Hacker News new | past | comments | ask | show | jobs | submit login
Nvidia to Acquire Arm for $40B (nvidia.com)
2120 points by czr 5 days ago | hide | past | favorite | 749 comments





This is terrible. Not really just because of Nvidia - which has a lot of problems I've previously commented on the rumors of this [1] - but Nvidia's ownership completely changes ARM's incentives.

ARM created a business model for itself where they had to act as a "BDFL" for the ARM architecture and IP. They made an architecture, CPU designs, and GPU designs for others. They had no stake in the chip making game, and they had others - Samsung, Apple, Nvidia, Qualcomm, Huawei, Mediatek, Rockchip and loads of others make the chip. Their business model was to make the ARM ecosystem accessible for as many companies as possible, so they could sell as many licenses as possible. In that way, ARM's business model enabled a very diverse and thriving ARM market. I think this is the sole reason we see ARM eating the chip world today.

This business model would continue to work perfectly fine as a privately held company, or being owned by a faceless investor company that wants you to make as much money as possible. But it's not fine if you are owned by a company that wants to use you to control their own position in the chip market. There is no way Nvidia (any other chip company, but as laid out previously Nvidia might even be more concerning) will spend 40 billion on this without them deliberately or inadvertently destroying ARM's open CPU and GPU ecosystem. Will Nvidia allow selling ARM licenses to competitors of Nvidia's business? Will Nvidia reserve ARM's best IP as a selling point for its own chips? Will Nvidia allow Mali to continue existing? Any innovations ARM made previously it sold to anyone mostly indiscriminatorily (outside of legal restrictions), but now every time the question must be asked "does Nvidia have a better propietary purpose for this?". For any ARM chip maker the situation will be that Nvidia is both your ruthless competitor, but it also sells you the IP you need to build your chips.

EDIT: ARM's interests up to last week were to create and empower as many competitors for Nvidia as possible. They were good at that and was the root of the success of the ARM ecosystem. That incentive is completely gone now.

Unless Nvidia leaves ARM alone (and why would they spend $40B on that??), this has got to be the beginning of the end of ARM's golden age.

[1]: https://news.ycombinator.com/item?id=24010821


Precisely, plus just consider the information that Nvidia will have on all its competitors who use Arm IP.

- It will know of their product plans (as they will need to buy licenses for new products).

- It will know their sales volumes by product (as they will need to pay fees for each Arm CPU sold).

- If they need technical help from Arm in designing a new SoC then the details of that engagement will be available to Nvidia.

How does this not give Nvidia an completely unfair advantage?


I wouldn't use the term "unfair" here. There's also just three x86 licensees in the world and people don't usually consider that an affront. You buy, you control, that's how the world works.

But I do think it's important that we recognize that we're going from a position of tremendous competitiveness to a much less competitive situation. And that will be a situation where ARM will be tightly controlled and much less inducive to the innovation we've seen in the last years.


> There's also just three x86 licensees in the world and people don't usually consider that an affront.

Hi! Counterexample here.


ME TO!

Especially because of the x86 oligopoly I would think that Arm is so much more important as an ecosystem.


The three licensees would be Intel, AMD and VIA

VIA doesn't have a license for the AMD64 instruction set, however. Intel and AMD did a cross-licensing deal so they have a co-equal position.

Which hasn't kept them from building 64-bit cores with everything including AVX-512..

https://fuse.wikichip.org/news/3099/centaur-unveils-its-new-...

IIRC you watch the "Rise of the Centaur" documentary they talk about the Intel lawsuit, and the corresponding counter suit that they won. Which makes the whole thing sound like MAD.

More interesting there is https://en.wikichip.org/wiki/zhaoxin/kaixian


Why doesn't someone like, say, TSMC acquire VIA in a deal where M&A experts do all their jugglery and cunning to make the licensing stay I wonder?

Because TSMC can't compete with their customers and of them is AMD, I believe. TSMC promised not to build their chips, AFAIK, for exactly non-competition reason.

VIA isn't privy to the cross-licensing agreement Intel/AMD signed which gives them full access to the other's technology portfolio (since AMD was a second source foundry for Intel, while Cyrix/VIA was not). However, the Court ruled (when they were still Cyrix) that they are allowed to sell any clean-room designed x86 hardware. So while they don't have access to internal architectural documents, they're allowed to reverse-engineer/duplicate x86.

In some ways, it actually frees them up a bit; as they don't have to reciprocate any efforts with Intel. Which they tried to leverage with their Padlock technology. Unfortunately, their marketshare limits any real practical usage of those benefits.


> they are allowed to sell any clean-room designed x86 hardware

Uuuuh, presumably the case was about patents, right? I don't see how cleanroom-ing is fine with regard to patents.


Source: https://law.justia.com/cases/federal/district-courts/FSupp/8...

Summary: Cyrix (and subsequently, VIA) have an implied license to Intel's patents; allowing them to develop x86 hardware.

How that differs from AMD: AMD and Intel have a full cross-license on technologies. This means, AMD can utilize Intel resources to integrate AVX-512, a ring-bus, etc (and vice versa). VIA can not. They must develop those technologies independently, in a compatible manner.


That case was about patents, but I'm not sure patents are as central here. Right now, there is a patent minescape of mutually assured destruction between everyone. If anyone exercises their patent rights, no one can make anything modern, and the whole industry would come screeching to a halt.

Oddly enough, integrated circuits have their own IP scheme. IP has copyright, patent right, trade secrets, trademarks, and mask works. You rarely learn about mask works, since they're so narrow.

https://en.wikipedia.org/wiki/Integrated_circuit_layout_desi...

I think a lot of this would hinge on that corner of the law.


Thankfully those patents are about to expire.

The original patents have expired.

The problem with waiting for the patents to expire is you're always stuck 20 years behind.


The ones necessary for x86-64, I mean.

Being 20 years behind isn't a particularly big deal if those 20 years are basically just SSE3, SSE4, AVX.


That's a good point. I guess it is "unfair" but that isn't necessarily an argument against it - as you say lots of things are unfair.

But, given the high market share of Arm in several markets allowing one firm the ability to use that market share to gain competitive advantage in related markets seems to me to be deeply problematic.


Noob question: how does the x86 licensing work? Does Intel still own the rights? Why would they license to AMD? Why don’t they license to others?

Very briefly:

- They don't license because they can make a lot more money manufacturing the chips themselves.

- AMD also has the right to x86 because Intel originally allowed them to build x86 compatible chips (some customers insisted on a 'second source' for cpus) and following legal action and settlements between the two companies over the years there is now a comprehensive cross licensing agreement in place. [1]

- Note that AMD actually designed the 64 bit version x86 that is used in most laptops / desktops and servers these days.

[1] https://www.kitguru.net/components/cpu/anton-shilov/amd-clar...


> some customers insisted on a 'second source' for cpus

IBM required it. It was their business MO to protect themselves from losing access to a technology or having the market be dictated by one company. Intel acquiesced so that they would be the architecture of the IBM PC-series.


Afaik they could implement the ISA but not the Intel interconnect bus so AMD licensed Hypertransport (as used in the Alpha) from DEC so you had different motherboards/chipsets.

All second source AMD (and other manufacturer) chips were pin-compatible with Intel chips. Compare the i386/am386, the i486/am486, etc. Even a few non-second source (e.g. the unique uArchs that came later) designs were still pin-compatible, such as the 5x86 and K6/K6-II/K6-III. In fact, the third parties were the first to break away from electrical compatibility with their Super Socket 7 motherboards.

Intel did change sockets as a means to disallow socket-compatibility; forcing consumers into their architecture if they bought their motherboard, but that had no effect on AMD's development. AMD had purchased a significant share of DEC's engineering portfolio and, along with it, their employees. Those employees then developed the K7 (Athlon) architecture around some of the Alpha's technological advantages, which included HyperTransport, a multi-issue FPU (fixing one of the major issues AMD had struggled with and bringing them ahead of Intel), etc.


Those cross licenses are the only reason why this laptop isn't running on Itanium, without them there wouldn't exist AMD64.

The 'only' reasons (plural) that your laptop does not run Itanium: - battery life: an Itanium in a laptop would run for less than an your - too hot: Itanium servers ran incredibly hot due to power. A laptop would probably catch fire in short order - Windows was not supported on Itanium - cost: would you pay $5K for an Itanium laptop?

some customers? I heard it was IBM.

In simplest terms, the AMD & Intel perpetual license goes all the way back to the early 80s and Intel has tried to legally harangue their way out of it ever since, with limited success.

Not really. Intel hasn't shown any legal or business moves to try to cancel it ever since they developed EM64T and integrated it in their entire product line. AMD existing protects them from antitrust considerations. They would just prefer AMD stay in the budget realm so they can continue selling high margin products (enthusiast, laptop and server chips) without competition.

Not that they would complain if AMD lost their license or ceased existing, they just don't seem to actively be trying to cancel the license at this point.


> You buy, you control, that's how the world works.

Only in countries with poor regulators like the states does it work like this.


No, it works like that all across the world in fact.

It works like that in Japan: say hello to the many giant conglomerates that rule their economy top to bottom.

It works like that in South Korea: say hello to Samsung, roughly 12% of South Korea's GDP in a given year (Walmart by contrast is equal to 2.5% of US GDP, and that's crazy big).

It works like that in Germany: say hello to a parade of big old industrial giants that have dominated their economy for most of the past century and will continue to.

It works like that in China, openly so: they intentionally go out of their way to promote giant national champions at the expense of everyone else.

It works like that in France: their largest corporations and richest individuals are even larger in relation to their economy and national wealth than they are in the US (say hello to Arnault, Bettencourt, Pinault and the Wertheimers).

It works like that in Russia: say hello to the countless, directly state protected oligarchs. Threaten their interests, you die. Their approach is super simple.

It works like that in Italy and Spain, which are both dominated by old, large corporations and family interests. Which heavily explains their forever economic stagnation.

It even works like that in Switzerland: ever see how large their financial companies are in relation to the economy? Who do you think actually runs Switzerland? Their banks are comically outsized versus the size of the economy.

It completely works like that in Brazil and India.

It works like that across all of the Middle East, to a much greater degree than most anywhere else.

It works like that in second tier economies, including: Poland, Mexico, Argentina, Romania, Turkey, Thailand.

It works like that in poor countries, including: Vietnam, Indonesia, Philippines, Ukraine, South Africa, Pakistan, Bangladesh.


Hey now, don't bring Pakistan in to this. We are owned by the military, not some puny corporation. Our biggest industrialists/companies barely hit a few billion dollars in yearly revenue, and they do it by keeping their head down and staying quiet.

Sales of important companies have to be approved by regulators, e.g. "Reuters: EU regulators approve ArcelorMittal to buy Italian peer Ilva".

https://www.reuters.com/article/us-ilva-m-a-arcelormitta-eu/...


