Arm wants a bigger slice of the chip business
economist.com
economist.com
Now that Apple and Qualcomm both design their own custom cores, Arm's custom core license business is suffering. The only other company that license their custom cores for phones is Mediatek. Arm still has hyperscalers in Google, AWS, Microsoft, and Meta for server core licenses though. But that might not last as long since everyone wants to differentiate and design their own cores. Apple and Qualcomm are succeeding in designing better custom cores than original Arm cores.
So Arm's business model has to change if they want to survive.
Either they make ISA license fees much higher (risk losing customers to RISCV) or they make their own CPUs and compete against their own customers even more.
I enjoy analyzing chip companies and businesses. I've never invested in Arm's stock because their customers are also their competitors. Ultimately, the questions for Arm investors are:
1. Do you believe they can design better cores than Apple and Qualcomm (and other big tech)?
2. Do you think that the threat of companies moving to RISC-V is real if Arm raises ISA royalties?
Very little software on android is written in assembly, and binary translation seems to be solved anyway.
An emulation and platform support dream. Run almost anything anywhere.
It'll probably be a dream
> I've never invested in Arm's stock because their customers are also their competitors.
Also applies to Nvidia. And there as been a lot of controversy related to exactly that reason. The most note-able being EVGA just throwing in the towel and sunsetting their entire GPU department. (And that predates the current LLM / AI boom)
I guess it's a showcase of a company leveraging this relationship
Can Nvidia design a better AI chip than their customers? Can AMD design a better server CPU than their customers?
Nvidia is proving they can. Arm has already lost the performance and efficiency crown to Apple and will likely lose soon to Qualcomm, if not already.
I think the GPU business is also fundamentally different than CPUs. For CPUs, it's all about interoperability. The ISA is the same. You can swap out any RAM easily. You can run Windows, iOS, Android, Linux. The only thing you're differentiating on is performance, efficiency, and the chip size.
For GPUs, everything is custom. There are far more ways to differentiate. Nvidia can differentiate through ISA, CUDA, chip design, cooling design, packaging design, rack design, networking chips, etc. Nvidia is more about selling systems than a chip.
Apple and Qualcomm don't sublicense so they are not really Arm competitors. Plus, ARM was always a convenient way to mutualise development costs without the risk of collusion. I am not sure it's changing.
You say people want to differentiate but Google constant failures to bring a state of the art offer with the Tensor shows that it's not that easy. Meanwhile Mediatek is competitive while remaining close to ARM design.
I don't think the market has changed that much on the high-end. The real question is how free will ARM be to export their future architectures. That's more where I see a risk.
That and the low-end being contested by Risk-V.
Apple and Qualcomm don't sublicense so they are not really Arm competitors. Plus, ARM was always a convenient way to mutualise development costs without the risk of collusion. I am not sure it's changing.
Arm makes far more money licensing their cores than their ISAs. If their entire business was reduced to just licensing their ISAs because their cores can't compete against their customers' custom cores, then they will be at a highly disadvantaged position.
The important things for Arm is not so much the few customers with competitive cores which are great marketing but the multitude of other customers who don't have the budget to design their own core. It's far from trivial as shown by the Google exemple.
If ARM goes down this path, it'll be a classic “innovator's dilemma” cycle leading to disruption.
ARM is clearly the superior incumbent technology, with a vast software base, billions of devices, extremely optimized designs, and huge amounts of engineering experience at the companies that use it; RISC-V is less capable on all counts.
And yet RISC-V is likely to eat ARM's lunch gradually—and then rapidly—over the next decade or so, due to being cheaper to build new cores on.
Define better. Because ARM is delivering better CPU Core and Value through their C1 and C2 without having its customer setting up its own R&D on CPU design. Not to mention the GPU and other IPs that comes with it.
> But that might not last as long since everyone wants to differentiate and design their own cores.
That is easier said than done. Had that been the case, the largest Hyperscaler of all, Amazon would have had their own CPU core on Graviton. Designing CPU Core is one thing, making them and integrating them properly to suit their own workload is an entirely separate effort.
>So Arm's business model has to change if they want to survive.
