence and the like is slightly more difficult, but it is still much easier than it seems. The incumbent position of those companies is maintained mostly because only they have access under NDA to the necessary information about the rules of the chip manufacturing processes used by the major foundries. There are no competitors for them mostly because of the absence of public information, not because those programs are particularly difficult to write.
If SMIC or any other Chinese company succeeds to have a competitive CMOS manufacturing process, there will be no difficulty for the likes of Huawei to develop appropriate chip design software tools for those processes.
The only real difficulties are in the physical chip manufacturing, especially in the EUV lithography. However the sanctions will force China to allocate more people and more money than anyone else for solving these problems. There is no doubt that eventually they will succeed. What is unknown is whether they will need 5 years or 15 years.
If China doesn't have access to foreign chips and/or foreign lithography technologies, they will have to develop their own methods and probably diverge towards different technical solutions. And that's always good for progress.
Why would they, until forced? Comparative advantage is very tempting; most of the world avoids making things they can instead import cheaper.
Because they were trying for years and failing? It’s not at all clear that the 7nm chips they produced recently have high yield and low cost compared to their competitors which is crucial so it’s not clear that they’ve yet succeeded.
they were already playing dirty and weren't top customers for those chip makers, regardless.
Honest question. If you could sell to somebody(and make profits) it would be foolish to create a situation where you might have to compete with them for the same thing(and now make losses)?
The US kills off Huawei’s access to Google Android and TSMC. Almost completely killing their consumer business.
But then hands them a lifeline: you can buy chips at Qualcomm (but not Mediatek) and Intel (but not AMD). Allowing them to maintain a presence in consumer phones and laptops.
Now 5 years later when they have finally built their own operating system and setup their own fabs - the US kills access to Intel and Qualcomm.
Why?
Having one less option is never a boon.
One the other hand this is good for consumers. For a few years we used to make fun of HarmonyOS being a knockoff of Android, but we have now reached a point where they are due to cut Android compatibility out. They have a new polyglot UI stack that seems quite performant. Basically we now have a 3rd big smartphone mobile operating system that is about to hit global markets.
However a few years ago Huawei started locking their bootloaders, so I have to wonder how much of the UI stack is actually going to be open source.
I don't know too much about it but it seems to have some similarities to react native. Raycast changed my opinion on react native based apps on macos, and I think linux could benefit from such a thing, although I have only little hope that ArkUI will run on vanilla Linux.
The present HarmonyOS is a placeholder for that.
Android is based on Linux (UNIX clone, 60s), IOS is based on Mach (80s design).
HarmonyOS is a contemporary design, following current state of the art in system architecture, far more advanced than these two systems.
They are, ironically, ahead.
EDIT: I just looked it up, that's awesome. It's pretty amazing that everyone was bashing Huawei and Harmony for so long, but they finally managed to pull it off.
Do you have any reference that breaks down the system architecture into more details than whats in the wikipedia article?
Completely different from that of China. US big tech are own by individual organizations out of US government and cooperate with it. Huawei benefits people from CCP directly.
After that we will see a flood of lower cost hardware coming to the rest of the world, CPUs, GPUs, AI chips. That will put pressure on Intel, Qualcomm, Nvidia and the likes.
I'd invest in US tech companies for the next few years and after that move to Asian companies.
Every major western semi partner currently projecting talent shortage of 100,000s in 5-10+ year time frame. VS PRC at ~500k/700k of domestic IC talent they think they need for indigenous industry according to 2018 white paper (IIRC same year they started prioritizing talent production to first-level discipline), and now academic system is generating ~30k IC talent per year. TLDR is everyone maybe making improvements, but PRC uniquely positioned to have requisit talent to catchup/exceed, and have advantage of coordinating entire semi industrial base in single jurisdiction. Semi talent development is still one area where west is actively dropping ball, and unless TSMC Arizona sort out their worker/cultural drama, might continue to if west can't elevate semi to desirable career.
Europe doesn't really have much domestic production in phones that could have benefitted from eliminating Huawei. Nokia is all but dead, Siemens is long dead, Sony Ericsson is now just Sony and on life support... it's all Apple, Samsung and various Chinese nowadays.
Apple doesn't benefit from banning Huawei, they're targeting completely different markets. Samsung is Korean. And Google (its Pixel line aside, which has an utterly negligible marketshare and effectively only exists to showcase bare-bones Android) doesn't care who is bringing in the Play Store bucks, for them it's no difference if it comes from Samsung, Huawei, Xiaomi, Oppo, Oneplus or whatever else.