Yes. But are they successful. It is AMD that kick intel not the other way round both architecture x86 64 bit) and now through working with others (Taiwan which is not in your list) to break their hardware.

Might and reality is not right.

We need competition. We adore competition. And even the countries you quote many do have competition. You just do not innovate and of course one can try to stop and rest. But the works does not.

Be water not mountain my friend.


Without AMD I could be enjoying Itanium nowadays.

Itanic was sinking from the start, it wouldn't be unlikely that Intel would've invented something like amd64 just a few years later if AMD didn't do the smart thing.

You can pick any colour as long as it is black.

American company buys a British company from a Japanese company.

How exactly is this only like in the states?


I dont even think this is about CPUs. This purchase is consolidation of two GPU companies. I've said it before, as Risc-V further commoditizes CPUs, the differentiator will be who has graphics. In that light, this is pure consolidation.

I have a lot of respect for RISC-V but I really struggle with an unqualified statement that RISC-V "is" commoditising CPUs in a meaningful way when there are basically no mobile, desktop or server chipsets with a meaningful presence in the market. All that may change over the next 10 years but who knows? For now x86 is dominant on servers and Arm in mobile.

In reality Arm commoditised in many markets CPUs by making reasonable designs available to all at reasonable cost, keeping control over the ISA and allowing firms to innovate in their implementations. You can have the same code running on a Raspberry PI, an iPhone and a 64 core Graviton2 server.

The Nvidia takeover threatens all this by giving control to a firm who could well 'unlevel' the playing field and even refuse to offer the latest IP to competitors.

RISC-V may provide a way out for firms unhappy with Nvidia but it could be a bumpy path. And its certainly not the case that Nvidia are paying $40bn to consolidate Mali with their own graphics IP.


Arm's Mali GPUs don't seem very important in the grand scheme of things. They're the choice for SoC makers who don't have anything else. Qualcomm has their own Adreno, Apple has their own thing, NXP buys GPUs from Vivante, Samsung is now going to include AMD Radeons in phones..

Sure but you have to assume that at some point in the future that Nvidia is going to start making those non-Nvidia GPUs very difficult/expensive to use in future ARM designs. They could do this legally, technically through manipulating the reference designs to tightly integrate Nvidia GPUs, cutting support for third party GPU integrating SoCs or any of the above.

> You buy, you control, that's how the world works

NO! That is how the world does NOT work! We seen it trillion times, I am disappointed the least seeing this negligent 'argument' for blocking dominance! Shame!


How does the world work then?

To play devil’s advocate a bit, are nVidia's incentives necessarily so different? Their goal will be to make as much money as possible, and it's clear that licensing has been a winning strategy for ARM.

Samsung comes to mind as another company that makes their own TVs, phones, SSDs, ect., but is also perfectly happy to license the underlying screens and chips in those products to other companies. From my vantage point, the setup seems to be working well?


It could be Nvidia just wants ARM's profits and leave them alone, but I don't understand why Nvidia would spend 40 billion dollars on that. They spend 40 billion on the control of a company, why would they do that if they just would leave them be? Surely they want to exercise that control in some way for their own goals. Especially a company like Nvidia which (in my linked comment) has a proven track record of not understanding how to collaborate with others.

EDIT: Let's be clear that ARM's incentive last week was to create and empower as many competitors for Nvidia as possible. They were good at that and was the root of the success of the ARM ecosystem. That incentive is completely gone now.

I'm guessing Samsung has a track record where I'd feel a little more confidence in the situation if they'd taken over ARM here, but in general ARM's sale to Softbank and thereby its exposure to lesser competitive interests has been terrible. They could have remained a private company.


> They spend 40 billion on the control of a company, why would they do that if they just would leave them be?

Well, the optimistic reason would be talent-share. nVidia has a lot of chip designers, and ARM has a lot of chip designers, and having all of them under one organization where they can share discoveries, research and ideas could benefit all of nVidia's products.


All I can say to that is that I don't share your optimism. Explaining this as an acquihire sounds nice, but that seems a thin explanation of this purchase given the key role ARM has for many of Nvidia's competitors and the ludicrous amount of money Nvidia put on the table here. We'll see what the future holds - I certainly hope the open ecosystem can survive.

I don't see it as much as Nvidia trying to strangle the competition that licenses ARM as much as I see it Nvidia having full control over the development of the architecture. This would enable them to own full solutions in their AI part of the business and tailor the architecture to their very specific needs.

Buying ARM not only allows them to control the direction of development, it also protects them should anyone else have bought it with hostile intent.

Personally I see this as more of an attack on Intel.


I agree actually that Nvidia's strategic motivation here is their long-term position w.r.t. Intel and AMD, maybe even Apple/Huawei/Samsung. And that's actually a good reason for this investment!

I likewise don't think destroying the ARM ecosystem is Nvidia's primary objective here - far from it. It might not even be a secondary objective. But I do think the ARM ecosystem will be slowly torn apart either as an innocent bystander or as a less prominent secondary objective (given real-world complexities, probably some combination of both). When Nvidia buys ARM, and they explicitly buy them to improve their competitive position, there's only one direction where I see the incentives going. Those will be against the open ARM ecosystem, which is a breeding ground for big and small competition against Nvidia.

Even if Nvidia's primary motives are relatively benign, I think they'll inevitably create a sitation where the ARM ecosystem can't continue existing in its current form. That's where the real tragedy will be.


Or to make sure they can't be left in the dirt by somebody else. If Apple, for example, would have bought ARM, I could imagine them quite hostile to NVIDIA.

How do we know that Samsung hasn't stifled some potential competitors by refusing to sell them screens or by selling them an inferior product?

Samsung makes (excellent) screens for iPhones, which are huge competitors to Samsung's own flagship phones, but Samsung still seems happy to take the profits from the screen sales. If there are smaller potential competitors that Samsung won't work with it's most likely because the scale is too small to be in their economic interest, not because they're rejecting profits in order to stifle potential competitors.

iPhones sell massively well though, to the point where it would be a big loss if Apple went to another company for their screens. You just can't screw with a company the size of Apple without consequences. It's not like they'd go "oh no, no more iPhone now I guess :'(" if Samsung decided not to sell them screens anymore.

The problem is more with smaller companies that could be destroyed before they even get a chance to compete. Those can be bullied pretty easily by a company the size of Samsung.


I don’t really see how it works both ways. It’s hard to imagine a bigger scarier competitor to Samsung than the Apple iPhone, yet we agree Samsung is happy selling screens to Apple.

I recall reading a breakdown at one point a few years ago that indicated that between screen, fabs (before Apple went for TSMC) and memory, Samsung made more money from each iphone sold than they did for each galaxy S phone of the time. Seems like a winning deal to me.

Because there are a lot more Androids and other devices that need screens. If they didn't make screens for Apple, a competitor would.

Apple is a much bigger company than Samsung, they can't realistically turn down a contract that big when LG is lying in wait to take all that money.

Not really. Revenues and Profits of Samsung group and Apple are roughly comparable and in many ways Samsung's revenue streams are more diversified and sustainable.. stock valuation notwithstanding.

"stock valuation notwithstanding" Perhaps you were thinking of Samsung Electronics, not Samsung Group.

There isn't really much outside Samsung Electronics in terms of revenue/profit.

Samsung SDS made $635 million in profit in 2019 which I think was higher than ARM? As far as I'm aware the 'Samsung Group' doesn't actually have all the companies under that so it's difficult to see what the overall revenue/profit is of the entire 'group'. It has quite a few other companies, like heavy industries, insurance, asset management, etc and I think they all have thousands of employees and are very large companies in their industries so I'd be surprised that when combined the other companies weren't significant. Did you have any data on this?

That is the key. There are and need to have alternatives. Real one.

Not the one to rule them all is the key to every innovation.


Samsung is a massive company ... the division that handles screens is managed independently from the one that makes smartphones.

It must be probably weird for Apple technical staff to communicate so closely with one division of Samsung while fighting for market share with another division of the same company.


> Samsung makes (excellent) screens for iPhones, which are huge competitors to Samsung's own flagship phones, but Samsung still seems happy to take the profits from the screen sales.

What markup does Apple pay for Samsung OLED displays compared to Samsung’s other OLED customers? I think this is highly relevant if you want to use it as an example. Because if the markup for Apple is 5x that of other buyers of Samsung OLED displays then you certainly can’t say Samsung is “happy” to sell them to Apple.

Same for nVidia-owned-ARM: if they’re happy to sell ARM licenses at 5x the previous price, then that will surely increase sales for nVidia’s own chips. I guess my overall point is: a sufficiently high asking price is equivalent to a refusal to sell.


demanding information that you know nobody will be able to produce is an unethical debating tactic.

obviously nobody but Samsung and their customers will know that information, and anyone who could reveal it is under NDA.

Apparently the prices are good enough that Apple doesn't go elsewhere.


Resources are limited. If a Samsung phone and a Motorola phone need the same screen and there's not enough to go around, what happens?

A bidding war of course.

On the surface, it's capitalism at work. In reality, Samsung winds up in a no-lose situation. If Motorola wins, Samsung gets bigger margins due to the battle. If Samsung wins, they play "pass around the money" with their accountants, but their only actual costs are those of production.

I'd note that chaebol wouldn't exist in a free market. They rely on corruption of the Korean government.


Screen manufacture is a highly competitive market. ARM licensing isn't.

The main difference I'd say is that ARM holds more of a potentially captive industry that is totally intertwined with its IP. Most of the products that Samsung sell to other industries are more or less fungible with products that other third parties sell but the ARM design rights are going to be incredibly difficult to replace. Entire ecosystems of electronics and software are built on it.

There's a whole lot of inertia for Nvidia to take advantage of here while the rest of the industry figures out where it's going.


The reason that ARM is only worth 30-40 billion while powering nearly all smartphones and much of the embedded space is because their business model is not that profitable. If Nvidia is looking for profits, ARM is a bad investment. If they are looking for synergies to boost nVidia profits, it will be done at the expense of other licensees. That's exactly what people fear.

Didn't Samsung stopping selling flash chips to ouside parties when it suited them?

I tend to agree, but there may be another angle to this which could prove beneficial to consumers. Right now, the only ARM chips which are actually competitive with desktop chips are from Apple, and are obviously very proprietary. If this acquisition enables Nvidia to begin producing ARM chips at the same level as Apple (somehow, who's to say how, that's on them) then that would help disrupt the AMD/Intel duopoly on Windows. Its been a decade; Qualcomm has had the time to try and compete here, and has failed miserably.

I doubt Nvidia would substantially disrupt or cancel licensing to the many third-rate chip designers you listed. But, if they can leverage this acquisition to build Windows/Linux CPUs that can actually compete with AMD and Intel, that would be a win for consumers. And Nvidia has shown interest in this in the past.