They can continue to do what they are doing today to survive perfectly fine. I assume you want change because you want them to grow.
Also, ARM's business model isn't just on Smartphone. There are billions of places if not trillions of places where ARM core exist.
After following that drama, it's difficult to see ARM as anything but a greedy profiteer.
Maybe they'll eventually make their own RV core designs too.
I am not a deeply technical embedded person, but I actually don't think that would be the death of ARM: my understanding is that they develop a lot of SoC-level interconnect/fabric standards and IP as well. After all, you have to do a lot of work to integrate your ARM cores into an actual platform...
You license ARM cores because you want a “custom” chip but do not want to start from scratch. You especially do not want to have to bootstrap a software ecosystem. When ARM had no competition, it is just a question of which ARM core you want.
Now, you can get the same thing from any RISC-V design house. Which means having real choice over the features you want. If ARM is just one of those RISC-V shops, how does ARM compete? By being the best? Not likely.
And, in the past, you could not totally outgrow ARM as they own the ISA. The Qualcomm lawsuit was an attempt to maintain tight control over this. With RISC-V, you can pack up and move your whole ecosystem elsewhere including taking it entirely in-house. This includes the ISA to an extent since anybody can add extensions.
Today, we are seeing RISC-V succeed where this flexibility matters most: in microcontrollers and in AI.
But as performance equalizes, volumes go up and costs come down, the use cases where ARM makes more sense dwindle.
That makes backwards compatibility the last real reason to use ARM. But does this matter on mobile where devices download the apps that match their arch? Not really. Does it matter in most embedded cases? Not really. Does it even matter in the server? More, but even there not as much as it used to. Does it matter for anything mostly GPU or NPU driven? No. So that leaves desktop and laptop. And, outside of Apple, ARM has not really built up anything to stay compatible with. RISC-V may have time to grow into that niche before being blocked.
We are going to exit 2026 with RISC-V chips that are fast enough. How fast will the costs come down? Perhaps a year or two?
What markets is ARM well positioned to continue its dominance in?
Welcome to Capitalism. It's not perfect but it's the best we've got.
Keep in mind that Apple is the other party in this interaction. Pot, Kettle, Black, etc.
TSMC should charge Apple a 30% fee to be in their FabStore.
TSMC already does make a huge amount of money off of being the indispensable fab for all of Apple's latest-and-greatest chips.
Each time TSMC launches a new or incrementally-improved manufacturing process, Apple's latest Ax SoC is nearly always the first chip manufactured on that process to reach consumers.
In 2 years from this date, I fully expect the safety critical RISC-V chips like the forthcoming High Performance Spaceflight Computer (HPSC) from Microchip, Inc.[0], and derivatives [1] [2] leveraging SiFive IP[3], and other RISC-V competitors [4] to take a dominant position in Space, Aerospace, Aviation, and potentially other less cost-sensitive industries where RTOS dominate.
This raises a few questions in my mind:
Could that extend to vehicles and other use-cases?
Will we see more derivatives with even higher performance beyond the already announced PolarFire2 designed specifically for terrestrial use?
I don't know how sensitive their overall BOMs are to high priced reliable chips designed for fault tolerance...
I don't know how fast the quality of the mass market chinese RISC-V chips will ascend the perceived quality gap, and expand offerings into newer profiles [5]
Where will the toolchain for RISC-V be on a specific chip basis?
Is Nvidia likely to expand their usage of RISC-V?
[0] https://www.microchip.com/en-us/products/microprocessors/64-...
[1] https://www.microchip.com/en-us/products/microprocessors/64-...
[2] https://www.microchip.com/en-us/products/microprocessors/64-...
[4] https://www.gaisler.com/products/gr765
[5] https://docs.riscv.org/reference/profiles/rva23/_attachments...
They already are in India [0][1], but SiFive helped build a portion of the ecosystem in India as well [2].
I'm sure there are similar applications in China but I'd need help pointing to a specific initiative. I know Tenstorrent is hitching their wagon to China especially after poaching Arm China leadership.