Immediately before the sanctions, Huawei had demonstrated better smartphone chipsets than those of Qualcomm, which Huawei could no longer manufacture after losing access to TSMC.
Most of the smartphones that have replaced Huawei have used Qualcomm chipsets, regardless of their vendor.
(Also Huawei was massive competition for Cisco and Nokia 5G equipment - as with phones, the equipment was recent - from claims of people that worked with it - and significantly cheaper than the western counterparts. Similar to phones.)
Parity trade barriers back in the 90’s probably would have avoided this whole mess.
> Huawei doesn’t rank in Qualcomm’s list of top 10 customers, according to Bloomberg supply chain analysis. It also doesn’t feature in Intel’s list of top customers.
This is how Year of Desktop Linux will finally happen, eventually forks will be required to work around export restrictions of the two US OS vendors, running on top of ARM and RISC-V units produced by non-US companies.
I heard an interesting point that China will not escape the middle income trap, they will however just rug pull the developed nations straight back into middle income or even lower income.
How would they do that?
So imaging being a non western company, are you really going to buy the more expensive western machinery or are you going to buy cheaper and probably better quality Chinese machinery? This probably means this will cut massively in western companies revenue and profits, making paying off or getting future loans harder etc.
It will not happen tomorrow but i think this process will be done before 2050. I don't even think its an explicit Chinese leadership goal. Its just the result of China being the biggest and most competitive manufacturing market on the planet. So its mostly market forces doing their thing within China's consumer good and private sectors.
Actual article: https://www.bloomberg.com/news/articles/2024-05-07/us-revoke...
No chance of Huawei beating the likes of Nvidia, Intel, Apple in the next few decades at least.
So it's not like the West is sitting around doing nothing.
Also China has started investing heavily in domestic AI chips. Huawei's Ascend is roughly comparable to the A100 I think, but you shouldn't forget that this has caused a flurry of GPU as a service company to pop up outside of China that serve GPU compute to Chinese customers.
The best this kind of legislation does is make it a bit more expensive for China to access GPU compute. It's worth remembering that a lot of tech was banned from Iran, but that never stopped Iranians from getting any of those components. It just made them twice as expensive.
https://www.trendforce.com/presscenter/news/20231211-11957.h...
It might not be the only way to get things to work, though clearly I have no real idea. :)
At least cpu's haven't been adhering to Moore's Law for a few years now.
That being said foundries are a hard business, there is a reason there has been so much concentration and companies throwing the towel.
If their government wills an independent and competitive semiconductor industry, then it will happen sooner than we think.
Hell, China will develop commercially viable fusion power if their government wills it.
The US semiconductor companies need Huawei more than Huawei needs the US semi companies.
Say goodbye to those companies while you can. They won't be with us for very much longer. You don't last very long when you can't sell to the World's biggest market.
even china bull Ray Dalio thinks China in a lost decade or more https://www.forbes.com/sites/williampesek/2024/03/29/ray-dal...
5% isn’t going to cut it.
It has a massive overhang of real estate and big demographic headwinds.
The middle income trap is looking more and more a possibility.
China doesn't need to reach the same GDP/capita as the West in order to attain the same level of economic power. Even with only 50% of Western GDP/capita, the Chinese economy would as large as the US and EU combined.
> The middle income trap is looking more and more a possibility.
The "middle income trap" is related to an inability to move up the value chain. Given that China already leads in many growing high-tech sectors, it already looks like China has blown past the middle income trap.
China has absolutely not blow past the middle income trap. 600M live on less than $140/month according to Chinese statistics.
If they hope to get anywhere close to the standard of living of developed countries they will need to quadruple the size of their economy which even at 9% would take 2 decades.
The middle income trap is about countries getting stuck, because they are unable to transition from cheap manufacturing to higher-value-added industries. China has already moved into high-tech sectors in a big way, so it looks like the middle income trap is already irrelevant for China.
Why is it supposed to fall apart exactly?
That's an irreversible problem that means we are essentially seeing peak China now.
As night follows day you’ll either have to accept India as an equal or risk loosing that market as well.
Nicely done. Russia did a similar thing a while back, so when visa and mastercard had to kill their business in Russia all of the credit cards inside the country remained undisrupted. I was skeptical back in the day over the whole engineering and integration of NSPK (national system for payment cards), but seeing it doing its thing was an awe.