Yes, its a massive disruption to the status quo. But it may be a good one for consumers.


But Nvidia has an Arm architecture license already - the same as Apple - so can build Arm chips to whatever design it wants (and it does in Tegra).

This is nothing to do with extending Nvidia's ability to use Arm IP in its own products.


Nvidia is a total powerhouse when it comes to chip design. Most people here are looking at this from the angle of "how does ARM benefit Nvidia", but I think its more valuable to consider "how does Nvidia benefit ARM". In 2020, given what we know about Ampere, I really don't think there's another company out there with better expertise in microprocessor design (but, to be fair, lets say top 3 next to Apple and AMD). Now, they have more of the stack in-house, which may help produce better chips.

Yes, ARM mostly just does licensing, but it may turn out that this acquisition gives Nvidia positive influence over future ISA and fundamental design changes which emerge from their own experience building microprocessors.

Maybe that just benefits Nvidia, or maybe all of their licenses; I don't know. But, I think the high price of this acquisition should signal that Nvidia wants ARM for more than just collecting royalties (or, jesus, the people here who think they're going to cancel the licenses or something, that's a wild prediction).

The other important point is Mali, which has a very obvious and natural synergy with Nvidia's wheelhouse. Another example of Nvidia making ARM better; Nvidia is the leader in graphics, this is no argument, so their ability to positively influence Mali (whether by actually improving it, or replacing it with something GeForce) may be beneficial to the OEMs who use it.


> Nvidia is a total powerhouse when it comes to chip design. Most people here are looking at this from the angle of "how does ARM benefit Nvidia", but I think its more valuable to consider "how does Nvidia benefit ARM". In 2020, given what we know about Ampere, I really don't think there's another company out there with better expertise in microprocessor design (but, to be fair, lets say top 3 next to Apple and AMD). Now, they have more of the stack in-house, which may help produce better chips.

In my view you have this completely backwards. I think the opposite is true and that Nvidia is not a powerhouse CPU designer at all. They make extremely impressive GPUs certainly, but that does not automatically translate to great capabilities in CPUs. In terms of CPUs they have so far either used standard ARM designs and have attempted their own Project Denver custom architecture which are not bad but have not impressed CPU wise either. In this area Nvidia would need ARM - primarily for themselves.

> The other important point is Mali, which has a very obvious and natural synergy with Nvidia's wheelhouse. Another example of Nvidia making ARM better; Nvidia is the leader in graphics, this is no argument, so their ability to positively influence Mali (whether by actually improving it, or replacing it with something GeForce) may be beneficial to the OEMs who use it.

I know you're only entertaining the thought, but the image of Nvidia shipping HDL designs of Geforce IP to Samsung or Mediatek in the short term future seems completely alien to me. Things would need to change drastically at Nvidia for them to ever do this.

Certainly Nvidia has the capabilities to sell way better graphics to the ARM ecosystem, and very likely only one line of GPUs can survive, but it just seems extremely unlike Nvidia to ever license Geforce IP to their competitors.


Nvidia actually did try to license out its GPU IP a few years back: https://www.anandtech.com/show/7083/nvidia-to-license-kepler...

I don't believe they ever closed a deal, but clearly Nvidia had some interest in becoming an IP vendor. Perhaps the terms were too onerous or the price too high.


Ah that's very interesting. I wonder what kind of circumstance would get in the way deals for a GPU line that (starting with Maxwell) is technically the best GPU architecture money can buy. Perhaps Nvidia were trying to win game console SoC deals?

> "In 2020, given what we know about Ampere, I really don't think there's another company out there with better expertise in microprocessor design..."

Since Ampere and GPUs in general are structured nothing like a microprocessor, I doubt you'll find anyone who agrees with that.


AMDs answer to Ampere won't be shabby based on info about next generation consoles. They're also widening the gap on CPUs with Zen to where ARM won't have an easy time making inroads on server/workstation.

On the Intel side, the process obstacles have been tragic, but they have plenty of hot products and plenty of x86 market share to lose, or in other words, plenty of time to recover CPU performance dominance.


As Linus said ARM will only be widely adopted if there is availability of affordable hardware that developers can actually buy.

Apple has pulled this off about 4 times because of their small market share and willingness to deprecate old hardware, software and the close control of the hardware they release.|

In the PC world - x86 will remain with us for a LONG time to come.


Which is why Arm in Macs is a crucial moment for the whole of the Arm ecosystem.

I hope so. And when AMD's graphics cards and Intel's processors become good again, they're welcome to reclaim a top spot. But, until then, they are woefully behind.

perhaps they can be part of an 'early insider' program to get access to the next gen architecture improvements. and use that to steer towards integrating their own gpus for a premium instead of the dinky little Malis.

The failed "Windows on Snapdragon" experiment is not because of the fault of Qualcomm but Microsoft's. Qualcomm consistently released chips for PCs since Snapdragon 835 and the chips are pretty competitive in raw power compared to Intel. But even some of Microsoft's own software is not compiled natively for ARM leading to bad performance. No 64-Bit emulation means a lot of programs users download does not work at all. .Net Framework which is probably the most popular .Net framework for Windows programs is not compiled for ARM either so people who want to compile their apps need to first transition to .Net Core, and this wasn't even an option a year ago (because of lack of missing parts in .Net Core). So Microsoft is to blame for the handling of the Windows on Snapdragon experiment.

> There is no way Nvidia will spend 40 billion on this without them deliberately or inadvertently destroy ARM's open CPU and GPU ecosystem

But why would a company spend that much money to buy a company and destroy it afterwards?


Nvidia will certainly try to get as much money out of ARM's R&D capabilities, existing IP, and future roadmap as they can. They will get their money's worth - at worst they will fail trying. In that sense, they won't destroy "ARM the company" or "ARM the IP". But Nvidia will have no interest in maintaining ARM's business model whereby ARM fosters a community of Nvidia competitors - they have an interest to the opposite. Therefore they very likely will destroy "ARM the ecosystem".

It’s not going to destroy the business it’s going to destroy the current business model. The point is that the only way this deal can make sense for Nvidia, is to use ARM’s IP as a competitive advantage over other competitors. Until now, ARM’s value proposition has been IP neutrality for the various user companies.

Oracle and Sun at $8B.

They figure they can make more than $40B by driving ARM to the ground?

Personally I don't think that's true in this particular case, but the strategy isn't exactly unheard of.



I think it's wonderful news that Arm is joining the ranks of giant American tech oligopolies. This is a win for freedom and increases prosperity for all.

/s


Both Apple and windows has arm implementation. In fact with Apple it is arm not risc/v we are looking at. Can Nvidia bring competition or instead destroy arm ... wonder if he wants to destroy what a cpu ip or competition it pay 40B to destroy. What does it want from the deal ? And what will it get from the deal ? Any past deal we can gauge ? Oracle buy sun is the worst scenario.

Exactly, there is no way ARM is worth 40 B to Invidia, unless they are going to use it to arm-twist their competition.

Just look at ARMs annual net, multiply by 10, multiply that by 2 assuming starry-eyed optimism about you being better at generating value from ARM IP, you’re still far from 40 billion.


I agree with the general sentiment here, but ARM is not exactly Snow White. It's an open secret that ARM was (and still is) selling the CPU design at a discount if you integrated their Mali (GPU). This isn't relevant to Nvidia today, but it was when they were in the mobile GPU space. Also this caused obvious problems for IMGtec and other smaller GPU players like Vivante.

" It's an open secret that ARM was (and still is) selling the CPU design at a discount if you integrated their Mali (GPU)."

Why is that bad? Not only it's common business practice (the more you buy from us, the cheaper we sell), it also makes sense from the support perspective. Support the integration between their cores and a different GPU would be more work for them than integration of their cores with their own GPUs.

That's why companies expand to adjacent markets: efficiency.

A completely different thing would be to say: "if you want our latest AXX core, you have to buy our latest Mali GPU". That's bundling, and that's illegal.


A question is how big those discounts are.

Microsoft have away a free browser with their operating system - leaving little room for other browser vendors to serve that market.

Each ARM design deal including a GPU for cheap leaves little room for other GPU vendors.


Bundling isn't necessarily anti-competitive provided ARM isn't taking a loss selling their chip. I'll admit that things aren't actually free-market here because copyright and patent monopolies apply.

There are three possibilities here: ARM's design is approximately the same as the competitory, ARM's design is inferior to the competitor, and ARM's design is superior to the competitor.

If faced with two equivalent products, staying with the same supplier for both is best (especially in this case where the IP isn't supply-limited). The discount means a reduction in costs to make the device. Instead of ARM making a larger profit, their customers keep more of their money. In turn, the super-competitive smartphone market means those savings will directly go to customers.

In cases where ARM's design is superior, why would they bundle? If they did, getting a superior product at an even lower price once again just means less money going to the big corporation and more money that stays in the consumer's pocket.

The final case is where ARM has an inferior design. I want to sell the most performance/features for the price so I can sell more phones. I have 2 choices: slight discount on the CPU but bundled with an inferior GPU or full price for the CPU and full price for a superior GPU. The first option lowers phone price. The second option offers better features and performance. For the high-end market, I'm definitely not going with the discount because peak performance reigns supreme. In the lesser markets, its a calculation of price for total performance and the risk that consumers might prefer an extra few FPS for the cost of another few dollars.

Finally, there are a couple small players like Vivante or Imagination Technologies, but the remaining competitors in the space (Intel, AMD, Nvidia, Qualcomm, Samsung, etc) aren't going to be driven under by bundle deals, so bundling seems to be pretty much all upside for consumers who stand to save money as a result.


I'm sure that ARM is not a saint here in the sense that they would also have an incentive to milk their licensees as much as possible. Now they will keep having that incentive, but also the terrible incentive to actively outcompete their licensees which is much worse.

>I agree with the general sentiment here, but ARM is not exactly Snow White. It's an open secret that ARM was (and still is) selling the CPU design at a discount if you integrated their Mali (GPU).

It is actually the other way around. ARM is more like giving the Mali GPU for free ( or at a very low cost ) if you use their CPU.

>Also this caused obvious problems for IMGtec

Yes, part of the reason why PowerVR couldn't get more traction and Apple were unhappy with their GPU pricing.


Tie-in sale is anti competition and not allow ? May be the per sec is started to go but still ...

Bundling is not necessarily problematic: it happens at your local supermarket all the time!

What would be an issue would be if Arm used their market power in CPUs to try to control the GPU market - e.g. you can't have the latest CPU unless you buy a Mali GPU with it.


>Will Nvidia allow selling ARM licenses to competitors of Nvidia's business?

Like AMD? Sure. None of the ARM IP compete with Nvidia much. Not to mention by "Not" Selling to AMD it create more problem for its $40B asset than anyone could imagine.