> Could that extend to vehicles and other use-cases
They already are in India [3][4][8] - this is something the US and Indian governments as well as American and Indian VCs and corporations are collaborating on together. One such collaboration has already IPOed [5], seen combat, and begun helping develop capacity within America [6]. And another has recently announced a mega-raise with General Catalyst [7]. RISC-V design is on the roadmap as well in this relationship.
[0] - https://www.iitm.ac.in/happenings/press-releases-and-coverag...
[1] - https://www.isro.gov.in/vikram3201.html
[2] - https://www.sifive.com/blog/sifive-expands-presence-in-india...
[3] - https://www.digitimes.com/news/a20260209VL216/risc-v-automot...
[4] - https://www.mindgrovetech.in/s2401-secure-iot
[5] - https://www.qualcommventures.com/insights/blog/ideaforge-fro...
[6] - https://firstbreach.com/news/first-breach-signs-jv-with-idea...
[7] - https://www.generalcatalyst.com/stories/our-investment-in-ra...
[8] - https://incoresemi.com/
Good to see that we're being noticed. But we all still need to deploy (in scale) to make anything worthwhile.
Keep on doing IPR focussed design.
Grow deeper roots into a foundry, work on integration of the tech into FPGA and big chip/platter stuff for AI.
If AI tanks, the work will find a value point. We all want more memory and more execution cycles per clock tick be they on one ALU or many.
Lots of work to come in optical related areas. ARM has green fields to dig, beyond the instruction set.
There was a lawsuit in 2020 or 2021 where some evidence was unsealed showing ARM gets 1% of the CPU cost. I can’t recall how that CPU cost was calculated - but I believe that was a part of their deal through the early 30’s. That’s less than a dollar per iPhone.
are there any estimates on how much apple is costed in server costs because of apple app store and if there are, can that be considered with the 30% that they are leeching and compare the two of them and compare that estimate with the retail example you brought.
So there are 2 million apps in app store. Assume average to be 100 MB so that is 2 million * 100 /(1024* 1024) = 190.734863 TB using $0.015 / GB-month Cf r2 with 0 egress costs I get the price as 2929.6128$ per month or 3 Grand/month ie. 36 Grand an year
Heck, consider the average app size to be FOUR times the cost. Then you get the salary of 1 apple engineer and that's STILL less than the average salary of 1 apple engineer.
https://www.glassdoor.com/Salary/Apple-Engineer-Salaries-E11...
And we are forgetting the 99$ something which apple app devs have to pay iirc at sometimes. I am an android guy and actually android has similar number of apps so the estimate is same actually.
But I would assume that the engineers aren't the bottleneck either. I assume that both apple/android stores aren't so complex to build to deserve 30% of transaction.
Some might be interesting about the malware/static-analysis but there was this hn post made recently about instagram blackhole which uncovered an malicious apple app which was available in app store so :/
The only reason one might think so is maybe considering the development cost of apple but I think that is an hardware decision and when we buy apple devices, it should be used to recoup the costs not trying to cash cow either. Its a trillion $ company trying to extract 3$ from a 10$ transaction you might do to anybody's patreon from the app :/ Yeah.
TLDR: No. I don't think a 30% transaction is justified.
Hope this helps.
a) Apple is getting really good at switching CPU architectures when needed.
b) Don't they already have a forever licence to the ARM ISA (since the 90s/Newton) as well as a substantial in-house design team? I guess the renegotiations are about future/roadmapped ARM archictural enhancements.
It would be sad if there was a substantial fork of the arm64 ISA.
Have a look at WCH. If you are used to the ST-HAL style you get that stuff running within a hour. They stuff works will fully opensourced compilers.
Additionally, a significant portion of Arm's China, US, and India engineering and product leadership has left to work on RISC-V startups and companies now.
That said, I can see Arm being leveraged by other Softbank owned companies like Ampere (which they already do) and Graphcore, with an eventual merger of all 3 into some form of a mega-corp due to operational overlaps and efficiencies, but this would be defensive in nature given the degree to which the industry has aligned with funding a RISC-V ecosystem and how RISC-V's governance and leadership consists of major players and leaders in the chip design space.
---
Edit: Can't reply
> It is also a bad look when they sue Qualcomm for selling chips in a way that Arm does not like.