>Will Nvidia reserve ARM's best IP as a selling point for its own chips? Will Nvidia allow Mali to continue existing?

Sure. Mali dont compete with Nvidia at all. Unless Nvidia will put up their CUDA Core for IP licensing with similar price and terms to Mali. Could they kill it or raise the price of Mali? Sure. But there is always PowerVR. Not to mention AMD is also licensing out Radeon IP to Mobile. Mostly because AMD dont / cant compete in that segment.

>Unless Nvidia leaves ARM alone (and why would they spend $40B on that??)

It has more to do with Softbank being an Investor. They were already heavily invested in Nvidia. And they need money, they want out. And seriously no one sane would buy ARM for $40B ( It is actually $35B, with $5B as performance bonus, the number $40 was likely used only for headline. ) As a matter of fact I would not be surprised if Softbank promise to buy it back someday. This also paint a picture of how desperate Son / Softbank needs those Cash. So something is very wrong. ( Cough WeWork Cough )

But I do understand your point. Conflict of Interest. Similar to Apple wouldn't want to build their Chip in Samsung Foundry.

While I would have liked ARM to remain independent. I am not as pessimistic as some have commented. And normally I am the one who had people pointing at my pessimism.

On the optimistic side of things. There are quite a lot of cost could be shared with the tools used for TSMC and Samsung Foundry implementation. ( Nvidia is now in bed with Samsung Foundry ) For ARM that means higher margin, for its customers that mean access to Samsung Foundry Capacity where previously they are stuck with TSMC. Nvidia also gets to leverage ARM's expertise in licensing, so their Nvidia GPU could theoretically enter new market. The real IP with Nvidia isn't so much about the GPU design, but its Drivers and CUDA. So may be Nvidia could work towards being an Apple like "Software" company that works with specific Hardware. ( Pure Speculation only )

There are lots of talk about Nvidia and ARM. While I dont think the marriage make perfect sense, It is not all bad. There are more interesting point no one is talking about. Marvell, AMD, Xillix and may be Broadcom. The industry is consolidating rapidly because designing leading edge chip, even with the cheap IP licensing is now becoming very expensive. And the four mentioned above have less leverage than their competitors.

Interest Times.


exactly

When will Europe realise that there is no second place when it comes to a market - the larger player will always eventually end up owning everything.

I can not put into words how furious I am at the UK's Conservative party for not protecting our last great tech company.

Europe has been fooled into the USA's ultra free market system (which works brilliantly for the US but is terrible for everybody else). As such American tech companies have brought EVERYTHING and eventually moth balled them.

Take Renderware it was the leading game engine of the PS2 era consoles, brought by EA and mothballed. Nokia is another great example brought by Microsoft and mothballed. Imagination Technologies was slightly different in that it wasn't bought but Apple essentially mothballed them. Now ARM will undoubtedly be the next via an intermediate buyout.

You look across Europe and there is nothing. Deepmind could have been a great European tech company - it just needed the right investment.


tech is only 10% of the US economy, and European nations are much more reliant on free-trade. Germany in particular, whose exports constitute almost 47% of their GDP, globally only comparable to South Korea for a developed nation.

I get that Hackernews is dominated by people working in software and software news, but as a part of the real economy (and not the stock market) it's actually not that large and Europe doesn't frame trade policy around it, for good reasons.

The US also doesn't support free-trade for economic reasons, but for political and historical reasons, which is to maintain a rule based alliance across the globe, traditionally to fend off the Soviets. Because they aren't around any more, the US is starting to ditch it. The US has never economically benefited from free-trade, it's one of the most insular nations on the planet. EU-Asia trade with a volume of 1.5 trillion almost doubles EU-American trade, tendency increasing, and that's why Europe is free-trade dependent.


Tech sadly is everything hence why it dominates the stock market. Information is the new oil and all that. Its much more than purely economic, whoever controls the tech controls those who use the tech. Hence why China has created its own tech companies - they arent stupid.

I think you're also getting mixed up between 'free-trade' and 'free-markets'. Free trade is about trade deals: NAFTA, WTO, EU, CPTPP, Mercour or whatever trade grouping you want - generally to do with the removal of taxes and standardisation of goods between countries.

Free markets on the other hand is do with the liberalisation of markets i.e removing government intervention (as much as possible) i.e regulations and restrictions of buying and selling of stuff - in this case companies (which can be covered in a trade deal admittedly)

What I'm advocating is that British gov (and most European gov's) restricts the selling of their tech companies based purely on the importance of the tech company.

Why?

Because as I say its do to with control. We're not able to make democratic, sovereign decisions when the fabric of how most things are done is controlled completely by someone else.


Only because of EU. Intra-EU trade is more comparable to trade between US states then true international trade.

I'd say the opposite, intra-EU trade is more like international trade, at least for B2C. Each country has its own national market situation, companies cannot easily expand to the entire EU because in every EU country they will find different local competitors who know the local market much better than they do. Every product has to be localised for the local language and culture. All marketing has to be localised.

Despite efforts to the contrary the EU functions as a glorified free trade zone, half a century of integration cannot beat 1000 years of fragmentation.


Just to be clear my statement has nothing to do the 'EU' which is largely a trade body. I specifically used the term 'Europe' and 'free markets' not 'free trade'. This has nothing to do with Brexit to avoid confusion.

Not true at all - EU does make things simpler but it's still very different legal systems, currencies and even languages.

Why? What's the difference between USA-Mexico trade in car parts and intra-EU trade in car parts?

And you really think more protectionism will help?

Maybe part of the problem is that due to so many regulations, there's not a healthy startup ecosystem and the compensation isn't remotely high enough to draw the best talent.


Regulation has little to do with it. Most of the tech industry is inherently winner-takes-all or winner-takes-most because of how easy it is to scale up tech solutions. US companies get a huge head-start because of their large home market compared to the fragmented EU market, and can easily carry that advantage into also dominating the EU market.

There is a reason Russia and China have strong tech companies and Europe doesn’t. That reason isn’t lack of money, lack of talent or regulations. The only way for Europe to get big tech companies is by removing or crippling big US companies so EU companies can actually compete. The US companies would be quickly replaced by EU alternatives and those would offer high compensation all the same.

Whether or not that is worth it from the perspective of the EU is not so black and white - tech is obviously not everything - but the current situation where all EU data gets handed to the US government on a silver platter is also far from optimal from the perspective of the EU.


Absolutely. The only bit I don't agree with is that tech isn't everything - to all intense and purposes it is (at this point in time).

The only way for Europe to get good tech companies is to create a business friendly environment. They seem unwilling to do that.

When is the time china and Russia has a lead in IT by not protectionist and copycat. I can only think of tik tok.

The strange thing is if you do not count brexit, Arm is one of the many example Uk can do it. And whilst we say Nokia, Sieman and japan fuji (sitting in Hosiptal now and thinking those mri, ...) non-chinese and non-Russia do dominate the tech world even they are not USA. But communist ideology totalitarian I found tik tok is really the exception.

Hence I think Eu has their problem. But not because they are not as good as Russia or china.


Tmall of Alibaba processed 544,000 transactions per second during the peak of its Singles' Day in 2019. I believe this has set a new world record for an e-commerce platform.

In case you didn't know about Singles' Day: https://graphics.reuters.com/SINGLES-DAY-ALIBABA/0100B30E24T...


Huawei. They lead in 5G equipment because they did the necessary R&D investment.

Their phones are pretty good, too (or were pretty good before they got cut off from their suppliers). Their edge was that they built really great cameras into their phones.


Yes the US is all about protectionism, same goes for China, same goes for the EU (also note my use of Europe, EU != Europe). Hence why all three major trade blocks dont have complete free trade deals setup between them (not that comment has anything to do with free trade deals - at least not primarily).

Ultimately free trade is where the world would like to get to purely from an economic basis but you have to do that in tandem with the rest of the world. If you go first everybody else has a economic advantage over you as possibly the UK will find out after Brexit actually happens. Also politics gets in the way of the world achieving full free trade. Some gov will always want votes by protecting an industry - like the UK's fishing industry for instance.


There are a lot of tech companies that start and are successful in Europe. They just get bought when they get big enough and killed later.

Skype is another example to add to the list.

Skype is effectively dead technology and isn’t even promoted any more.


> Maybe part of the problem is that due to so many regulations, there's not a healthy startup ecosystem

Reaganomics talking point since the 80s, yet the U.S. constantly relaxes regulations, recently it released even more environmental ones and it looks in parts like Mars.

But of course, cut regulations, cut corporate taxes, cut benefits, cut, cut, cut. There's never a failure model for such capitalism apparently. 2008 even was blamed on regulation, rather than lack of thereof.

Am quite frankly done with this line of argument.


I know what you are referring to in the US, but that doesn't explain the intra-Europe differences in regulation when it comes to doing business. Like Denmark is number 4, Germany is number 22, and Greece is number 79 on this index? [1] I don't think Denmark (or Norway, at number 9) is exactly an unregulated capitalist paradise where the environment is being destroyed in the name of business.

I say this because I think Europe with a set of consistent regulations and ways of establishing business would serve as a good counterweight to the freewheeling, "anything goes" nature of US capitalism. But I think the fragmentation is its Achilles Heel.

------------------------

[1] https://www.doingbusiness.org/en/rankings?region=oecd-high-i...


I don’t really see your point. Even the examples you list are nonsensical.

* We are no longer in the ps2 era. EA now uses Frostbite, which was developed by the Swedish studio Dice. It is alive and well, powering some 40-50 games. https://en.m.wikipedia.org/wiki/Frostbite_(game_engine)

* Nokia was dead well before MS bought them.


I sort or agree but those aren't exactly great example.

I dont see how Renderware would compete with Unreal. Even their owner EA choose Unreal. They were great in PS2 era, but next Gen console ( PS3 ) they were not.

Nokia was dead even before Stephen Elop became the CEO. So the Microsoft acquisition has nothing to do with it.

IMG - Yes. But I would argue they were dead either way. They couldn't get more GPU licensing due to ARM's Mali being cheap and good enough. They couldn't expand into other IP licensing areas. Their MIPS acquisition was 5 years too late. Their wireless part couldn't compete with CEVA. And they somehow didn't sell themselves to Apple as an Exit. ( But then Apple lied about not using IMG's IP. While Steve Jobs often put a spin thing, I find Time Cook's Apple quite often just flat out lying )


I'm a game rendering engineer that uses Unreal (and used Renderware) so I know a little about this subject.

If Renderware hadn't been brought by EA (and hence controlled by your competitor), the rest of the industry would probably have kept using Renderware as it was the best option and development would have continued. It would have been built on to deliver next gen experiences.

It mirrors pretty much perfectly what is wrong with the ARM Nvidia deal.