That's why Qualcomm is also betting on RISC-V as well after acquiring Ventana [6] and is participating in India's DeepTech initiative [7], which has been targeting RISC-V startups as well as Renesas [8] in Japan+India taping out a 3nm RISC-V processor for automotive and IoT usecases. And also why FuriosaAI in SK has been working on RISC-V, as well as the multitude of fabless players in China.
It's the same thing that happened with IBM POWER vs x86 decades ago with an added sovereignty component.
---
Edit 2: After thinking some more, I think a case could be made for Arm to survive but not thrive in the same manner as Minitel continued to kick around for so long due to France's stress on technological sovereignty. Long term, I think RISC-V will eat a large portion of Arm's commodity and embedded computing market share, but Arm (and moreso Softbank) is attempting to position itself as critical to British [9], Malaysian [10] (they remain a major semiconductor hub), and even Indian [11] attempts at design sovereignty.
I can see a British-Japanese alignment around eventually merging Softbank properties like Arm, Graphcore, Ampere, and Rapidus into a British-Japanese version of Intel such that Graphcore+Ampere leverage Arm's ISA for HPC and Embedded/Telecom usecases respectively and Rapidus becomes their foundry.
Additionally, I can see the Japanese government pushing it's players to heavily leverage Arm as well - especially given that all the major players in Japan already cooperate, have an ownership stake in, or are partially owned by Softbank.
---
[0] - https://www.digitimes.com/news/a20260213PD208/arm-risc-v-com...
[1] - https://www.cas.cn/cm/202601/t20260126_5097208.shtml
[2] - https://www.digitimes.com/news/a20260216VL205.html
[3] - https://www.pib.gov.in/PressReleseDetailm.aspx?PRID=2224839&...
[4] - https://www.pib.gov.in/PressReleasePage.aspx?PRID=1820621&re...
[5] - https://m.blog.naver.com/nanambook/223316051806
[6] - https://www.qualcomm.com/news/releases/2025/12/qualcomm-acqu...
[7] - https://idtalliance.org/
[8] - https://www.digitimes.com/news/a20250923VL201/renesas-3nm-re...
[9] - https://www.gov.uk/government/publications/building-a-sovere...
[10] - https://newsroom.arm.com/blog/arm-malaysia-silicon-vision
[11] - https://www.digitimes.com/news/a20250918VL202/arm-design-chi...
I have read at that time the documents presented by both parties, but essential details of the contracts were missing from the public documents, so it was impossible to know which of Qualcomm and Arm was right.
Nevertheless, the argumentation presented by Qualcomm seemed far more plausible, unless it would have been contradicted by some of the redacted out contract details.
Arm has not shown in public any information that could have proven that they are right, so they, or their supporters, may not say that the judge has made a wrong decision.
Supposing that the decision was wrong, Arm has preferred to keep secret their arrangements with the customers, instead of proving that they were right, presumably because they might lose more money if other customers learned the exact details of the contract with Qualcomm, so they could request similar terms.
While I believe that it is right for Qualcomm to have won the trial, I strongly dislike the fact that Qualcomm now designs their own cores instead of licensing them from Arm.
The reason is that the Arm cores have excellent documentation, on par with that of the Intel-AMD CPUs, while the Qualcomm cores, like the Apple cores, have non-existent documentation. Moreover, Qualcomm is so silly that they have always obfuscated even what Arm cores they used in their older products. Whenever I evaluated some smartphone with Qualcomm SoC it was impossible to find any useful information on the Qualcomm site, but I had to go to various third parties to learn what is actually inside the Qualcomm SoC, to be able to compare it with alternatives.
So Arm requested increased license fees, which Qualcomm refused, claiming that the license contracts that both Qualcomm and Nuvia had remain valid, with their already negotiated license fees.
Orthogonal question for Dang or someone who knows -
Do downvotes, account flaggings, and/or high posting volumes trigger this? I run into it frequently whenever I get downvotes. I almost never used to get this before 2022 or so.
I've gotten into plenty of flamewars with Dems, Republicans, Anti-Vaxxers, Pro-Vaxxers, AI Luddites, AI Fundamentalists, China bots, China hawks, Apple fanatics, Apple haters, far-right, far-left, pro-WFH, anti-WFH, pro-immigration, anti-immigration, and others on HN.
I just don't care about filtering my opinions and use HN as a way to kvetch and impart some information I may know about.
I and a couple other long-time HN users have noticed a significant degradation in HN after a persistent boom in sign-ups began around 2022-23.
It's hard to answer lol.
10% of the stock is floated.
90% of the stock is owned by Masa who used it for collateral for his 18 billion loan for Stargate. THat is against 33 banks who have a strong incentivise to dump in a margin call situation.
Their revenues are circular for the last 4 years, with 30% growth purely coming from Softbank shuffling their own money.
They are gonna be the canary in the coal mine for when the AI bubble implodes.
Boots with uefi
Not all ARM use cases need that, but it would be a huge mistake to not develop integrated options.
And also an opportunity to make adjustments to their business model.
On ARM, every processor has its own bootloader, blobs needed for initialisation. Even the systems have different architecture. In the end, you need a special software setup, which is not supported more than a few years. See phones, Raspberry PIs and derivatives, Chromebooks.
This is because of a supporting set of standards (BIOS/UEFI/ACPI) that are well-supported on x86 systems, but technically independent of the x86 ISA. BIOS, and the general compatibility you're talking about, is a historical artefact of the IBM PC being so dominant in the market that other companies created compatible computers. UEFI and ACPI actually exist in the ARM ecosystem now. If ARM continues to grow outside of mobile devices, you could eventually see the kind of broad compatibility you're talking about. It's not super likely though. All signs point to the consumer computing ecosystem becoming more closed, rather than more open.
[0] https://developer.arm.com/community/arm-community-blogs/b/ar...
Considering the original ARM use case was a desktop-computer shaped thing with some degree of expandability, they had to solve the same problems that the PC did in bringup/enumeration/device abstraction. These should be solved problems.
At some point between Archimedes and iPhone, they lost this functionality. I assume there was a moment where they assumed that they were only doing SoCs with fixed peripherals and jettisoned all their knowledge and tooling in the space.
> ...jettisoned all their knowledge and tooling in the space.
ARM made a smart decision not to compete against Intel/AMD at the cutting edge, and instead completely dominated the low-power CPU market.
> These should be solved problems.
They are. The solutions have nothing to do with the ISA. The device discovery functionality you're talking about (ACPI/BIOS/etc) is provided by the motherboard, and (in the case of ACPI) made available to the OS by the bootloader. SoC's don't need sophisticated device discovery.
Quick question but How does Risc-V compare to this?
Paradoxically, Linux core maintainers prefer the ARM situation (as do RMS-grade FOSS fans). For them going x86 route means constantly getting blame for crappy code they didn't wrote. Not that I'm unsympathetic, but it really goes against users' interests. And again, BSDs and smaller OSes often simply doesn't have resources to support the myriads of platform hardware.
Where Arm is in markets that do demand compatibility (i.e. server) the standards like UEFI and ACPI are there and work. Where it's in markets like embedded, you still see the embedded profusion of different random stuff. Where other architectures are in the embedded market, you also see a wide range of different not very compatible hardware: look at riscv for an example.
IP-locked ISA is something nobody sane would want.
No, they capture exactly the value they create.
As a long time (40 years) subscriber to The Economist, I expect better of them.
Analogous to discussion 3 days ago
https://news.ycombinator.com/item?id=46994869#46995258
We're fast approaching 1 trillion Arm CPUs manufactured - because it was a good design for a good price.
No doubt unprecedented for something so complex to be produced in such volume. I predict that nothing else will ever achieve such scale.
I think in this case semantics matter.
But false about my meaning, which was precisely the opposite of "a euphemistic way to describe pushing producer surplus towards 100% and consumer surplus to 0%"
My original comment "They capture exactly the value they create" was my (failed) attempt to restate the basic premise of capitalism, that profits tend towards zero, using the words at play in this discussion.
See airlines who create enormous value, but remain terrible businesses. I think it remains to be seen if the AI companies can really capture the value being created as well.