Nokia yes wasn't doing well in the smart phone sector but was doing excellently in the feature phone sector. Hence why HMD Global is now doing very well selling those handsets.


And if Europe develop a company that threatens US leading tech / surveillance companies like Facebook / Google, or becomes the leader of the next tech wave, be prepared for US government actions to take it down. See: Japanese semi conductor industry in the 80s, Alstim, Bombardier, Tiktok.

One could argue that's what happened to NOKIA - the only non-US company controlling the largest growing market at the time. For anyone saying that it was dead already, you don't understand that despite minimal presence on the US market it controlled over 60% of the world's market and that could have lasted much longer due to loyalty of customers and upcoming MeeGo (that even US reviewers liked) if it wasn't Eloped.

Everyone I know here in Europe thought Nokia was dead when MS bought it.

Yes, when they bought it it was already dead, but I was talking about the time they were #1 and just got their new ex-MS CEO with a $20M bonus in case he sold the company. Then he proceeded to dismantle all that worked.

It doesn't even have to be the US government itself doing this; US companies are becoming large and strong enough to do as they please without regard to what lesser organizations and even countries have to say.

Perhaps governments around the world should do what the US did (and still does) to foreign companies before its too late.


They have yet to do this with rare earth metals, which represent a huge strategic threat to both the US and much of the rest of the world to rely on China for a vast majority of the supply.

It has been said that other countries, including US, in fact have more rare earth deposits. If they choose to relax environmental laws and disregard patents held by Chinese corporations, more rare earth would be produced.

Yes, other countries have them. They're not willing to produce them at the higher cost required to obey environmental laws. I'm not sure how patent laws are a huge issue here: rare earths may have increasing market demand, but they've been extracted and processed for countless years. Only innovations China has made in the last ~20 years would be covered by patents. I'm sure there have been some innovations, but I'm also sure our other mining industries in the US would be happy to leverage their patent portfolio in a patent war if China wasn't willing to work on reasonable licensing terms.

A global strategic bottleneck in China for these things doesn't require patent or environmental law violations, it just requires us to pay more for them. For a critically strategic resource like this, the US should ensure a consistent supply chain independent of geopolitical concerns with China. And if China were to cut off supply then concern for their patents goes out the window too.


Wait - it was OK when ARM was owned by a Japanese company, but suddenly bad when a US company buys them? Being anti-US is not a legitimate point. If anything, nVidia will create new market opportunities and expand existing ones for ARM. They have already said they will be expanding their UK presence. Maybe don't react so emotionally next time.

No it wasn't OK when Softbank bought them, I vehemently disagreed with the sale then, as this was the next step after Softbank acquired them.

Although I dont agree with selling to Softbank at least they didn't have a dog in the game. What is bad about Nvidia is that they have a dog in the (chip) game. A major reason you went to ARM was for a non biased design team - you knew you were getting their best if you paid for it. Now I'm afraid you don't.

Also my comment is not anti US, it just so happens that the US has all the big tech companies and foolishly the European countries, especially the UK believes it can compete in a level playing field with the US even though the US's GDP is about 9 times bigger than the UK's - god knows how much bigger its equities markets are.

At the EU level the US is not but then this kind of stuff isn't decided about at the EU level - maybe it should be, not that that will help the UK see the error in its ways after Christmas.

As for emotional reaction with the greatest of respect did you read your message before you posted it?


ARM was second place in laptops for quite a while (arguably still.)

If Europe doesnt like American business they can make their own company or use open source... Risc V is a great alternative to arm anyways.

Bad for the US as well.

A monoculture is bad. Any monoculture.


DICE, the Swedish game developer, bought by EA and is now belly up

brought = bought ?

lol yes and in the next line you see I used 'bought' just a simple typo.

Brought?

No one has seemed to notice the following two things:

"To pave the way for the deal, SoftBank reversed an earlier decision to strip out an internet-of-things business from Arm and transfer it to a new company under its control. That would have stripped Arm of what was meant to be the high-growth engine that would power it into a 5G-connected future. One person said that SoftBank made the decision because it would have put it in conflict with commitments made to the U.K. over Arm, which were agreed at the time of the 2016 deal to appease the government." (from https://arstechnica.com/gadgets/2020/09/nvidia-reportedly-to... )

and

"The transaction does not include Arm’s IoT Services Group." (nvidia news.)

which appear to contradict each other.

I'm not sure about the significance of this. I would have guessed Nvidia would have wanted the IoT group to remain.

Also, to first order, when a company issues stock to purchase another corporation, that cost is essentially "free" since the value of the corporation increases.

In other words, Nvidia is essentially paying $12 billion in cash for ARM up front, and that's all. (The extra $5B in cash or stock depends on financial performance of ARM, and thus is a second-order effect.)


It is not "free", it means current shareholders of Nvidia are paying for the remaining money. Their stock is diluted on fresh issue of shares.[1]

The $12B comes from Nvidia the company, the remaining money comes from Nvidia's shareholders directly.

[1] Only if the valuation of ARM is "worth it" the fresh issue of shares will not cost the current shareholders anything. This is rarely the case , if Nvida overvalued(or less likely undervalued) the deal then current shareholders are giving more than they got for it.


And note that the $12B is also owned by Nvidia's shareholders (they own the company which owns the cash), so they're paying for that too.

It's just different forms of shareholder assets being traded for other assets, by the shareholders (or rather, their majority vote).


Effectively all of it is funded by the shareholders who "own" the corporation.

In the current example if the cash is coming from a debt instrument is it not bank funding it now?

it is typically about who is fronting the money now, it could be your bank making loans, or from cash reserves you have, fresh stock issue, selling another asset, or even the target's bank as in the case of LBO.

The shareholders always end up paying for it eventually in some form or other. Differentiating it by the current source helps understand the deal structure and risks better.


It's still the shareholders even if debt financing is used. The shareholders, through the company, have to pay off the debt. It increases the risk of bankruptcy.

When doing a deal, it comes down to price and source of funds. Changing either can drastically change how good or bad the deal is.


Absolutely. I was just noting it for anyone reading.

To me, "paying for" and "diluted" connotates negative emotions.

Cash paid in this instance is treated no different than cash in their normal operating expenses. If either generates profits in line with their current expected returns, the stock price stays the same, and everyone is indifferent to the transaction.

Same goes for stock issuance. If the expectation of the use of proceeds from the issuance are in line with the company's current expected returns, everyone is indifferent.

Your statement is still true, and the stock market jumped today on the news, so I feel my connotation is misplaced.


> If either generates profits in line with their current expected returns, the stock price stays the same

This is a pretty naive depiction of Wall Street.


Yes

There were two separate IoT business units: Platform (https://pelion.com) and Data (https://www.treasuredata.com/). The Platform unit fits the Segars post-acquisition comment about end-to-end IoT software architecture, https://news.ycombinator.com/item?id=24465005

> One person close to the talks said that Nvidia would make commitments to the UK government over Arm’s future in Britain, where opposition politicians have recently insisted that any potential deal must safeguard British jobs.

So the deal has already been influenced by one regulator. That should encourage other regulators.

> SoftBank will remain committed to Arm’s long-term success through its ownership stake in NVIDIA, expected to be under 10 percent.

Why is this stake necessary?


The UK government never learns. Kraft made similar promises when buying Cadbury regarding jobs,but quietly reneged on them over time. The Kraft CEO was asked to show up before an UK parliament committee, but of course declined, and that was the end of the story.

Except the Cadbury-Kraft debacle led to major reforms in how the UK regulates foreign takeovers.

In the case of Arm, the guarantees provided back in 2016 were legally binding, which is why we’re here, four years and another acquisition later, with Nvidia now eager to demonstrate it is standing by those commitments.

Maybe in this particular instance they did learn something?


They should learn to prevent the sale, period. Promises to retain some jobs (for ever? What happens if profits decline? What happens if the core tech is sold and ARM becomes a shell? Do the rules still apply?) address a tiny fraction of the problems presented by the sale of one of our core national tech companies.

I would have liked to see ARM remain owned in the UK. I think it's proven itself capable of innovation and organic growth on its own.

But how can we evaluate the whether that will continue?

What if ARM is not sold, and then (for whatever reason) stagnates, doesn't innovate, gets overtaken in some way, and enters gradual decline?

Perhaps that's unlikely, but prevent the sale, period is feels too absolute.


Who is to say ARM was “owned in the UK” prior to the SoftBank acquisition?

Prior to that it was a publicly traded company, presumably with a diverse array of international shareholders.


That feels like giving up, to me. We should have the confidence that British industry can development and flourish on its own merits without being sold off to foreign interests.

What have confidence when we can just look at ARM financials?

There are more ARM chips sold each year than those of all its competitors together. Yet ARM's revenue is 300 million $.

Why? Because ARM lives from the ISA royalties, and their revenue on the cores they license is actually small.

With RISC-V on the rise, and west sanctions against china, RISC-V competition against ARM will only increase, and it is very hard to compete against something that's good / better and has lower costs (RISC-V royalties are "free").

I really have no idea why NVIDIA would adquire ARM. If they want a world-class CPU team for the data-center, ARM isn't that (Graviton, Apple Silicon, Fujitsu, etc. are built and designed by better teams). ARM cores are used by Qualcom and Samsung, but these aren't world-class and get beaten every gen by Apple Silicon. If they want ARM royalties, that's high risk business, and very low reward (there is little money to make there).

The only ok-ish cores ARM makes are embedded low-power cores (not mobile, but truly IoT < 1W embedded). Hard to imagine that an architecture like Volta or Ampere that perform well at 200-400W would perform well at the <1W envelope. No mobile phone in the world uses nvidia accelerators, and mobile phones are "supercomputers" when compared with the kind of devices ARM is "ok-ish" at.

So none of this makes sense to me, except if NVIDIA would want to "license" GPUs with ARM cores to IoT and low power devices like ARM does, but that sounds extremely far-fetched, because nvidia is super-far away from a product there, and also because the margins for those products are very very thin, and nvidia tends to like 40-60% margins. You just can have those when buying IoT chips for 0.12$. Its also hard to sell a GPU to these use cases because they often don't need it.


> If they want a world-class CPU team for the data-center, ARM isn't that (Graviton

Graviton uses Neoverse CPU cores, which are designed by ARM. To say that ARM is not a world-class CPU team is unfair. Especially as Ampere just announced an 80 core datacenter SoC using Neoverse cores.


The main source of revenue for ARM is, by far, royalties. Licenses are paid once, royalties are paid by unit shipped. And they shipped billions last year.

Revenue is not $300, we don't know what ARM's revenue is because it hasn't been published since 2016. And back then it was like $1.5 billion. $300 million was net income. Again, in 2016.

I think you've already been adequately corrected on your misconceptions about ARM's CPU design teams.


Apple mobile beating arm mobile is pretty irrelevant considering they can barely run the same type of tasks it's a miracle they both have the same benchmarks since they're operating systems are totally different. Beyond that it doesn't even really matter which one's faster because you got all that overhead of the os's that are so different. I would also like to say that Apple users are going to buy the newer chip whether it's faster or not. Because of planned obstinance at a software level. I've never heard an Apple user talk about speed or The benchmark of the phones because that marketing segment is clueless to that. My guess is that Nvidia needs arm for their self-driving stuff.

>" If they want a world-class CPU team for the data-center, ARM isn't that (Graviton, Apple Silicon, Fujitsu, etc. are built and designed by better teams)."

The latest Fujitsus HPC offering the A64FX is also ARM based though.[1][2] And it sounds as though this is replacing their SPARC64 in this role .

[1] https://en.wikipedia.org/wiki/Fujitsu_A64FX

[2] https://en.wikipedia.org/wiki/Fugaku_(supercomputer)


Fujitsu is using custom cores, but AWS Graviton (1 and 2) actually uses Arm Neoverse cores. The Ampere Altra will too.

The core is not designed by ARM.

Nvidia could still asset strip ARM, and then let ARM decline organically with redundancies justified by the decrease in revenue.

Sounds like with all those new regulations it will be a limiting factor for new jobs not going to set up my company there if I got to jump through hoops like that.

It absolutely outrages me. I've posted elsewhere on the thread about this, but I want to do something to prevent this sort of thing in future. Setting up a think tank seems like a viable idea. See my other response in the thread if you are interested and want to get in contact.

Welcome to vulture capitalism it's been gutting American industry for about 10 20 years used to be four or five paper mills around here but now they're all in China and everyone who worked there is on drugs now

Welcome to vulture capitalism it's been gutting American industry for about 10 20 years

More like 30-40. https://en.wikipedia.org/wiki/Private_equity_in_the_1980s


Get the deals in writing, with explicit liabilities if contract is broken. There, I fixed it.

What, and the profits, investments and growth made by the company will stay inside the country? There's a much broader problem here.

What's this "quietly reneged over time"! Kraft simply welshed on the deal instantly! There were explosions and a change in the law.

it's all posturing unless the government makes the company to agree to punitive severance packages to be put in an escrow account to be released to the company after 20 years or to the people if fired before that.

> Why is this stake necessary?

Edit: it’s not necessary/a requirement.

They’re noting that after the transaction, SoftBank will still fall under the 10% ownership threshold that requires more reporting from the SEC [1]:

> Section 16 of the Exchange Act applies to an SEC reporting company's directors and officers, as well as shareholders who own more than 10% of a class of the company's equity securities registered under the Exchange Act. The rules under Section 16 require these “insiders” to report most of their transactions involving the company's equity securities to the SEC within two business days on Forms 3, 4 or 5.

[1] https://www.sec.gov/smallbusiness/goingpublic/officersanddir...


Are they getting 10% of nvidia or keeping 10% of Arm

From the press release directly, it appears to be 10% of NVIDIA:

https://nvidianews.nvidia.com/news/nvidia-to-acquire-arm-for...

> Under the terms of the transaction, which has been approved by the boards of directors of NVIDIA, SBG and Arm, NVIDIA will pay to SoftBank a total of $21.5 billion in NVIDIA common stock and $12 billion in cash, which includes $2 billion payable at signing. The number of NVIDIA shares to be issued at closing is 44.3 million, determined using the average closing price of NVIDIA common stock for the last 30 trading days. Additionally, SoftBank may receive up to $5 billion in cash or common stock under an earn-out construct, subject to satisfaction of specific financial performance targets by Arm.

Since NVIDIA currently has 617 million shares outstanding, if the earn-out were to be fully in common stock, this would bring Softbank to 8.8% of NVIDIA from this transaction alone (plus anything they already have in NVIDIA as public-market investors).

(I believe that the Forbes analysis in the sibling comment is mistaken in describing this as a "10% stake in new entity" - no new entity is mentioned in the press release itself.)


> no new entity is mentioned in the press release itself

The Forbes article is based on a joint interview today with the CEOs of Arm and Nvidia, who could have provided more detail than the press release, specifically:

> Arm operating structure: Arm will operate as an NVIDIA division

This level of detail can be confirmed during the analyst call on Monday. Operating as a separate division would help assuage concerns about Arm's independence. The press release says:

> Arm will remain headquartered in Cambridge ... Arm’s intellectual property will remain registered in the U.K.

Those statements are both consistent with Arm operating as a UK-domiciled business that is owned by Nvidia.


Given the quality of the reporting so far it’s possible that they are getting 10% of the UK.

10% of Arm division of Nvidia, https://www.forbes.com/sites/patrickmoorhead/2020/09/13/its-....

> Softbank ownership: Will keep 10% stake in new entity

That would mean Nvidia acquired 90% of Arm for $40B, i.e. Arm was valued at $44B.

Is Softbank invested in Arm licensees, who may benefit from Softbank influence? Alternately, which Arm licensees would bid in a future auction of Softbank's 10% stake of Nvidia's Arm division?


Does this mean they are or aren't buying treasure data? What would happen if not?

That would be up to Softbank.

So they aren't buying treasure data then?

Looks like they continue as independent company, owned by Softbank, https://blog.treasuredata.com/blog/2020/09/14/nvidia-to-acqu...

> The transaction does not include Arm’s IoT Services Group, which is made up of Treasure Data and Arm’s IoT Platform business.

> Back in July, Arm announced a proposed transfer to separate Treasure Data from its core semiconductor IP licensing business, enabling Treasure Data to operate as an independent business. That separation is on track and will be completed before the close of the NVIDIA transaction. Most importantly, you should not see any disruption in our service or support as a result of this news.


>Also, to first order, when a company issues stock to purchase another corporation, that cost is essentially "free" since the value of the corporation increases.

This isn't correct. If investors thing Nvidia overpaid, its share price will decline. There are many examples of acquiring companies losing significant value on announcements to buy other companies even in pure stock deals.


In addition, it's not even a valid argument if the cost was entirely in cash.

One would be making the argument, "the cost is essentially free because although we spent $40B, we acquired a company worth $40B". Obviously, that's not any more correct than the case of paying in stock.


Correct, if it's shares, they just dilluted the value of the stock. If I owned 1% of the company, after printing and handing out $40bn worth of stock, then I keep the same amount of stock, but now it's (e.g.) 0.5%, which means I just lost 50% of that investment. Which means I will receive 50% of the dividend (since out of the next year's dividend 'pool' someone will get a big chunk. Which means NVidia just screwed the existing shareholders (for the time being, and once the numbers/$ are merged I should be getting my dividend by two parts, one from ARM and one from NVIDIA).

If they gave away cash, that's a different story, it all depends.. if they were sitting on $1tn cash, and the spent $40bn, that's no biggie. I mean we went through COVID, what worse can come next?


> If I owned 1% of the company, after printing and handing out $40bn worth of stock, then I keep the same amount of stock, but now it's (e.g.) 0.5%, which means I just lost 50% of that investment. Which means I will receive 50% of the dividend

Well, you’re now getting 50% of the dividend produced by the new combined entity. If the deal was correctly priced, your share of the Arm dividend should exactly replace the portion of your Nvidia dividend that you lost through dilution.


That's not how equity works. Paying cash or shares, the end result is exactly the same for shareholders barring perhaps some tax implications.

I pay the car dealer $20,000, I get a car worth $20,000. Was the car free?

Technically, yes. Your net worth is unchanged. The problem is that the car is no longer worth $20,000 as soon as you drive it off the lot.

actually as soon as u get the car its not a 20k$ car anymore.

I agree that paying in new shares doesn't make the acquisition free. But the value of shares is an expectation of future cash flows. If that expectation is currently on the high side for Nvidia and on the low side for Arm then paying in shares makes the acquisition cheaper for Nvidia than paying cash.

Nvidia is essentially telling us that they think their shares are currently richly valued. I agree with that.


Not necessarily. They could just prefer to be diversified. It can be rational to take a loss of value in the pursuit of diversification (depends on the portfolios of the stakeholders).

They could also think their shares are valued accurately but believe the benefits of synthesis would increase the value.


How would any of that make payment in shares beneficial compared to cash payment?

$12B cash appears to be all they have available. They can't pay more cash.

https://www.marketwatch.com/investing/stock/nvda/financials/...

Their only other option would be taking on debt.


Right, so Nvidia with their A2 rating and their solid balance sheet in a rock bottom interest rate environment still found it favourable to purchase Arm with newly issued shares.

That's telling us something about what they think about their share valuation right now.


That's not how interest works. "Low rate environments" are low rate because debt is less attractive than in a high rate environment. Debt isn't automatically preferable when interest rates are low - if it were, rates would rise. There's no free lunch.

All it tells you is that they think it's preferable to taking on debt, which in some sense is the position you always start from. Debt has a deadweight cost that you have to overcome.

But we can test your theory. You're saying they thought their shares were overvalued. The market's reaction was the opposite - announcing the deal bumped their share price 7.5%.


Perhaps there are tax implications?

>If investors thing Nvidia overpaid, its share price will decline

You realize that logic no longer matters in this economy. There's an oversupply of printed money and stonks literally only going up.


It's rare for the acquiring companies stock to not decline, regardless of whether the market thinks they overpaid.

I worked at Parse when it was bought by Facebook. The day the news broke, Facebook’s market cap grew by multiples of the acquisition price. I remember being gobsmacked that Facebook effectively got paid to buy us.

Unless Facebook itself actually issued stock at the new price, Facebook did not get paid. It were the Facebook shareholders that got paid.

Really, it shows that the market valued Parse much more than the cash it cost Facebook. If Parse was bought with stock instead of cash, that's almost cooler. Since it allowed Parse to capture more of the surplus value they created. (Since the stock price popped).


More sophisticated analysis is required. You should look at the beta of Facebook at the time and determine whether the market reacted positively to the acquisition, or if the move was inline with Facebook's exposure to the market

Keep in mind that market cap is a fictional aggregate. It does not represent the real price that all shares could be sold for.

I see this going a few ways for different players:

The perpetual architecture license folks that make their own cores like Apple, Samsung, Qualcomm, and Fujitsu (I think they needed this for the A64FX, right?) will be fine, and may just fork off on the ARMv8.3 spec, adding a few instructions here or there. Apple especially will be fine as they can get code into LLVM for whatever "Apple Silicon" evolves into over time.

The smaller vendors that license core designs (like the A5x and A7x series, etc.) like Allwinner, Rockchip, and Broadcom are probably in a worse state - nVidia could cut them off from any new designs. I'd be scrambling for an alternative if I were any of these companies.

Long term, it really depends on how nVidia acts - they could release low end cores with no license fees to try to fend off RISC-V, but that hasn't been overly successful when tried earlier with the SPARC and Power architectures. Best case scenario, they keep all the perpetual architecture people happy and architecturally coherent, and release some interesting datacenter chips, leaving the low end (and low margin) to 3rd parties.

Hopefully they'll also try to mend fences with the open source community, or at least avoid repeating past offenses.


> The perpetual architecture license folks that make their own cores like Apple, Samsung, Qualcomm, and Fujitsu (I think they needed this for the A64FX, right?) will be fine

There is one thing they would need to worry about though, which is that if the rest of the market moves to RISC-V or x64 or whatever else, it's not implausible that someone might at some point make a processor which is superior to the ones those companies make in-house. If it's the same architecture, you just buy them or license the design and put them in your devices. If it's not, you're stuck between suffering an architecture transition that your competitors have already put behind them or sticking with your uncompetitive in-house designs using the old architecture that nobody else wants anymore.

Their best move might be to forget about the architecture license and make the switch to something else with the rest of the market.


> Their best move might be to forget about the architecture license and make the switch to something else with the rest of the market.

This assumes that there isn't some other factor in transitioning architecture - this argument could boil down in the mid 2000's to "Why not go x86/amd64", but you couldn't buy a license to that easily (would need to be 3-way with Intel/AMD to further complicate things)

Apple has done quite well with their ARM license, outperforming the rest of the mobile form factor CPU market by a considerable margin. I don't doubt that they could transition - they've done it successfully 3 times already, even before the current ARM transition.

Apple under Cook has said they want to "to own and control the primary technologies behind the products we make". I doubt they'd turn away from that now to become dependent on an outside technology, especially given how deep their pockets are.


It's kind of puzzling that Apple didn't buy them. They don't seem to be particularly aggressive/creative in the M&A department

They have a perpetual license. What value is it for them to buy ARM?

Many of Apple's direct competitors are ARM customers, and US politics is already turning a large anti-trust eye towards FAANG. That might have been a factor.

Apple buying arm would never get approved. They have way to many direct competitors that heavily rely on SoCs that has arm cores.

Apple could transition in 10 years to RISC-V, just like how they transitioned 10 years ago to x86, 10 years before to PPC, 10 years before to .........

They did dump ZFS when they decided they didn't like the licensing terms.

> Apple has done quite well with their ARM license, outperforming the rest of the mobile form factor CPU market by a considerable margin.

But that has really nothing to do with the architecture. They just spent more money on CPU R&D than their ARM competitors. They could have done the same thing with RISC-V, and if that's where the rest of the industry is going, they could be better off going there too. Especially for Mac, since they're about to do a transition anyway. They could benefit from putting it off for another year while they change the target architecture to something the rest of the market might not be expected to avoid in the future.

There is also no guarantee that their success is permanent. They might have done a better job than Qualcomm this year, but what happens tomorrow, if Google throws their hat into the ring, or AMD makes a strong play for the mobile market, or Intel gets their heads out of their butts, or China decides they want the crown and gives a design team an unlimited budget? There is value in the ability to switch easily in the event that someone else is the king of the mountain for a while.

> I don't doubt that they could transition - they've done it successfully 3 times already, even before the current ARM transition.

Just because they can do it doesn't mean it's free.

Really if Intel was shrewd, they'd recognize that they've lost Apple's business already and just sell them an x86 license, under some terms that Intel would care about and Apple wouldn't, like they can only put the chips in their own devices. Then Apple could save themselves the transition entirely and do another refresh with processors from Intel while they put their CPU team to task redesigning their own chips to be x86_64. It would give both Apple and Intel a chance to throw a punch at Nvidia (which neither of them like) while helping both of them. Apple by avoiding the Mac transition cost and Intel by maintaining the market share of their processor architecture and earning them whatever money they get from Apple for the license.

And it gives Intel a chance to win Apple's business back. All they have to do is design a better CPU than Apple does in-house, and Apple could start shipping them again without having to switch architectures. Which is also of value to Apple because it allows them to do just that if Intel does produce a better CPU than they do at any point in the future.


The problem I see with your argument is that it ignores their superior mobile ARM SoC. If they are going to unify on an architecture, it would certainly be easier to migrate Mac to ARM than iOS to x86.

Your assumption is that the architecture matters and they couldn't build an x86 SoC which is just as good, but they could.

And you don't want to move to ARM if everybody else is moving away from it.


Samsung, Qualcomm, and MediaTek all currently just use off the shelf A5x & A7x cores in their SoCs. Unless that part of the company is losing money I don't expect nVidia to cut that off. Especially since that's likely a key part of why nVidia acquired ARM in the first place - I can't imagine they care about the Mali team(s) or IP.

What happens if/when Nvidia launches SoCs that compete with Qualcomm and MediaTek. Will it continue to offer the latest cores to competitors when it will make a lot more money on its own SoCs? This is the reason for the widepsread concern about Nvidia owning Arm.

I don't know if Nvidia is eager to re-enter the SoC market. It wouldn't be a clear path to more money, since they would need to then handle modem, wifi, ISP, display, etc... instead of just CPU & GPU. And they'd need to work with Android & its HALs. And all the dozens/hundreds of device OEMs.

They could, but that's more than just an easy money grab. Something Nvidia would already be familiar with from Tegra.

What seems more likely/risky would be nvidia starts charging a premium for a Mali replacement, or begins sandbagging Mali on the lower end of things. But Qualcomm already has Adreno to defend against that.


Looks like Nvidia never left the SoC market to begin with.

The latest Tegra SoC launched March of 2020.


They released a new SBC aimed at autonomous vehicles. They haven't had a mobile SoC since 2015's Tegra X1. Which only made it into the Pixel C tablet, Nintendo Shield, and Nvidia's Shield TV (including the 2019 refresh)

You're forgetting the TX2 as well as the Jetson Nano

Which ended up in what mobile products?

Mediatek are in the lower end market, something Nvidia's culture doesn't like competing in. Qualcomm holds the key to Modem. Which means Nvidia competing with Qualcomm wont work that well. Not to mention they have already tried that with Icera and generally speaking Mobile Phone SoC are low margin business. ( Comparatively Speaking )

Completely take your point on Qualcomm.

On mobile SoC margins I guess that margins are low because there is a lot of competition - start cutting off IP to the competition and margins will rise.

I suspect that their focus will be on the server / automotive to start off with but the very fact that they can to any of this is troubling for me.


Qualcomm used to design their own cores up until the last generation or two, but you're right they use the reference designs now.

EDIT: correction, make that the last generation or four (oops, time flies)


Qualcomm hasn't used an in-house core design since 2015 with the original Kryo. Everything Kryo 2xx and newer are based on Cortex.

That was a really sad time honestly. QCOMM went from leading the pack to basically using reference designs (which they still arrogantly brand as Kryo despite essentially being a tweak of a reference design.

It all happened because Apple came out with the first 64-bit design, and QCOMM wasn't ready. Rather than deliver 32-bit for 1 more year, they used an off the shelf ARM 64-bit design (A57) in their SoC called the Snapdragon 810, and boy was it terrible.


From what I gathered, they made at least _some_ risky architecture choices in their custom architecture that turned out not to be sustainable over the next generations. Also note that their Cortex core is indeed customized to a significant extent.

"and may just fork off on the ARMv8.3 spec, adding a few instructions here or there"

No, they may not. People keep suggesting these kinds of things, but part of the license agreement is that you can't modify the ISA. Only ARM can do that.


Well, regardless of whether this amendment is kosher or not, AMX definitely exists. Perhaps the $2T tech behemoth was able to work out a sweetheart deal with the $40B semiconductor company.

> There’s been a lot of confusion as to what this means, as until now it hadn’t been widely known that Arm architecture licensees were allowed to extend their ISA with custom instructions. We weren’t able to get any confirmation from either Apple or Arm on the matter, but one thing that is clear is that Apple isn’t publicly exposing these new instructions to developers, and they’re not included in Apple’s public compilers. We do know, however, that Apple internally does have compilers available for it, and libraries such as the Acclerate.framework seem to be able to take advantage of AMX. [0]

my123's instruction names leads to a very shallow rabbit hole on google, which turns up a similar list [1]

Agreed upon: ['amxclr', 'amxextrx', 'amxextry', 'amxfma16', 'amxfma32', 'amxfma64', 'amxfms16', 'amxfms32', 'amxfms64', 'amxgenlut', 'amxldx', 'amxldy', 'amxldz', 'amxldzi', 'amxmac16', 'amxmatfp', 'amxmatint', 'amxset', 'amxstx', 'amxsty', 'amxstz', 'amxstzi', 'amxvecfp', 'amxvecint']

my123 also has ['amxextrh', 'amxextrv'].

[0] https://www.anandtech.com/show/14892/the-apple-iphone-11-pro....

[1] https://www.realworldtech.com/forum/?threadid=187087&curpost...


That's untrue.

(famously so, Intel used to ship arm chips with WMMX and Apple for example ships their CPU today with the AMX AI acceleration extension)


WMMX was exposed via the ARM coprocessor mechanism (so it was permitted by the architecture). The coprocessor stuff was removed in ARMv8.

Now custom instructions are directly on the regular instruction space...

(+ there's the can of worms of target-specific MSRs being writable from user-space, Apple does this as part of APRR to flip the JIT region from RW- to R-X and vice-versa without going through a trip to the kernel. That also has the advantage that the state is modifiable per-thread)


In ARMv8 you have a much cleaner mechanism through system registers(MSR/MRS).

Apple has been using system registers for years already. AMX is interesting because it's actual instruction encodings that are unused by the spec.

That's like saying that my Intel CPU comes with an NVIDA Turing AI acceleration extension. The instructions the CPU can run on an Apple ARM-based CPU is all ARM ISA. That's in the license arrangement, if you fail to pass ARM's compliance tests (which include not adding your own instructions, or modifying the ones included) you can't use ARM's license.

Please, stop spreading nonsense. All of this is public knowledge.


No. I reverse-engineered it and AMX on the Apple A13 is an instruction set extension running on the main CPU core.

The Neural Engine is a completely separate hardware block, and you have good reasons to have such an extension available on the CPU directly, to reduce latency for short-running tasks.


Are the AMX instructions available in EL0?

Is it possible AMX is implemented with the implementation-defined system registers and aliases of SYS/SYSL in the encoding space reserved for implementation-defined system instructions? Do you have the encodings for the AMX instructions?


AMX instructions are available in EL0 yes, and are used by CoreML and Accelerate.framework.

A sample instruction: 20 12 20 00... which doesn't in any stretch parse as a valid arm64 instruction in the Arm specification.

Edit: Some other AMX combinations off-hand:

00 10 20 00

21 12 20 00

20 12 20 00

40 10 20 00


very interesting, thanks!

The AMX is an accelerator block... If you concluded otherwise, your reverse-engineering skills are not great...

Let me repeat this: part of the ARM architectural license says that you can't modify the ISA. You have to implement a whole subset (the manual says what's mandatory and what's optional), and only that. This is, as I've been saying, public knowledge. This is how it works. And there are very good reasons for this, like avoiding fragmentation and losing control of their own ISA.

And once again, stop spreading misinformation.


Hello,

Specifically about the Apple case,

After your tone, not certainly obligated to answer but will write one quickly...

Apple A13 adds AMX, a set of (mostly) AI acceleration instructions that are also useful for matrix math in general. The AMX configuration happens at the level of the AMX_CONFIG_EL1/EL12/EL2/EL21 registers, with AMX_STATE_T_EL1 and AMX_CONTEXT_EL1 being also present.

The list of instructions is at https://pastebin.ubuntu.com/p/xZmmVF7tS8/ (didn't bother to document it publicly at least at this point).

Hopefully that clears things up a bit,

And please don't ever do this again, thank you. (this also doesn't comply with the guidelines)

-- a member of the checkra1n team


You may be correct, but do you really have to be so attacking?

Can you provide a link to the "public knowledge" for those who don't know?

https://www.electronicsweekly.com/news/business/finance/arm-...

Broadcom has an architectural license. They do also license core designs.

Really if nVidia locks up the lower end cores, then a lot of stuff breaks. Billions of super tiny ARM cores are everywhere. ARM has few competitors in the instruction set space for low end, low power, low cost cores. AVR, PIC, and MIPS are what come to mind. And AVR/PIC are owned by Microchip corporation.

These ARM chip unit licenses are dirt cheap, there's hundreds of small manufacturers, and their chips go in everything, and in unexpected places. And these aren't just little microprocessors anymore. They're even in SoCs as little coprocessors that manage separate hardware components in realtime.

The amount of penetration ARM has in hidden places cannot be underestimated. And there isn't a quick replacement for them. Not one freely licensed to any manufacturer.


It seems to me that if Apple felt that Nvidia would limit them, they could have outbid them for ARM! So I think you are correct.

Apple would not get antitrust approval (iPhone maker controls all android chips????). So that’s why.

Were it a serious enough threat to their ARM license, they'd find a way to buy ARM and keep it independent.

Exactly, Apple is already straddling the line (and imho way past it) on anti-comp laws.

I'm not defending Apple, just thinking that can't we say this for many of the biggest tech firms? They are way past the line on anticompetitive business.

Yes, Apple is not alone in this. Google is another example, and they are very aware of this and are acting very carefully

What is their monopoly or which of their competitors are they colluding with?

Neither of those are required to violate anti-trust laws.


I think Apple is not committed to ARM at all. Bitcode, Rosetta 2, "Apple Silicon" - it all suggests they want to keep ISA flexibility.

Exactly. Apple’s strategy here is very clear:

Offer customers iOS apps and games on the next MacBook as a straight swap for Boot Camp and Parallels. Once they’ve moved everyone over to their own chips and brought back Rosetta and U/Bs they’re essentially free to replace whatever they like at the architecture level.

In their reveal I noticed that they only mentioned ARM binaries running in virtual environments. It makes sense if you don’t want to commit to supporting GNU tools natively on your devices (as it would mean sticking with an established ISA)


I would be quite surprised if LLVM ever lost the ability to compile C for the platform.

Exactly my point though.

Apple is large enough that if they want to break from ARM in the future they can do so by forking an LLVM backend. That's not a large job if it's a very small change, but once it's been done once they have plenty of resources to provide ongoing support (like they do for webkit).

The dividends of doing so are potentially massive. Given that they've been able to make really large gains with the A series chips to date on mobile (not least because they've been able to offload tasks to dedicated co-processors that general-purpose ARM cores don't ship with), the rewards for having chips that are a generation ahead of other like-for-like computers will outweigh the cost of maintaining the software.


Wow, but that cost - it's not a small thing to transition ISA, and don't forget that this transition is one of the simpler ones (more registers, fairly few devices). The risks of transitioning everything away from arm would be much greater.

I guess they have some ISA flexibility (which is remarkable). But not much; each transition was still a very special set of circumstances and a huge hurdle, I'm sure.


At the low-level driver interface, transitioning ISA is a big deal. But I would guess that, at higher levels, most of the work is just changing the target of your compiler?

As in, most of the work occurs in the low-level parts of the Operating system. After that the OS should abstract the differences away from User-space.


No way; not at all.

First of all: there's lots of software that's not the OS. The OS is the easy bit: everything else: grindy, grindy horrorstory. A lot of that code will be third-party. And if you think, "hey, we'll just recompile!", and you can actually get them to too - well, good luck, but performance will be abysmal in many cases. Lots and lots of libraries have hand-tuned code for specific architectures. Anything with vectorization - despite compilers being much better than they used to be - may see huge bog downs without hand tuning. That's not just speculation; you can look at software that's gets the vectorization treatment or was ported to arm from x86 poorly - perfomance falls off a cliff.

Then there's the JITs and interpreters, of which there are quite a few, and they're often hyper-tuned to the ISA's they run on. Also, they can't afford to run something like LLVM on every bit of output; that's way too slow. So even non-vectorized code suffers (you can look at some of the .net core ARM developments to get a feel for this, but the same goes for JS/Java etc). Webbrowsers are hyper-tuned. regexengines, packet filters, etc etc etc

Not to mention: just getting a compiler like LLVM to support a new ISA as optimally as x86 or ARM isn't a small feat.

Finally: at least at this point, until our AI overloads render that redundant - all this work takes expertise, but that expertise takes training, which isn't that easy on an ISA without hardware. That's why Apple's current transition is so easy: they already have the hardware; and the trained experts some with over a decade of experience on that ISA!. But if they really want to go their own route... well, that's tricky, because what are all those engineers going to play around on to learn how it works; what's fast, and what's bad?

All in all, it's no coincidence transitions like this take a long time, and that's for simple (aka well-prepared) transitions like the one's Apple's doing now. Saying they have ISA "flexibility", like they're somehow interchangeable is completely missing the point on how tricky on those details are, and how much they're going to matter on how achievable such a transition is. Apple doesn't have general ISA flexibility, it has a costly route from specifically x86 to specifically ARM, and nothing else.


Extremely aggressive optimizations are really special though, and they tend to require rewrites when new CPU extensions release (and workarounds to work on older hardware). If you rely on super low level ultra-aggressive micro optimizations your code is going to have a relatively short shelf life, different ISA or not.

The vast majority of the code written for any given computer or smartphone doesn't have this level of sophistication and optimization though. I'd wager that for most code just changing the build target will indeed mostly just work.

It won't be painless but modern code tends to be so high level and abstracted (especially on smartphones) than the underlying ISA matters a lot less than in the past.


That's not my experience. Plain old SIMD code ages fairly well - but if you port to a new arch, you will need to look at it again.

This isn't a question of precise instruction timing, it's a question of compilers being pretty bad at leveraging SIMD in general, even in 2020. Also, while I'm sure lots of projects have hand-tuned assembly, even higher level stuff like intrinsics help a lot, and need manual porting.


Absolutely this - even some aggressively optimised stuff like game engines don't have insurmountable amounts of micro optimisations in them.

Also - why would Apple massively up and change uArchs? Even if they did decide to turn Apple Silicon into not-ARM, I'd wager it would look a lot more like ARM than for example, x86 does.


Do it more and more and they'll have the tools to efficiently manage them.

Also likely the small tweaks they will want from time to time should be "easy" to follow internally, if you can orchestrate everything from top to bottom and back.


I doubt Apple is dumb enough not to have basically a perpetual license for ARM.

ARM was launched as a joint venture between Acorn, Apple, and VLSI. I believe that since day 0 Apple had perpetual access to the IP.

They sold all of the ARM shares in mid 90s to prevent themselves from going bankruptcy.

Not to mention starting a JV has nothing to do with perpetual IP access. You will still have to pay for it.


They can certainly have a contract with Arm that allows them to renew their arch license in perpetuity that nvidia won't be able to void.

I obviously don't know that for sure but the idea that Apple would stake their future on something they don't have a legal ironclad position seems unlikely.


I would also agree, the thing is - businesses breakup and come together all the time. If it makes sense and both parties can agree despite of past disagrements and lawsuits, they will partner.

Just because Apple and nVidia has bad relationship at the moment regarding their GPUs is probably orthogonal to what they'll do with this new branch of nVidia, that is ARM.


>Just because Apple and nVidia has bad relationship at the moment regarding their GPUs is probably orthogonal to what they'll do with this new branch of nVidia, that is ARM.

Yea, they got fucked by Nvidia business practices multiple times. There's a saying about shame on you, shame on me, etc. Unless the entire Nvidia business unit also gets replaced in the same transaction, it doesn't matter how much of a faux separation of concerns they want to market.


What need does Apple have with ARMs R&D going forward? They have their own chip designers, build tools, etc.?

True about frenemies the entire time that Apple was suing Samsung, it was using Samsung to manufacture many of its components.


But if your chip heavily builds on arm's IP you need a license for that at least as long as you can't replace the infringing parts of the design. Which sounds very much impossible if you also want to progress on other aspects of having the best chips.

Apple uses the ARM ISA. It doesn't use ARM IP - as in, the design of the chip. Apple designed their own damn chip!

Since they're not branding it as ARM in any way, shape, or form, and they have a perpetual architectural license to the ISA, I suspect they could do pretty much what they please - as long as they don't call it ARM. Which they don't.


During the design were they careful not to create any derivative works of arm IP and/or not to infringe on any of arm's patents?

Apple co-founded the company. I imagine that they have a right to that IP.

Even if not, then yes, I would imagine that a behemoth of IP design like Apple would probably have considered IP rights during that design...


They definitely call it ARM in developer tools and technical documents.

I doubt that will be true going forward. They first used the “Apple Silicon” branding for the Mac and yesterday they used the branding for the Watch and iPad (?)

Curious as I don't know the terms of a perpetual architectural ARM license. But, is it valid only for a specific architecture, say v8 or v9, or is it valid for all future architectures as well? Or is it one of those things, where it depends per licensee and how they negotiated?

> nVidia could cut them off from any new designs

Why would they do that anyway? The downsides are obvious (immediate loss of revenue), the risks are huge (antitrust litigation, big boost to RiscV or even Mips), the possible benefits are nebulous.

Those who are most obviously at risk are designers of mobile GPUs (Broadcom, PowerVR ...).


If they do it that directly, sure. But on a large enough (time)scale, incentives are the only thing that matters. And they'll certainly think hard about putting their fancy new research that would help a competitor into their openly licensed chips from now on.

Nvidia mend. Lol

Applications are open for YC Winter 2021

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact

Search: