š¾ Archived View for dioskouroi.xyz āŗ thread āŗ 29422071 captured on 2021-12-03 at 14:04:38. Gemini links have been rewritten to link to archived content
ā”ļø Next capture (2021-12-05)
-=-=-=-=-=-=-
________________________________________________________________________________
The reason to block the merger is pretty straightforward:
1) Nvidia designs expensive chips and sells them by the millions. [1]
2) ARM designs cheap chips and sells them by the _billions_. [2]
3) Nvidia's "normal" product departments are incentivized to prevent ARM from competing with Nvidia's expensive chips, in various ways. This would hinder ARM's innovation and growth opportunities vs being independent.
I am glad to see this motion to block the merger, for the sake of a competitive market for CPUs, GPUs, etc.
[1]
https://www.techradar.com/news/nvidia-hits-new-highs-in-gpu-...
[2]
https://www.tomshardware.com/news/arm-6-7-billion-chips-per-...
ARM doesn't _sell_ chips. They license the instruction set. Ask Apple or Qualcomm if the final ARM SOCs they build and sell are cheap...
They are responsible through licensing for huge number of cheap chips on the market. It really doesn't matter they don't sell it themselves. If they stop licensing it would have the same effect as if they werse selling them directly and they stopped selling.
Wouldnāt the effect be Apple and others creating a new (and hopefully better) replacement for ARM?
That's the idea behind all the revolutions, that good things will spontaneously come after we destroy something we have. They rarely do.
It's better to create a better thing first, even unfairly, and let it outcompete the incumbent.
ARM is hardly a "good thing". These things rise and fall almost purely by chance since the differences between ISAs are menial at best. If the stars had aligned differently this could very well have been a SuperH world. Or a MIPS world.
What's that about sparc? Or a PA-RISC world, maybe even an AlphaWorld. (Granted, the last two were way less likely than the others)
None of those 3 even tried to have a product for applications with low cost and low power consumption.
The only competition was from Motorola/IBM POWER (and MIPS in some niche products), but neither Motorola nor IBM ever tried to offer low cost products.
During 2000 to 2005, I have ported some products with embedded computers from POWER to ARM.
The porting was not very easy, because the ARM CPUs had much lower performance, but they were much cheaper, so the conversion was done anyway.
The ARM advantage has always been in its business model, which created a lot of competing CPU vendors, willing to give you the best price, not in any technical advantages of its CPU architecture.
The NVIDIA acquisition certainly threatens this business model, despite all contrary claims of the NVIDIA management.
Alpha and PA-RISC died before the smartphone revolution (end 90s vs mid 00s). x86-32 killed them, before AMD64 (x86-64) took off.
Laptops were all x86-32/64. There has been an UltraSPARC laptop (one AFAIK), and Apple switched away from POWER because of power inefficiency. Also before the smartphone revolution.
Intel was busy with Moblin and PowerTOP back in the 00s. This before iPhone.
MIPS and RISC-V also exist. Without ARM it would've happened but slower/later.
ARM singlehandedly made the smartphone happen. No other leading company had anything remotely similar in terms of power draw. Many still don't.
This is ridiculous. ARM was already pretty much well-settled-in in the PDA way before the smartphone, so it definitely didn't "single-handedly made it happen". And during the early PDA era, when they got settled in that market, it most certainly did not win due to "low power draw". Back in the day where SH(-3) PDAs where still a thing, the battery life was measured in days out of 2 AAA cells (even for ARM devices). That was destroyed due to forced backlight usage in most screen techs rather than any change in ISA.
Today MIPS is still used in network devices, SH is still used in things like optical drives, SPARC is still used on some servers, x86 is still used on many servers and desktops and mobile devices, and so on and so forth. Do you think that there is something technical that warrants the choice of these architectures for their usecases ? Do you think that there is something in SH that makes it a better ISA particularly for CD-ROM drives ? Or that there is something in SH that made it a better ISA for some videogame console generations, then something changed and the better ISA was PPC, then ARM, then x86 again ?
It's mostly arbitrary marketing decisions, which manufacturer happened to be doing well that day, and some "historical reasons"/network effects which define these choices. There's very little difference at this ISA/architectural level, and a lot in the actual design and manufacturing level, where ARM is not doing poorly but hardly shines. Apple, Intel, AMD, Qualcomm, even nVidia all do better than ARM there.
If we the wind had blown even slightly differently some day 20 years ago, we may very well be using anything else, even say Itanium in our smartphones. And before you complain that Itanium/EPIC are hardly suited for low power, remember that Transmeta made their money selling _low-power_ VLIW processors emulating x86 instructions. And actually they were lower power than their x86 competition those days. Perhaps Torvalds would still be there rather than making Linux...
Do many smartphones run on Intel, AMD or Nvidia chips?
Qualcomm makes ARM chips for phones.
There were many other CPUs with the same or lower power draw.
The advantage of ARM has always been the cost. For given requirements of performance and power draw, you could find much cheaper ARM CPUs than the existing alternatives.
This happened because there were many competing ARM CPU vendors.
Motorola could give you a PowerPC CPU with the same or better performance and power draw, but they were not willing to lower their prices, falsely believing that there are no alternatives.
Therefore everybody abandoned the other CPU architectures and switched to the cheaper ARM solutions.
I'm sure many companies could do low power performant and cheap chips when high resolution small screens, lithium batteries and large capacitive touch surfaces matured enough to be harvested.
The thing is only Arm did actually designed appropriate chips at the right time and licensed those designs in the best way possible. Arm was IBM PC of the smartphone era and I don't understand how people might not appreciate that.
Actually, it's kind of the other way around.
When Nokia and Ericsson switched to ARM the company was probably saved from bankruptcy.
"just create a new instruction set"?
It's taken the better part of a decade for workable alternatives to CUDA to emerge.
I hope youāre not calling ROCm a workable alternativeā¦
itās really not close. I love that they are trying, I just honestly feel they need to be trying with 10 times as much funding/resources behind them.
Or do you mean Appleās Metal Compute? useful only with inconvenient (at scale) Apple hardware.
If workable CUDA alternatives were really here, I feel people would be talking about them more given how large the hate is for Nvidia on multi fronts from multiple different user groups.
I was, but mostly through XLA or Vulkan layers. Also, Iād define workable as a pretty low bar.
Thatās fair, for me it never worked with any of the hardware or software I wanted it to and Iām not enough of a GPU expert to really fix that myself, so for me and many others I found with similar problems it definitely has yet to become a workable solution, but I was probably a bit harsh if I came across like I thought it didnāt work at all, Iāve seen it demoed and I know it works, but for such a restrictive set of hardware and software that Ive never seen it in use myself outside of demonstrations.
As for Vulkan, I tend to think of CUDA at the driver level given how closely coupled it is to Nvidiaās hardware. So I wasnāt really thinking about broader cross compatible APIs like Vulkan which tend to get implemented on top of the drivers providing low level access like CUDA givesā¦ but with the exception of Apple (because we know they will never change their mind) Iām hoping the broader industry gets behind Vulkan compute shaders enough we finally get something that delivers on the promises made when the OpenCL effort began. I want clean understandable abstractions over the top of all the different SIMD and MIMD capabilities we have these days it doesnāt have to magically compile my code down to FPGA hardware but how about actually being able to get the best matrix multiplication performance out of both my CPU or GPU or both if I have a CPU with the right kind of integrated GPU that they can efficiently enough share memory and both work together without slowing each other down when working on an embarrassingly parallel matrix multiplication task. Is this really too much to ask of software in an an era that gives us near magical JIT performance in multiple languages, and manages to build cross compile/transpilation tools that can convert entire assembly code bases into JavaScript and all the other nice things we have as software developersā¦ it never felt like too much to me but the fate of OpenCL appears to disagree.
I never understood the dominance of CUDA. I did GPGPU more than half a decade ago, and I feel like stuff has hardly changed. There are thread groups, group shared variables, global, and group shared atomics etc. Underneath, all GPUs run the same-ish wide SIMD architecture, so I'm pretty sure most stuff is semi-performance portable.
DirectCompute, OpenCL and CUDA and probably all the others expose this same programming model. Why is CUDA so dominant?
_> Why is CUDA so dominant?_
nvidia decided they wanted to be the market leader, and spent $$ developing tools, training materials, and libraries (like cuDNN) which they gave away for free.
Other GPU vendors chose not to outspend them.
They also hire software engineers to integrate CUDA into open source projects. On the one hand, those projects can now take advantage of advances in hardware. On the other, users will have to buy NVIDIA to use it, so how open is it? Also NVIDIA will not provide any help with CI, so those costs are borne by the project ā or the code isn't tested.
Which alternative? When it comes to ML there is no alternative to CUDA.
JAX/ XLA seem quite popular, especially since you can use them in Colab on TPUs. IIRC, there are people who have managed get jax compiled with ROCm support.
Tensorflow lets you deploy to quite a few backends.
oneAPI support in GPUArrays.jl seems to be coming along.
By that logic there will never again be another workable alternative. Why not start now rather than wait a few decades?
Innovation is not an on-off switch. Market forces create an incentive to _attempt_ to innovate, but they don't magically make innovation happen.
There's nothing that prevents you from doing that as it stands.
It's likely that there are many new designs being spun up everyday by various people, but that doesn't mean that we want to remove competition from the market.
Apple has a very broad and safe license for their Designs. They would largely be unaffected by the merger and likely not change anything.
That said, the replacement ISA already exists IMO: RISC-V
Apple supposedly has a indefinite 'architectural' license to design and manufacture ARM chips. Presumably that cannot be revoked.
Considering they've even gotten away with breaking the rules (custom ISA extensions), I imagine nothing ARM does could realistically cause Apple to be unable to continue manufacturing ARM-compatible chips.
Yeah, turns out when you're one of the founders of ARM you get a pretty broad license!
I would imagine that Apple (which put money into Acorn and ARM JVs over thirty years ago and I _think_ was the first third-party company to do so, by a very wide margin) has a much older license than anybody else.
Does that also include access to new versions of ARM not yet released?
Why would Apple even care about those?
It's not correct to say they just license the insn set. They license ready to use cpu and gpu core designs (like the Cortex family in most Android phones).
It depends on the customer. Apple and Nvidia indeed license just the ISA and build their own CPU implementation from scratch. At the other extreme, you can make a complete chip just from ARM IP - they have way more than CPU and GPU IP, they can sell you pretty much everything else, too. At the middle are chip makers taking eg ARM CPU IP but no other ARM IP and making other IP themselves or licensing from other vendors.
ARM also licenses CPUs, GPUs, and just about everything that one would need to build an SoC.
And those elements are used to enable the manufacture and sale of cheap chips. Semantics aside, ARM is _the_ critical player in several market segments.
I don't see that that really changes anything here.
Well the argument is that ARM is cheap and good for consumers because it's cheap. I wanted to make it clear that you can't go to ARM and buy <cheap arm chip>. You have to go to Qualcomm or Apple or Broadcom and buy <not quite so cheap arm chip>. So that kinda impacts the argument.
Whatever Nvidia wants to do with licensing they have to combat market forces already in play. There's a market for somewhat less expensive arm chips, for one, so they'd lose that revenue. Two, if they want to royally fuck arm shops and become the only ARM vendor, well, good luck... you just killed the ISA you bought for 40 billion...
40 billion isn't much to sabotage the vast majority of your competitors
But we have pi and many others. If all these are slowly killed may be they moved to more expensive chips.
The business model does not fit.
Depends. Apple does not sell chips. Apple sells a vertically integrated platform that has many components from cloud services to physical devices. The price of one component in the physical device is not that important.
Many companies licensed ARM ISA. If Nvidia wants to hinder competition by only licensing them low performance cores, isn't it still possible for them to design their own cores?
> isn't it still possible for them to design their own cores?
Most companies license the cores, so they can build ARM-designed cores and integrate those into an SoC.
In order to design your own core, you need to license the ISA itself, and thatās a wholly different kettle of fishes: there are only about a dozen architectural licensees, and not all of them license both 32 and 64b ISAs.
Arm is British, the FTC and Nvidia are American. Can someone that's not a technocrat explain to me the problem here?
Why don't we have newer chip companies?
If barrier of entry is the issue isn't that what people should be striving for?
Looking at GPU cost and availability, I agree with you.
This is the most megalomanic thing I've ever heard of. ARM has high chances to be both the next x86 and the most common embedded architecture. While at the same time Nvidia GPUs are already the de facto standard for GPU computing. So Nvidia would have monopolies in 2 verticals. No reasonable and competent antitrust commission could ever allow that - although probably everyone was betting the latter wasn't the case...
I totally agreeā¦ whatās interesting is that some of the reasons that FTC provide are totally not whatās interesting (self driving).
I have a MacBook Pro with M1 ARM chip, and I love it, itās extremely power efficient, I donāt really have to think much about charging it. It wouldnāt be possible if NVIDIA could force its chips on Apple.
Apple has a perpetual ARM architecture license. Nvidia wouldnāt be able to force chips on Apple at all.
This is of course related to the fact that ARM was initially a joint venture between Acorn, Apple, and VLSI Tech way back in 1990. Apple was a major early investor and has been using ARM ISA chips in products since the 1990s.
If anyone is curious, third parties can license specific designs and drop them in their chips, or architecture licensees can design their own chips, like the M1, based on the instruction set architecture. There are around a dozen or so companies with an architecture license.
But I believe there is a new instruction set released by ARM every couple of years. Would Apple get automatic access to say ARMv10 architecture because of their architectural license? Or would they have to renegotiate with ARM (or NVIDIA?)
If they have to renegotiate with NVIDIA to get access to ARMv10, it would be quite nasty if NVIDIA decided they wanted a bigger piece of Apple's pie, especially if Apple has developed competitive GPU technology that made NVIDIA's somewhat obsolete.
I haven't read the license of course, but what I have read from others is that Apple has a perpetual license that gives them access to all future architectures as long as they fulfill their side of the deal (fees, etc.). That deal is with ARM, so if someone buys ARM the purchaser is still bound by the terms of existing deals.
Seems like a good reason to block it. Apple could only do it with a perpetual license. Competitors would (potentially) have to abide by rules that Apple avoids.
Several companies have the same 64-bit architectural license, including
Nvidia,AMD, Qualcomm, Samsung, Broadcom, Cavium, HiSilicon, Applied Micro, and Apple. There are probably some others.
How long can you use the M1 laptop on single charge?
I go about 2-3 days without charging it under moderate (regular) use. While developing (compiling, lots of tabs, etc.), Iāll use about 80% in a day.
For reference, the last zoom call I was on (2 hours) with nearly max brightness only drained the battery ~5%. My Intel MBP wouldāve been down at least 20%.
Not just that; how long does it last while idle/sleeping? Pretty much forever; the MacBook Air I have for random OSX tasks and testing Linux on sporadically goes over a week between charges.
These things are very power efficient while actually computing, but they are _extremely_ power efficient while idle. The whole SoC uses a couple dozen milliwatts while in active idle state (e.g. screen off, machine up and responding to pings via WiFi).
https://support.apple.com/kb/SP824?locale=en_US
The official number is 17 hours, bu I never had to test it in real life. The crazy thing is charging to 50% in 30 minutes, and 50% is enough to get me through my trips.
That's great. Yeah, once the on-battery time is slightly greater than one workday requires, increasing it further stops being interesting.
For battery longevity, in general I would charge to 90% and then use it until 20%, then recharge to 90% again. This makes the most out of the battery charge cycles, and avoid the increased stress of close-to-0% charge (bad if there for longer time) and close-to-100% charge (less bad but still better not to go there often). But maybe M1 battery system is different for some reason, I don't know.
Itās not different. In theory recent macs are supposed to ālearnā about your habits and lock charging at around 80% when youāre not going to move a while (āoptimised battery chargingā), but from what Iāve heard itās not super reliable. And Iāve never seen it work properly on my iphone, whose usage is about as reliable as you could be (plus iOS has all the alarms so it knows exactly when Iām going to wake up) so I didnāt exactly have high hopes.
A more reliable solution is to manipulate pmset directly, or use something like AlDente which handles it for you (the paid version has lots of extra features but the free one suffices to avoid charging above 80% by default).
Obviously when charging time arrives you need to realise yourself and plug in somewhere between 20 and 40%.
The only annoyance is the magsafeās LED remains amber even when the mac stops charging at 80%.
On my iphone it pretty reliably charges up to 80%, waits a few hours and then charges to 100% when I put it on the charger for the night.
On my m1 air Iāve only known it to charge to 80% after a period of several weeks where it was docked. As soon as I started using it away from power for significant amounts of time it went back to charging to 100%.
> On my iphone it pretty reliably charges up to 80%, waits a few hours and then charges to 100% when I put it on the charger for the night.
I just checked what happened last time. I went to bed around 2300, it quickly charged to 80% (getting there a bit after 2400), then stayed there until 0200, at which point it decided to charge to 100 which it had reached by 0400.
As it does every workday, and has since before I had that specific phone (which is 2 years old), the alarm woke me at 0600.
So over a 7h night it managed to spend more time at 100% than it did at 80. Iām not going to say Iām impressed, because Iām not. Based on usage pattern I _could_ see it fail to reach 100% by 0600, but I donāt think Iāve woken at 4 once in the last 3 years.
>The crazy thing is charging to 50% in 30 minutes,
There is zero crazy about that part, it's just 1C charging which is not even extra hard on the batteries.
I use an M1 Air for my daily development, and under full CPU load the battery lasts about 2.5 hours, while under no load and low screen brightness it will last for 15-20 hours.
Was that an artificial load test? The heaviest continuous load Iāve put my m1 air under was playing one of the tomb raider games, which puts a heavy load on cpu and gpu, and it still lasted 3 to 4 hours on battery.
But you canāt play call of duty on it.
(Important, not just trolling)
I've been using Win11 ARM via Parallels to play older 32bit games and some newer 64bit games. It's enabled me to continue my addiction to playing on P1999 Classic Everquest Server. At the moment the main problem has been a lack of GPU power (M1 Mac Mini). Although some games run into larger issues, I assume, with the GPU itself like Borderlands 2 & 3 having a stutter every 30 seconds or so regardless of graphical settings.
Although I imagine the problem with a lot of newer crap will be draconian DRM measures freaking out at being ran in Parallels. Which are games I will never buy.
Its coming
https://boilingsteam.com/box64-can-now-run-crysis-on-arm-wit...
>(Important,
For whom ?
Strategically, why wouldnāt the US govāt want a US company owning ARM? Why is the FTC pilling on to block this? How would I, as a consumer, be harmed by a graphics company owning an instruction set and design specification?
Stopping two US companies from merging because it would create a monopoly and negatively impact US consumers is one thing. But this isn't even remotely the case here? Doesn't the US government get more tax revenue if Nvidia makes some extra dough charging royalties for that IP? Iām confused.
The press release directly and prominently answers this question multiple times. It suggests that under Nvidia the new Arm would be disincentived to innovate in any area that negatively affects Nvidia businesses. Given that Arm is a critical supplier for many industries, this could result in consumer harm via āreduced product quality, reduced innovation, higher prices, and less choiceā
Also, I donāt believe the FTC mission is to increase US tax revenue, so that aspect seems irrelevant.
> the new Arm would be disincentived to innovate in any area that negatively affects Nvidia businesses
If the standard is "they don't compete against each other now, but would be disincentivized to compete against each other in the future," that could be used as a rationale to block literally _every_ merger.
_> that could be used as a rationale to block literally every merger_
Don't you threaten me with a good time. Imagine, companies actually forced to collaborate via the ordinary mechanism of contract arrangements rather than by assimilation into a horrifying megaconglomerate.
A merger is a contract :)
The problem is when some companies form a cartel of sort (with our without contracts) anyway. The same way Intel was not licensing x86, but since AMD already had it they were both incentivized to keep users on x86 and keep arm out.
(Of course the nvidia-arm acquisition is more about nvidia trying to build an army, oh pardon, portfolio. As Intel and AMD already have. They acquired Mellanox, plus they have GPUs and CUDA chips. Intel and Apple kind of have everything of course. AMD has Radeon, but AMD is probably still trying to catch up financially. There's also Qualcomm with their snapdragons and modems. And there are probably others, but the important thing is that snapdragons are arm based. And while selling high-end ML and gamer/miner cards is certainly not bad, but nvidia is probably trying to expand into a different market too, let's say high-volume but older semi tech.)
Indeed, but while most contract arrangements need to be explicitly re-upped, a merger is permanent; a subsumed company cannot un-merge itself, because it no longer exists as an independent entity. This fact incentivizes both sides to remain competitive, because each side knows that they are replaceable if they don't keep costs down and quality up relative to their competitors. Furthermore, while a cartel is the worst-case scenario between two independent companies, cartel-like behavior is the _only_ scenario for two merged companies; it's unfathomable that any subsumed company could decide _not_ to do whatever its new owner tells it to.
How is SoftBank selling ARM? (And how did Google acquire Motorola and then sell them off to Lenovo, for instance?) Is there a difference between how they acquired ARM and how Nvidia wants to acquire ARM that makes it a merger instead of an acquisition or something? I honestly don't know and am curious and think there's an interesting point here if how I'm interpreting your comment is correct.
> that could be used as a rationale to block literally every merger.
Even though that would hurt me businesswise I would happily sign for that future. And while we're at it: make it illegal for one company to hold shares in another. Just natural persons holding shares in companies would be fine.
This is entirely impractical. Companies wouldnāt even be able own subsidiaries in other countries that can legally act as employers or tax payers there.
It would make it impossible for governments to own minority stakes in local subsidiaries of global corporations to veto some things. It would make it impossible to impose capital requirements on subsidiaries of big financial institutions. It would be impossible for larger competitors to rescue failing peers and keep them operating as a going concern.
And it would very likely create the most suffocating oligopolies on a national level as well as on an industry level as competitors would find it extremely hard to enter markets with entrenched incumbents if the only way to do it was to build from scratch.
> Companies wouldnāt even be able own subsidiaries in other countries that can legally act as employers or tax payers there.
So? That would enable local companies that occupy the same niche to do so. All these multinationals are not necessarily a good thing.
> It would make it impossible for governments to own minority stakes in local subsidiaries of global corporations to veto some things.
Only if you believe that those local subsidiaries should exist in the first place and that need not be the case.
> It would make it impossible to impose capital requirements on subsidiaries of big financial institutions.
There would be no such subsidiaries.
> It would be impossible for larger competitors to rescue failing peers and keep them operating as a going concern.
These could be structured as asset sales. And a whole lot of trickery that causes these peers to fail would go out the window.
You are making the mistake of looking at the glass 'half empty', when in fact the better way to look at it is what we would get for it in return. Capitalism, but with a much more direct link between UBO and the companies they have a hand in, far less opportunity for nation state level wealth to end up concentrated with a very small number of people and less opportunity for companies to play shell games with their income.
Capitalism, like money is a great invention. But like everything else when taken to extremes it is a net negative, in moderation it could well be uniformly good. But the whole externalization game needs to stop or it will harm us greatly.
So what you're saying is that companies should never employ people or put capital to work in more than one country or do anything else across borders that requires a legal entity that can be locally regulated, taxed and held accountable.
All failing companies should be liquidated and sold off in bits and pieces, employees fired, contracts and debts voided, pension schemes closed.
Entrenched local incumbents could only ever be disrupted by individuals starting new companies, funded exclusively by other individuals. Presumably, none of these individuals would be allowed to own a controlling stake in another company, otherwise it would effectively be one company.
You would have to convince those wealthy individuals to start companies in small countries in spite of the fact that the growth potential of their company would be far more limited than if they started a company in a big country. The only incentive would be local protectionism.
You would also have to make sure that companies don't enter into contractual relationships that effectively make them act as one (such as McDonald's franchises)
You would effectively hand absolute power to local family clans and oligarchs closely intertwined with local bureaucrats and politicians. It's a recipe for stagnation and corruption.
This is not a glass half full and it's not moderation. It's a glass smashed in an act of vandalism and blind rage. It's the sort of revolution that countries take a century to recover from.
Turning the clocks back a couple of centuries is not progressive either.
> So what you're saying is that companies should never employ people [in more than one country]
No, that's not what I'm saying, besides that could be easily covered through employment law, but we already have the concept of internationally operating free-lancers.
> or do anything else across borders that requires a legal entity that can be locally regulated, taxed and held accountable.
Not if that legal entity is going to have a company as its shareholders. But there would be no reason why it could not exist and have natural persons as its shareholders, they could even be the same shareholders as in some other company abroad.
You are still arguing about this from the perspective of everything that it can't do, which is pretty easy to deal with because most of those things it can do. What it can't do is to hide the UBOs, which is really the only big change. Everything else can easily be worked around.
_>You are still arguing about this from the perspective of everything that it can't do_
No, I also told you what I think it would do. Create a corrupt local oligarchy that is extremely hard to disrupt. It would create a mediocre, low productivity, stagnant economy shaped by the most extreme form of protectionism I have ever heard of.
_>Everything else can easily be worked around._
Your use of the word "easy" for something that would require a root and branch redesign of company law, tax law, employment law, competition law and scores of other laws in every country on earth tells me that you really haven't thought this through. Not to speak of changing the ownership structure of tens of thousands of existing companies.
Most shares are held by companies: either pension fund, to create your ETFs, mutual funds, or holding for you in normal broker account, or financial institutions, to provide liquidity, to put to work the pension fund huge inventory (most hedge fund short-selling is done with pension funds' dormant inventory using an investment bank middle-man), or by companies needing to store money somewhere other than cash.
I don't know many individual directly handling shares with the exchange and the settlement. How do you avoid the enormous volume of trading that happen due to retirement account reuse for shorting ? Would you forbid that too, removing the incentive for company to be honest because nobody can short anymore ?
Arm and Nvidia are in direct competition today. ARM was 39% of the tablet GPU market in 2020 for example.
NVIDIA's chips count as part of that ARM marketshare, and in fact ARM itself does not manufacture chips at all, nor does NVIDIA design CPU ISAs.
The grandparent was talking about GPUs. NVIDIA and ARM (Mali) both design GPUs.
And they also both design CPUs. ARM just doesn't manufacture theirs themselves.
> If the standard is "they don't compete against each other now, but would be disincentivized to compete against each other in the future," that could be used as a rationale to block literally _every_ merger.
Try it this way: How many _other_ companies are _likely_ to compete with them in the future? If the answer is a thousand, the loss of potential competition is not very significant. If the answer is less than ten, it is.
They compete more than ever. Google, Amazon, FB, Apple are all developing AI, CPU and GPU chips that compete directly with Nvidia using ARM
You know why? Because working with Nvidia has become such a pain in the ass. That it's better just to do it yourself.
They charge you 10x for a gpu that's a little better than their consumer GPU but won't sell you the consumer GPU because they want you to pay 10x.
Now they will definitely do the same with ARM licensing.
How do they stop companies from buying consumer GPUs? I have used some minor firmware tweaks to unlock full commercial capabilities out of Nvidia GPUs in the past. Perhaps that is against the tos and a large company couldn't get away with it like an individual can?
> Perhaps that is against the tos and a large company couldn't get away with it like an individual can?
A while back they made a change to the terms of the GeForce driver, barring it from deployment in datacenters except for blockchain processing.
https://www.datacenterdynamics.com/en/news/nvidia-updates-ge...
https://www.nvidia.com/en-us/drivers/geforce-license/
Wasn't the ostensible justification here making the cards more available for the enthusiast gaming market they were (to some extent) intended for?
That's an honest question, I can't fully recall.
If you're Google/Apple/etc, you don't want to be relying on some unofficial firmware tweak where the next driver update could completely brick your business and force you to never update the driver again.
Market power, among other things are the requirement. No one cares if mom and pop store buys its neighbour. Thereās a whole specialty called competition law.
I wonder how this standard would have been applied to past mergers?
https://en.wikipedia.org/wiki/List_of_largest_mergers_and_ac...
It should have been.
CVS buying Aetna was a particularly sick joke.
Don't both ARM and NVIDIA design CPUs and GPUs? Seems like they compete against each other now.
And why is SoftBank any more positioned to be a champion of ARM quality? I just donāt see how it would not be in Nvidiaās interest to promote a healthy ARM any more or less than any other owner. Is the goal for ARM to become independent again? Are Nvidiaās competitors making ARM graphics cards and SoCs?
It does not need to be. Softbank owning ARM is the status quo, the devil we know, and they have a track record of not running ARM availability/quality into the ground. They don't seem to have motivation to do so. Also, they are invested in telecommunications, so having healthy and strong ARM position for mobile devices is in their interest.
NVIDIA, on the other hand, is suspicious. It is in their interest to use and manipulate ARM to release/restrict products in such a way that will strengthen their already strong market position with accelerators, and push down its competitors like AMD and Intel.
Intel and AMD already have a monopoly on x86-x64, and both make GPUs and CPUs. Nvidia only makes GPUs.
They were terrible at the CPU game with their Tegra line. Not sure if that's still going on.
I'd rather Nvidia be kept away from Arm at all costs.
Or, they realized they're terrible and instead of trying to be terrible again they're trying to be not terrible this time by acquiring people that actually know what they're doing? A good ARM Tegra sounds like a product that would benefit the market and thus consumers.
Terrible enough to power one of the best selling consoles of all time despite its shortcomings.
Nvidia's Jetson series computers and their various SOCs (as seen in the Nintendo Switch (Nvidia Tegra), Tesla's cars, etc) are not just GPUs.
It is not possible for 2 competing companies to have a monopoly. Perhaps you suspect some sort of collusion between them? Seems unlikely given the kicking Intel have had the past few years.
AMD has barely beaten Intelās Sandy bridge, finally. Not enough for me to buy one.
They donāt need to collude, theyāll ācompeteā by siting on their ass like Intel did, barely being better yet jacking up the price, only āHaswellā is renamed āZen 3ā.
Wait a second, isn't Ryzen 5xxx much faster in single thread than Sandy Bridge ever was? Isn't the available number of cores greater as well?
> Intel and AMD
> monopoly
Pick one?
AMD decides it wants to make something like the TX1. Do you think that could happen if Big N owned ARM? Would the M1 series have progressed as quickly knowing the bad blood between Apple and Nvidia? Softbank probably isn't the best management but its neutral.
If ARM is truly too important to _the economy_ to be owned by a āselfishā business, shouldn't we just declare it public domain and move on? Why canāt Nvidiaās competitors who theyād try to abuse and squeeze go use some other ISA? Is ARM really _that good_, or is it just popular?
It's not about the economy, it's about competition and consumer harm from the lack of it.
As much as it's easy to forget, the free market is supposed to be a tool for the benefit of consumers (through both price and innovation), and while it's not very common, every once in a while the government actually looks at that as important and steps in.
It hasnāt been common recently, under Obama/Trump. The Biden Admin FTC seems like itās serious about using antitrust laws to prevent the market from being too lopsided in favor of a few big players.
More like Bush II/Obama/Trump. The FTC has been neutered for a while, particularly when the rulings about decreased competition vs consumer choice were made in the mid-2000s.
The reason for competition law is to regulate the free market to ensure that consumer choice is not limited by the uncontrolled acquisition and merger of suppliers.
The merger of NVidia and ARM is inherently non-competitive unless NVidia accepted limitations/restrictions on their licensing freedom which they are not willing to propose or accept.
That's leaving aside the international and other ramifications of a company that is involved in the entire vertical of chip production and sale through to retail having control of the licensing of the most used ISA and other IPRs that are available to its competitors throughout that vertical.
A company doesn't need to be owned by the govenrment (which I assume is what you mean) for the government to guarantee it is free from being bought out by it's competitors. Whatsapp/Instagram were great examples of this - where the leader in a sector managed to buy its primary competition. As far as I understand it no one can credibly claim Instagram's growth has been prioritized relative to Facebook - why? Becuase Facebook prioritizes facebook.
This isn't a question of "will the economy collapse if X buys Y" it's a question of "Which will yeild the best outcome, Y being acquired by X or no". The exact definition of "best outcome" is where the entire battleground though.
It is quite difficult to see who actually could acquire ARM though - because of ARM's quite unique position, any of the potential acquirers are either customers (Apple, Nvidia, Qualcomm) who would be vertically integrated to beat out competitors or direct competitors (AMD, Intel). I think almost the only people who could buy ARM at this point would be TSMC.
For the market to stay competitive, ARM must stay independent - we have enough oligopolies and monopolies already
Why does ARM need to be acquired by someone? If Softbank wants to unload it why can't they do an IPO?
Mostly because Jensen Huang was willing to pay a nice premium over what the market would otherwise offer. There are some _cough_ synergies _cough_ that Nvidia could have instrumented for higher profits, Jensen had a clear and enthusiastic vision about how to get there.
Softbank can still sell ARM to anyone including the public market, they just won't get as much money for them that way. Presumably.
Yes to both, if itās not competitive theyāll use RISC-V or another architecture.
Changing architectures is no small thing regardless of how many times Apple has done it.
Youāre right. I donāt even count the new M1 as an architecture change, they switched to mobile like how C2D came from pentium M, their software market in ARM was already mature and it would be significantly harder. At the same time, they made Rosetta and switched to x68 and ARM around the same time, only ARM won.
Why does ARM need to be owned by anyone? If they are profitable, then there's no need to participate in M&A except for polishing C level resumes.
Because Softbank made some really bad investments and needs money. (It made 12 billion loss with a 57 billion revenue last year)
I guess the other option is to list it on some stock exchange.
Exactly, last i checked the important part of free market was producing products consumers enjoy and generating revenue, not corporate scheming and strong-arming competitors
> Strategically, why wouldnāt the US govāt want a US company owning ARM
It's worth noting ARM Ltd is a British company currently owned by a Japanese company. The U.K. and Japan are two of the closest, if not the closest, defense and industrial partners the U.S. enjoys. And unlike many U.S. allies, they're more-or-less close by choice. IOW, there's a deep reservoir of trust across the spectrum--military, legal, political. Both the U.K. and Japan tend to exercise their independence far more freely than other U.S. allies precisely because of the mutual respect afforded among the three. There's much less tension and apprehension among those three than as between, say, the U.S. and France. A critical supplier like ARM being in the hands of the U.K. and/or Japan is good enough from the perspective of the U.S., absent some extraordinary complicating context.
You dont think this had anything to do with it then? ;-)
"When Britain Nuked America....Twice!"
https://www.youtube.com/watch?v=-Wx6npt421c
And this?
"U.S. satellites are being attacked every day according to Space Force general (thedrive.com)"
https://news.ycombinator.com/item?id=29402923
I think its a good decision because just being dominated by US tech companies can be demoralising, besides no one has a monopoly on ideas or innovation which all develops at different rates.
And if ARM did get bought up, would this also drive other countries towards China who are developing their own CPU's?
"How China plans to lead the computer chip industry"
https://www.bbc.co.uk/news/business-50287485
From a US Military perspective, you don't want to pee off your military assets/partners do you?
Wow, fascinating video.
> _why wouldnāt the US govāt want a US company owning ARM? Why is the FTC pilling on to block this?_
The U.S. government isn't a monolith. FTC is charged with, first and foremost, maintaining our markets.
Also to add:
Allowing nVidia to become a monopoly would bite the US in the ass strategically, as nVidia loses incentive to innovate.
That's how the US, effectively, ended up having no viable aircraft when it entered WWI, after being the country that invented the thing. (Wright and Curtiss, IIRC, locked the market with patents).
Or, more closely, how the US telecom/Internet infrastructure is atrocious, despite ā and because of ā the Internet and the telephone being invented here. Go figure, Ma Bell wasn't the bees knees.
Strategically, the US needs someone to keep nVidia up on their toes. Innovation cuts into profit margins when you have a monopoly.
Ma Bell made some great innovations here and started a lot of the computers tech, after it was destroyed the baby bells combined again, Cingular has exclusive rights to iPhone at a time.
Nvidia is reasltically already a monopoly with CUDA. They donāt need to innovate and they force universities to use their more expensive GPUs.
> They donāt need to innovate
They're constantly innovating, and have monthly software releases. Don't blame NVIDIA because AMD can't offer a solid enterprise offering that is attractive to customers.
They donāt need to for CUDA. I didnāt say they didnāt. Iād love more open ARM devices like Jetson.
> _Cingular has exclusive rights to iPhone at a time_
But not the other baby bells, which is the point. Also, Apple went to Cingular after being rebuffed by Verizon. (Apple wanted no crapware. Verizon said no.) Competition working.
Is your point that monopolies aren't bad, or that anti-monopoly laws are useless?
Because both of these are provably false.
As is that nVidia is already a monopoly.
Natural monopolies arenāt bad, donāt last long and they fall when itās leapfrogged, like Kodak when digital came. The USPS is an unnatural monopoly which stifles any carrier and serves mostly to distribute spam.
Anti monopoly laws are useless, x86-64 is a duopoly, and any monopoly laws arenāt able to stop Amazon and their vertical intergration, because the argument is that eBay and other sites exist. Network oligarchs collaborate to fix prices.
> The USPS is an unnatural monopoly which stifles any carrier
The USPS is required to deliver to everyone, even when it doesn't make monetary sense to do so. They are an important piece of US infrastructure.
They are also incredibly constrained by legislation. They aren't allowed to just raise prices as needed.
No itās not. You can do it digitally for official documents, and they mostly deliver spam, and the cost is in your taxes. Iād rather not have them.
Last time I checked there was no official government ISP the was required to provide some baseline of service to every physical address in the country at a reasonable price. And to keep it equivalent you'd only pay for bandwidth used - no flat fee for the hookup itself. The internet is not (yet) a replacement for what the USPS actually is.
USPS doesnāt serve everywhere either. Where do you think that inefficiency is paid for? Why are taxes subsidizing spammers that you canāt even refuse? What if you only had government email that was full of spam? What if it was physical spam that taxes pays for that follows victims around the US with no way to stop it? Is that what you support?
The USPS is self-funded by those periodicals. Your taxes donāt go to USPS
https://www.washingtonpost.com/business/2020/04/15/postal-se...
> The agency is burdened by hundreds of billions of dollars in debt and falling revenue, and Congress and the White House have signaled an unwillingness to grant more funding without major restructuring, lawmakers say.
> The law envisioned the Postal Service as a self-sustaining agency whose revenue could cover the expenses associated with an aging workforce involved in a physical occupation: delivering packages and parcels to every address in the country.
> Not even two decades later, it canāt. The Postal Service has racked up $160.9 billion in debt from whatās owed prepaying retiree benefits. On top of that, it has many yearsā worth of operating deficits, as its top revenue generators no longer covered the costs of delivering the mail.
Then why do they need bailouts? Are these bailouts self funding? What money gave them that credit when they were long unsustainable? This is an unnatural monopoly that doesnāt allow competitors and also is a failure who mostly delivers worthless physical spam you cannot escape.
There are places where it delivers where there is no internet.
At what cost? What benefit is sending non refusable spam to the ends of the US to anyone besides spammers? You donāt accept spam email do you? Would you support it if it physically existed and your taxes paid for it, including costly methods like sending it to the middle of nowhere?
If governments primary means of working is paper paper must be able to be delivered to every household.
"Natural monopolies arenāt bad, donāt last long and they fall when itās leapfrogged"
Natural monopolies are things like electric power grid and water supply and train and roads, they last forever precisely because they are natural and they were never leapfrogged in history. 'Natural" part stands for nessesary inveatment in infrastructure being redundant and causing zero-sum game.
Kodak wasn't much of a monopoly and certainly wasn't a natural one
Read up on how antitrust actually works, your assesment of USPS is equally inaccurate
You listed infrastructure, Iām referring to examples such as standard oil, which sold oil the cheapest.
That's not a natural monopoly
I wish we could run experiments on this sort of thing and compare. I want to know how an alternative "no IP law" USA would have developed. Or even one that just didn't grant patents as freely.
My totally uninformed opinion is that basically the only IP law that's worth a damn is trademarks (which should absolutely be very strongly enforced. It's crucially important that consumers know who they're buying from for the market to work properly), and _maybe_ copyright on /entertainment/ focused creative works for the life of the author and no longer.
Look at the US during the 1800s freely "borrowing" stuff from the UK, which was then the world leader in IP development.
All "developing" nations don't (and shouldn't) GAF about IPRs during that phase in their development. Notice that China didn't GAF for years, but does now, both from a control POV and from a competition POV.
Countries only worry about IPRs when they have sufficient IP development to have something to sell.
Realistically, with the cost of developing this stuff only going up, only monopolies or duopolies are really going to exist at the cutting edge. That's the trend and I don't really see a way of reversing it.
Cost of developing what? New ISAs? Chip Designs?
Their argument seems to largely be that ARM does not produce chips directly, but instead licenses designs to other firms using a "neutral, open licensing approach." They feel that this will stop should Nvidia acquire ARM.
The also FTC contends that ARM induces competitive behavior in Nvidia. And a merger would stifle that competition.
Additionally, and probably most importantly, the FTC contends that Nvidia's competition shares sensitive information with ARM. And part of what Nvidia is looking for with this acquisition is this information. I suspect this is the true reason behind the lawsuit; there's probably a good bit of industry support behind it.
Personally, I'd rather ARM be owned by a massive American firm. And out of all the American firms who would be interested in ARM, Nvidia is the most likely to continue to innovate, rather than merely engage in rent seeking behavior.
Why would you want ARM owned by a "massive" American firm. Is there one that you would trust to keep the evolution of ARM IP at the same rate, while also offering IPRs (ie licensing etc) in a way that continues to encourage innovation in the use and evolution of that IP?
Massive firms are not an inherent Good Thing. There are some benefits due to their ability to invest, but there are also multiple downsides to conglomerates, particularly in strategically vital industries and areas.
Well, it's not that I _want_ it owned by a massive firm, but the fact is that only an extremely large company would be in a position to buy it. I was primarily pointing out that, I'd personally favor it being owned by an _American_ company, because the most obvious alternative is a Chinese one. I can't think of many British/European technology companies who could afford ARM and could justify the price premium.
The alternative would be to go public again. But I think the temptation is too great, and some massive company would take it over. ARM is the kind of company which is much more valuable as a part of a conglomerate than it is as an independent company.
That was my exact reaction. I'm not sure how much I like Nvidia acquiring Arm, but on some level as an American my instinct is to encourage it.
If it is the FTC's actual position that the deal would harm consumers or the industry as a whole, it's certainly admirable that they would ostensibly prioritize that over US strategic interests.
This makes me wonder if their analysis shows that the merger would do sufficient harm within the industry as to actually run counter to US interests. If ARM is shaping up to become a pillar of the Western world/economy while China and its sphere of influence consolidate around RISC-V, then anything that harms Arm's market position is also a geopolitical risk to the West. The US government pushing for such a merger, at a time when China is investing heavily in semiconductor manufacturing capabilities while eyeing a conquest of Taiwan/TSMC, would therefore be shooting itself in the foot. Better to grow the pie than risk blowing it up for a slightly larger slice.
I don't think it's the US strategic interest for Nvidia to own ARM. International corporations don't have any particular loyalty to the country where their corporate HQ is located, and fewer semiconductor companies just mean higher prices and fewer choices. Also the acquisition would accelerate movement by Nvidia's competitors away from ARM. The only pie that is grown by merging two successful companies is the wealth of the stockholders, everyone else is worse off.
"then anything that harms Arm's market position is also a geopolitical risk to the West"
This is some seriously flawed thinking extremely convenient for corporate interests. Mixing up marlets with national security leads to the kind of atrocities that make totalitarian states proud.
"Coca-Cola Co.'s Colombian bottlers are working with death squads to kill, threaten and intimidate plant workers,"
The risk is that Nvidia could produce ARM chips themselves without having to pay money while raising the rates on other companies. This could create an unfair advantage for Nvidia and allow them to charge consumers above-current-market rates on ARM chips.
Politically, 'big tech' is on the defensive and this is one way for FTC to show some teeth.
That being said, nVidia is currently not playing in ARM's space, so it's hard to argue that this acquisition will harm US consumers or raise prices for them. Hence my feeling is that this is politically motivated, at least to some extent.
Also, we (the US) are entering a period of intense competition with China that may last decades and may even include acts of war. Semiconductors are a key area of competition. More US control of key semiconductor assets is in the US interest. However to be totally fair that doesn't typically concern anti-trust law. But it should concern the current administration and drive some of these decisions about where to focus. It would be very different if this was about social networks and funny cat gifs.
> That being said, nVidia is currently not playing in ARM's space, so it's hard to argue that this acquisition will harm US consumers or raise prices for them. Hence my feeling is that this is politically motivated, at least to some extent.
Huh? They currently use arm designs in the Nintendo switch and their own shield line. They also utilize them in their mellanox Ethernet adapters and network switches. What makes you believe that won't give them a reason to increase prices for competitors that also use ARM chips, if not outright refuse to grant them a license?
> nVidia is currently not playing in ARM's space
Look who ARM licenses to and what sort of products those clients manufacture. They are in direct competition right now in a number of areas.
> Doesn't the US government get more tax revenue
As a frequently observed phenomenon, no.
The intellectual property and brand which generate revenue for ARM will be held in some other country, so profits can be directed there and no tax payable in the USA.
how many US companies rely on ARM being a neutral party though?
while you might win a little on one side of the ledger how much would you lose on the other?
You're not paying attention if you think Nvidia is just a "graphics company". As a chip supplier they are known to be aggressive to the point of bullying and their negotiations tend towards the extortion end of the business spectrum.
I like Nvidia (because I don't have to work with them), but I don't think they are the right steward for ARM in any way.
So the headline should read: Nvidia too much of an asshole to own ARM? Letās update our laws too. Iām serious, if we donāt want companies to be assholes then weād need to regulate how they negotiate and a lot of other things.
Caring only about laws and not about reputation of being a good ecosystem partner to others is a recipe for others to look elsewhere. NVIDIA has a really bad rep due to its arrogant behaviour towards customers and other companies (Linux project, Apple). Maybe this ARM drawn out saga is just karma coming back from that.
Nvidia has burned its bridges with a lot of companies over the years (Microsoft comes to mind, i'm pretty sure they will never use an nvidia chip inside an xbox ever again)
For many different reasons too ! Some were a bit petty/misguided :
- Microsoft was mostly contractual on the Xbox, Nvidia didn't want to lower the price of the chips over the lifespan of the 1st Xbox which was a huge no-no. And some technical people also had a terrible experience (some people in the DirectX team had strong thoughts about working with Nvidia).
- Sony was mostly about terrible tech support and not wanting to share enough technical details which hampered AAA devs a whole lot, to the point most ended up using the SPUs in very creative (and technically fascinating) ways to compensate (the GTA V engine on PS3 was truly impressive in that regard).
- There were some minor (but acrimonious) back and forth with Apple over a failing gen of GPUs (G80s if memory serves), but the large fallout came for the same reason as for the rest of the mobile industry : the Kepler licensing initiative.
Long story short, Nvidia tried to assert patents on mobile GPU (ala Microsoft on Android, as a way to "sell" their exit of the mobile market to investors). They then sued Samsung and Qualcomm [1]. Samsung countersued Nvidia, which had to settle as Samsung was close to winning a ban on imports from some Nvidia products [2]. Many of the patents Nvidia tried to assert were thrown out and they had a terrible legal time.
The damage that ill thought strategy did to Nvidia is hard to measure but there's not a single mobile company that wanted that acquisition to go forward and many are probably quite relieved after this.
[1] :
https://www.zdnet.com/article/nvidia-launches-patent-suit-ag...
[2] :
https://www.reuters.com/article/us-nvidia-samsung-elec-idUSK...
Companies are sociopaths by definition. You can't stop them being assholes, what you can do is regulate their asshole-ness so that it minimizes that effect on consumers and markets in general.
"Free trade" and "Free markets" has never meant "totally unregulated".
That's completely flawed reasoning that ignores the actual purpose of those things, which is to encourage increased/better supply of products and services to consumers at the best possible price.
Maybe they recognize global nature of this issue since Earth has just one Arm, and if it is damaged all people on Earth will be disadvantaged, including Americans, who, even though some of them don't recognize that, live on Earth not planet America.
> Doesn't the US government get more tax revenue if Nvidia makes some extra dough charging royalties for that IP? Iām confused.
Maximizing tax revenues is not the FTCās job.
anything that would improve the US this admin does the opposite 90% of the time. look at the remain in mexico thing which was done away with and now after the epic border fail is going to be stood up again.
The strategy is for the rich and powerful to sell out this country piece by piece, starting with Hunter Biden's Cobalt mine deal:
https://www.wionews.com/world/joe-bidens-son-helped-china-ge...
Great way to cause a chip crises and kill our car industry, but if they keep going at this pace, soon there won't be anything left to sell.
https://www.extremetech.com/computing/326447-arm-china-seize...
People have gotten really into the idea of antitrust without really thinking about what problems "trust-busting" was meant to address or whether it's relevant to the ones they have now. Well, I guess half-baked nostalgia is all the rage these days.
Looking at your other replies your argument seems to boil down to āreasearch needs all the money in the world, consolidation helps with moneyā, but donāt really address where you think that money will come from, and for what value delivered.
The main concern voiced here is that Nvidia-Arm could get away will less research while earning more money.
Thatās basically what happened when Wintel was at peak and there was no viable alternative to Intel chips. They didnāt bring incredible innovation from the ungodly amount of money they gained; they just crushed all their other competitors through various means, got sued, and still made the field plateau for a decade or so before we saw real innovation.
Isnāt it the actual state of play though? At more and more levels there are only one or two players ā e.g., ASML, Samsung/TSMC, AMD/Intel, etc. There are lots of others around but theyāve written off cutting-edge business. Thatās even though there are a number of exciting developments in hardware going on. I just think the barrier to entry has grown so large that itās difficult to avoid.
Right now, regarding CPUs we have:
- Intel
- AMD
- Apple
- Qualcomm
Any of them is pushing the enveloppe in some way or form, Iād argue they are ācutting-edgeā
As an NVidia shareholder let me say thank goodness. NVidia is having to pay more than the market rate for ARM in order to get it from Softbank instead of them just selling ARM on the open market. But because nobody in their right mind would trust NVidia with ARM, ARM under NVidia is going to be worth less than what the market would pay for an independent ARM. And everything NVidia wants to do with ARM they can do by just buying an architectural license like everybody else.
The whole deal would have been a huge destruction of value for the sake of Jensen Huang's ego.
One further reason why this takeover is problematic:
Nvidia would gain unprecedented information on its competitors businesses.
- sales volumes for all arm based products
- plans to license arm cores for new designs
- some details of new designs - eg possibly process technology etc
I was pretty skeptical on attempts to block this merger, still am tbh.
But I'm also very happy this is going to get the scrutiny it needs, both nvidia and arm are pretty vital to the tech industry right now, I can see it going either way, ARM needs better GPU patents, NV needs better SOC patents, the risk of a merger being a disaster is pretty high.
> ARM needs better GPU patents
Remember also that "GPU" technology is not strictly about graphics.
If you browse the marketing website, listen to the keynotes, and generally learn of the NVIDIA vision of the future, it is something like this. There will be tons of little robots and similar endpoints ("edge computing"). Such a robot will be powered by a little ARM chip with an NVIDIA coprocessor for its on-board AI (computer vision, mobility, and the like). The robot will talk over 5G or one of its successors ā the signal processing for the base station perhaps powered by something like NVIDIA AI-on-5G. This communication link will connect it with compute appliances in the nearest utility closet or server room, wherein NVIDIA simulation and route-planning software maintains a "digital twin" of the real world, and helps orchestrate their operations, while operators can render that model on their local machines. And all these devices will have computer-vision and route-planning models on board, models that are trained in the big machines in the datacenter.
We have a few of these systems in computing today, but AI coprocessors will become very interesting as we move deeper into this world.
> Remember also that "GPU" technology is not strictly about graphics.
Itās not even primarily about graphics anymore.
Graphics seem to be the 2nd if not 3rd most common use for GPUs.
https://www.newyorker.com/magazine/2021/12/06/lina-khans-bat...
A rather long read but itās starting to make sense .
Came here to post the same thing. Very good read.
I must signal my approval of this piece. They have a great audio narration available for free as well. I'll be following this new FTC closely.
It's surprising the US gov would take action to stop this. One one hand yes, this isn't great for the industry. But on the other it would mean that ARM will become an American owned company which gives the US government enormous new tech war leverage against China.
It seems extremely unlikely that the UK will allow this deal anyway so it is somewhat moot what the FTC does.
The original ARM sale to Softbank was made while everybody was distracted by Brexit, and is regarded as somewhat of an embarrassment. However Softbank appears to have played themselves because any attempt to sell ARM to a non-British owner will likely be denied as politically unacceptable.
The worst thing Softbank did is selling Arm China stock to China local. This is serious failure for western world.
>However Softbank appears to have played themselves because any attempt to sell ARM to a non-British owner will likely be denied as politically unacceptable.
Softbank gets to keep a cool $1.25 billion in cash from Nvidia if the deal falls through. It's standard M&A practice and so a win-win situation for them.
Is it that general? Its one thing if a deal fails because a party gets cold feet, or can't reach an agreement. Its another if the parties are all agreed and ready and a regulator steps in, isn't it?
I wonder if vague fears about China are part of the reason the FTC has been sitting on their ass doing nothing about the tech industry for the past decade+?
It would be ironic if true, because their failure to intervene has hurt America's ability to innovate and created opportunities for China.
The arm china saga is already absolutely insane.....
Anyone know what rights Apple has to ARM? Do they have a license that pretty much allows them to do anything?
The are one of 15 companies that has an architectural license, which is the most flexible type of license. I don't think the details of the license are public (not even all 15 of the licensees are known). They were also an initial investor in ARM, so I suppose it is conceivable that they still own a small portion of it, together with SoftBank, although that seems unlikely.
If this merger does happen, my hope is RISC-V will win in 5-10+ years horizon. Which I think would be net-net better for the whole industry.
As a former hardware engineer I support the move by the FTC! I have an honest question though: does the lawsuit have any realistic change to block that merger?
This is an aside but: what made you switch away from hardware engineering? And what are you doing nowadays?
It seems it will at least delay it while the lawyers fight it out in court.
It absolutely does. The FTC doesnāt bring lawsuits forward unless they have pretty good reason to believe they will succeed. I suspect they have been gathering evidence to make this case for a while now.
Always such a weird argument. I assume Nvidia also doesn't try to aquire large companies unless they have pretty good reason to believe they will succeed.
This is a good thing, from any way you look. The last thing we need is more consolidation, and less competition, in this space.
On the other hand, a NVIDIA ARM acquisition could really spark RISC-V adoption. Maybe.
There are very few positive signs Nvidia would be a good caretaker of ARM IP, continue to push new innovated and _open_ designs and mass adoption.
Nvidia's entire business and philosophy regarding their chip designs are kind of antithetical to what ARM was achieving.
All you have to see is NVIDIA's trashy effort on Linux and locking-down what-would-otherwise-be-open hardware from being usable by open-source drivers...
It's funny because NVIDIA used to work well on Linux, way back in the day. Then they just decided to not care, then they decided to be hostile. Now they're backtracking but only because they are forced to. I thought the amount of ML would have made them backtrack, but it seems that it is more AMD and Steam.
The quality of Nvidia's linux drivers hasn't changed much. They are about as good as they were 10 years ago.
But everything else has changed around them. The quality of other drivers has improved, the linux ecosystem and what it expects out of drivers has changed, and the nvidia windows drivers have gotten features that the nvidia linux drivers don't.
Wasn't there a situation where the Linux ecosystem decided to use an API that Nvidia officially opposed and never wanted to support? Resulting in Wayland/KDE/GNOME (or similar) not running quite a lot of features on Nvidia? Something about Nvidia only supporting EGLstream while everybody else wanted GBM?
It's very foggy in my mind, so I might be misremembering details.
You have the details precisely correct.
Nvidia are finally moving to support GBM (iirc, it's because someone from KDE asked an nvidia engineer to help get something working with wayland+eglstreams, and he found out it was impossible, as the community had been saying for a while. Slight citation needed on that as I can't find a reference).
Technically they have support in their latest 495 drivers, but I can speak from experience when I say it's not seamless for a user. (I have put a lot of effort into getting it to work and it recently broke again).
Found a reference[0]. I'm not gonna pretend that "thing I read by some random person on a forum" is an unimpeachable source, but it's where I heard it I think.
[0]
https://www.gamingonlinux.com/2021/10/nvidia-beta-4952905-ro...
I cannot agree with this more! I'm so frustrated, since I actually managed to get nvidia + wayland (sway) working on my laptop, and an update of some kind broke my external monitors this week (I haven't had time to roll everything back individually to figure out what yet).
It's an incredibly frustrating experience, and I really wonder if nvidia realises the long-term impact they're having by alienating the tech crowd (i.e. the ones that give recommendations to friends, and decide what hardware their companies will buy).
It is important to remember that Hacker News and Phoronix and Linux kernel developers are within a communication bubble (like any forum).
It may be because, in practice, NVIDIA looks at the size of us all... and yawns because we're tiny and because most tech reviewers still recommend them and most gamers still buy them. And they have a point - AMD has better Linux support, but NVIDIA's got CUDA (which has basically killed OpenCL and AMD's ROCm is less than a proof of concept in quality), DLSS, Raytracing, much better video encoders, a bunch of stuff that people are willing to put up with subpar Linux for.
And ARM had constantly updated Linux kernels and drivers from Qualcomm before didnāt they?
ARM works very, very closely with the Linux kernel. The issues with updated code in the kernel isn't because of ARM CPU IP.
ARM hasn't been much better than Nvidia for GPU code.
Which is why I said CPU IP. That being said, it's a little hard to blame them there, GPUs are a bit of a patent minefield.
ARM has helped found Linaro (linaro.org) ~10 years ago to fix the problem with mainline Linux integration for all the participating SoC vendors.
You should expect the following manufacturers to have good mainline support:
https://www.linaro.org/membership/
I am sure your mileage will vary between vendors (I am pretty sure Qualcomm was NOT one of the founding members, but it's there now), but this at least signals an investment of money and resources.
ARM was never good for open hardware designs.
It wasnāt better before what you think Nvidia will do either. Look at the state of Linux kernel updates on ARM chips.
The openness of the hardware designs has little to do with Arm - it's the SoC designers adding closed GPU drivers for example. Nvidia taking over Arm won't solve that issue.
You seem to be making a lot of bad faith arguments in this comment section.
Why are you defending this deal so ferociously?
What is in bad faith? Itās assumed that ARM support is good now. Iād like to see more Tegra type chips, an open ARM device with updates Linux drivers from the manufacturer. Iād love to be pointed to a better one if you have it. I think what nvidia does with Linux GPUs is awful, but it is very different how they treat their ARM processors.
Please point me to good open ARM processor manufacturers that uploads good Linux updates, Iāll be happy to be wrong that the Jetson really isnāt the best.
All the SoCs available from
should have good mainline Linux support.
"Good manufacturers" today contribute directly to the upstream kernel, and frequently through Linaro kernel trees first.
Would you recommend any of these boards over a Jetson? Iām glad these exist but most of the specs look awful, and Iāve tried rockchip in pine hardware and it sucks.
I don't have much experience with the latest and greatest, and it surely depends on the types of workloads you plan to put on them.
Rockchip has a number of differently performing chips (RK3399Pro seems to perform well in comparison with Jetson:
https://www.cnx-software.com/2019/05/15/toybrick-rk3399pro-b...
,
https://www.seeedstudio.com/blog/2019/12/05/rk3399pro-vs-ras...
). Pine only recently introduced a RK3399-based board (non-Pro, basically lacking an NPU) version, so you might want to go with them for availability reasons.
Which is why the parent said that it would be good for RISC-V processors, not for ARM processors.
Yeah, just look at what they did with GSync
Eh. Who are the biggest consumers of ARM's IP? Apple, Samsung, Qualcomm?
Apple could move to RISC-V (over another two decades), but even if they did, their move wouldn't actually benefit the RISC-V community (and moreover, they'd probably just do their own ISA; at ultra-vertically integrated gigacorporation scale, why not?)
Samsung & Qualcomm? What's the assertion, that Nvidia would let ARM's ISA languish so much they'd be forced to move? Hasn't Qualcomm already been doing that, just with the actual chips; and no one seems to care? Except the people who run iPhone vs Android benchmarks of course.
And Linux on the desktop will defeat Windows' monopoly on the desktop.
RISC-V adoption will be driven by the capability of RISC-V and it being a competitive choice compared to ARM and others. NVIDIA acquiring ARM wouldn't spark RISC-V, it would reduce ARM adoption.
Neither is a good outcome.
As a partial subscriber to the idea that ARM is probably better than RV64 for high performance that probably might not be all that good.
It seems like it might be bad for computers, but good for embedded.
Since when do Nvidia and ARM compete?
Nvidia competes with other ARM customers.
But not with ARMā¦ so instead of Softbank making royalties Nvidia makes them. Iām struggling to understand how this merger would eliminate competition, consolidate the market, and put consumers at risk of harm.
You really don't see how NVidia controlling what price their competitors have to pay for ARM technology and when they get access to it at all could have an impact on the market?
Right now SoftBank controls that price. If they could raise it and make more money off of ARM and piss on the market then why aren't they doing it? Theyāre a business trying to maximize profits after all, itās in their interest regardless of whoās competing for what.
Softbank doesn't benefit if it weakens one of their licensees. NVidia would benefit from weakening other ARM licensees, even if that would hurt ARM income.
I do see how this might be a conflict, thank you for pointing it out. In general Iām not convinced that those competitors couldn't just leave and use some other instruction set. If ARM is too important to business to be owned by a selfish company, then declare it public domain alreadyā¦
>then declare it public domain alreadyā¦
Well that ship has sailed, as ARM is owned by the Japanese Softbank now.
Previously, it was a UK firm, and now the UK is trying to introduce legislation to prevent this from happening again.
ARM was UK's "oh shit" moment. That influenced their move to declare nuclear weapons industry "public domain", i.e. nationalize it:
https://www.upi.com/Defense-News/2020/11/02/Britain-to-natio...
Wow, the concept of privately owned nuclear weapons supplier is really something insane. And it is real. Wow.
Not only it's real, it's how half of the nukes worldwide are made.
The rest are pretty much made in Russia.
Makes you ponder the implications of existence of the privately owned military-industrial complex, huh.
https://www.dontbankonthebomb.com/nuclear-weapon-producers/
I don't know how accurate that website is. For France, it lists the following:
> Donāt Bank on the bomb identifies the following companies as producers of key components for the French nuclear weapons arsenal:
> Airbus Group (the Netherlands),
> BAE Systems (United Kingdom),
> Leonardo (Italy),
> Safran (France), and
> Thales (France)
(Neither of which has anything to do with nukes directly)
Which isn't incorrect but misleading. There are two French nuclear-capable missiles (ASMPA and M51, for air and submarine launches respectively) and one warhead ( TNA). ASMPA is a cruise missile designed by MBDA (which is the successor to government-owned missile companies), M51 was mostly designed by the military with some parts subcontracted to various subcontractors like MBDA, Safran, Thales, etc.
However, the nuclear warheads themselves are designed by a government-owned company, CEA, and everything nuclear related in France is at least majority government owned.
> Iām struggling to understand how this merger would eliminate competition, consolidate the market, and put consumers at risk of harm.
Uh, how about this: NVIDIA raises the licensing fees substantially. NVIDIA wonāt have to pay the licensing fee if they own Arm, which would expand their profit margin at the expense of Arm licensees, and raise prices for consumers.
Or they could refuse to license Arm IP to current licensees.
Both of these scenarios are pretty obvious.
This is exactly why Nvidia wants to buy ARM. Itās straight out of the Jack Tramiel (of Commodore) playbook. Commodore bought MOS to guarantee supply of the 6502 then raised the price for all of their competitors. Itās one of the reasons why a Commodore 64 was less than half the price of the nearest competitor.
Arm design CPU microarchitecture. NVIDIA design CPU microarchitecture.
Good that they remember their tasks of enforcing antitrust rules when it comes to mergers.
When will they remember them in cases of disruptive growth like Google and Amazon and do AT&T?
Going through my stock app the other day I thought Nvidia market cap had to be a typoā¦ since when is a graphics card company one of the biggest companies in the world? 800B is 4x the market cap of Intel, not that far off Apple. How many people realize this?
Because the market thinks Nvidia's potential for future earnings is similar to Intel, Apple, and similar tech companies (Microsoft, Amazon, Google, Facebeta or whatever-it-is).
While I can guess what the collective market is thinking, it's anybody's guess. If you disagree, then short Nvidia?
That's assuming that the market is always rational. It might be in this case but it not always is. See the more or less recent mortgage crisis for an extreme example.
I, for example, don't agree with Tesla's current valuation but I wouldn't be stupid enough to try to compete with Musk's memeing on Twitter.
Nvidia has the market cornered but anyone else can do the same thing theyāre doing for less. Itās just branding and being first to market at this point.
Iām not taking issue with the valuation. I never short tech companies that could have an unpredictable breakthrough success. For all I know they sell millions of GPUs at high prices to justify it. Iām just saying if you asked what are the biggest companies in the world I would never have guessed Nvidia.
Well, clearly lots of investors are aware of this...
It's easier that the US blocks it for diplomatic reasons than the UK. The UK will almost certainly block it if the US doesn't. It could be interpreted as a snub if the UK interferes with the deal, so the US has first dibs on blocking the merger.
Apple, Qualcomm, Intel, AMD probably lobbying FTC against this deal.
Finally, someone gets it. There is a lot of other well intended, intellectually driven arguments and opinions in this thread, but I am afraid that the people actually making the calls here are mostly likely being guided by their own self interests.
No no, this is what I'm trying to get to the bottom of. I don't see any way this is strictly _substantially_ bad for consumers or reduces competition in a way that antitrust law would apply, and I don't see any way this is at all bad for the US strategically. So the FTC is issuing a complaint in response to ..something.. that goes against the interest of our own government and has nothing to do with consumers (hell I'd love to see an Nvidia N2 SoC that competes with the Apple M1). Bingo, it's _other big tech firms lobbying_. These other big tech firms most certainly have y'alls interest at heart and are just looking out for you, small fry. Give me a break!
The point is that this was arbitrary use of power.
Not if the intention was good or bad.
[0]:
https://www.collinsdictionary.com/dictionary/english/arbitra...
Nothing is stopping Nvidia from creating a N2 SoC that competes with the Apple M1. They donāt need to own ARM to do that. They can even create their own SOC with a Nvidia GPU.
Usually I'd agree, but we're likely wrong on this one. I recommend this New Yorker piece detailing the new FTC [0].
[0]
https://www.newyorker.com/magazine/2021/12/06/lina-khans-bat...
The FTC mentioned chips in vehicles multiple times.
The FTCās complaint alleges that the combined firm would have the means and incentive to stifle innovative next-generation technologies, including those used to run datacenters and driver-assistance systems in cars.
and here:
According to the complaint, the acquisition will harm competition in three worldwide markets in which Nvidia competes using Arm-based products:
High-Level Advanced Driver Assistance Systems for passenger cars. These systems offer computer-assisted driving functions, such as automated lane changing, lane keeping, highway entrance and exit, and collision prevention;
I wonder if this action is related to the automotive chip shortage concerns in any way.
Unlikely, it's more that the vehicle manufacturers would be affected directly if these things became less competitive. Vehicle manufacturers are essentially component assemblers that buy their components at various levels of assembly from 3rd party suppliers.
It also is a major US industry so that enables the FTC to describe in the court cases the impact on competition outside of the specific chip design/manufacture industry.
If I was working for Qualcomm... I would say that the FTC suing right now is almost maliciously late.
Edit: So much so, that I would almost start my defense with claiming that the FTC's charges are in bad-faith.
I don't get the logic.
So US government is ok with ARM China getting kidnapped, but has a problem with merge of 2 US companies?
https://www.extremetech.com/computing/326447-arm-china-seize...
> So US government is ok with ARM China getting kidnapped
Two factors here: ARM isnāt a U.S. company, and neither the U.K. nor U.S. would seem to have a great deal of leverage to bring to bear on the situation, realistically.
A late, but potentially still well timed event that will tie up the deal through to next summer when the exclusivity period expires and SoftBank gets a 1.25B breakup fee.
Is that good? The Nvidia-ARM merger has progressed and been in the news so many times that it's been partly responsible for Nvidia's stock growth. It's a bit strange to correct a market by blocking a merger while choosing to sue at the most disruptive time (correcting/bear-ish turned market, inflation policy changes, very close to the exclusivity period expiration).
The FTC should be doing its job but it should be doing it the proper way - not just delaying the deal until it falls apart. It should also do its job promptly... the UK was looking at this deal nearly a year ago - that should have been a huge flag to accelerate their own investigation. What's the point of wasting months of people's time and money.
Rule of law still matters. Itās interesting to see how differently ARM is treated versus ARM China.
I'm conflicted. As an NVDA holder, I'm disappointed. As a tech lover, this is for the best.
Is it? Nvidia gave plenty of good examples of what it has been able to accomplish with ARM tech in their previous keynotes. If anything, I've come around from thinking this was simply a power play on Nvidia's part, to seeing this merger as a win-win-win for Nvidia, the ARM ISA, and ISA competition as a whole.
Whatever neat things Nvidia is doing with ARM can be done with ARM being independent.
Nvidia is building ARM supercomputers. The only other company that has shown interest in ARM supercomputer chips is Fujitsu. The ARM ISA has mostly been left to mobile chips, low-powered computer kits, and experimental desktop computing (StrongARM,XScale, etc.) decades since its inception. If ARM, inc. or Softbank were interested in high-powered chips, either of these companies would have done it by now. But they haven't. Nvidia is the kick in the rear needed to push the ARM instruction set further.
AWS Graviton and Ampere are both based on Arm Neoverse designs.
Plus Apple M1.
Arm ISA does not need Nvidia for a ākick in the rearā.
All very recent developments, but they all have limitations. Graviton is meant for data centers, Ampere is a GPU solution. M1's top configuration, while impressive in the prosumer space, is barely past the teraflop barrier. I doubt Apple and Amazon have any plans to build HPCs.
If only HPC counts as powerful then youāve already quoted that Nvidia and Fujitsu are using Arm under the existing ownership structure.
> If ARM, inc. or Softbank were interested in high-powered chips, either of these companies would have done it by now. But they haven't.
Just factually incorrect given what is happening at the moment with Neoverse.
(And Ampere Computing not the Nvidia GPU design)
Can you show evidence of this?
There has never been a good open ARM processor for mobile computing ever, the Jetson is the best open one while the Apple chips and M1 is the technically best one.
You answered your own question. Nivida doesn't need to own the entirety of the mobile processor market.
So then theyāre no good reason to block it. They would do a better job than it currently it currently is in now, mobile processors arenāt all ARM either.
> mobile processors arenāt all ARM either
How many non-Arm mobile processors are there?
Tegra from a tech lover standpoint is awesome. Nobody making ARM is good with drivers off the top of my head. Qualcomm with their shitty non updated drivers is BS compared to the updated Jetson.
And nvidia can keep making tegra, even without buying out ARM
Great if you donāt need better hardware. But theyāll never be able to make an APU like AMD.
Huh? Apple doesn't own ARM either, and have a GPU with an ARM CPU, Nvidia has an architectural license just like Apple does.
GPU with ARM CPU isnāt an APU.
Apples CPU design is absolutely an APU, I advise knowing what the market actually looks like before constantly posting.
Care to post a source?
Why doesnāt the SoC mention APUs but mentions Intel CULV?
https://en.m.wikipedia.org/wiki/System_on_a_chip
It's a pity MIPS died as a consequence of Apple killing PowerVR. It could provide some competition to ARM.
PowerVR is still around and selling designs, some of which are reportedly still used by Apple. Apple is also still paying licensing fees for PowerVR.
MIPS just got long in the tooth and died because the company (Wave Computing) that was licensing MIPS designs just didn't have enough customers. They have now switched to RISC-V.
MIPS hadn't been providing much competition to ARM[1] in the decade or so leading up the 2016 bust up between Apple and Imagination, so why would it suddenly start doing so after?
[1] In general terms, I know it has (had?) a few niches like routers (IIRC).
Okay let's assume the worst: Nvidia proves once and for all they're a giant green asshole and squeezes the life out of all existing ARM licensees to the point where they switch to a new ISA or die. What is the value of the investment you just made when nobody cares about ARM anymore?
Or, 40 billion to force the market to agree on a new ISA a little sooner than it would have naturally? Smart play. It's nice job security for the engineers that are going to need to port everything, though...
Good. Nvidia is a pretty lock-in oriented and anti-competitive company in general. So them getting ARM would have been bad.
Any ARM hardware that isnāt? Jetson was way more open compared to all other ARM hardware I know of.
Mots are pretty bad I guess. For example Qualcomm is horrible.
Iām excited for project mainline from google, maybe they can make a good generic kernel for all ARM processors.
https://www.xda-developers.com/android-project-mainline-modu...
An improvement for sure if they'll require upstreaming drivers (as in making them FOSS), because Android is causing a rift as long as drivers have blobs due to bionic vs glibc.
Does ARM have any real competitors?
ARM alone is way too powerful
There goes my NVDA calls. I should have sold when it hit the ATH.
Administrative overreach. If the FTC really believed they could win in a regular federal court where they did not hire and fire the judges, they would do so.
DPU SmartNICs,
Which other companies beyond NVIDIA sell DPUs ?
Marvell, Fungible, Broadcom, and Intel (though they may brand them as IPU or something else, they're in the same market/use niche), possibly more.
is softbank (ARM owner) in need of cash?
I wonder how much a business' valuation is determined by sale potential. If you have a business that is worth a lot, and yet nobody can or will buy it, how much is it really worth?
Obviously it's still worth something, it's just that the valuation becomes really fuzzy except for the assets.
Formerly worked in Investment banking: it is a huge huge determinant of value.
Particularly for a lower-dividend, higher growth company like Nvidia, the vast majority of the present value comes from the terminal value (what someone else will pay you for it when you're done holding the investment), made even more extreme by low interest rates.
Couldn't it transform into a high-dividend company?
> _Couldn't it transform into a high-dividend company?_
This is cigarette-butt investing. It ignores terminal value. If the terminal value is already close to zero, it's the right move, an orderly wind-down and re-allocation of assets. If there _is_ terminal value, it's a pillaging. Cases like these, where the terminal value starts looking more theoretical than practical, are how those incentives shift.
What happens in the situation where there isn't a "terminal" point for a company?
Coca-Cola has no reason to "terminate" itself, so why shouldn't it distribute profit excess to its operating requirements back to shareholders?
The terminal point doesn't necessarily have to apply for the company, or even the financial asset based on the company.
There is an simple model in finance [1] that collapses an infinitely growing stream of cash flows into a finite present value. So to your example of coca-cola, even if you assume they will exist and distribute growing profits forever, you can still find a terminal value.
[1]:
https://corporatefinanceinstitute.com/resources/knowledge/va...
A big win for consumers.
I think the comments along the lines of "This is weird, why wouldn't the US govt want a US company to own ARM?" are incredibly simplistic.
First, it's at least possible that some government agencies actually do care about their charters and don't make it 100% about any one-sided geopolitical advantages. It's not hard to see how this deal will be bad for American consumers and businesses at large, and the FTC's purpose is to prevent that harm.
But more than that, even if you _do_ take the position that they only care about realpolitik, the US government's desire to reign in big tech is about the sole thing that has bipartisan support. This action is well in line with keeping big tech from usurping the power of government.
> It's not hard to see how this deal will be bad for American consumers and businesses at large, and the FTC's purpose is to prevent that harm.
I am having a hard time understanding this. Care to explain rather than sidestepping and calling out my comment as overly simplistic? It seems the FTC is acting to protect businesses (ironically other big tech), not consumers. Thatās what I donāt get.
> This action is well in line with keeping big tech from usurping the power of government.
āItās politicalā isnāt really an explanation either. Thereās certainly something to discuss about at what point a company becomes too big and valuable, but that doesn't seem to be the stated motivation here. And Iād be interested in understanding what the framework is for applying those restrictions and how e.g. a company like Apple slid by without getting dismantled.
EDIT: To put my confusion another way, at the end it says āThe FTC acts when it has reason to believe the law has been or is being violated.ā What law has allegedly been violated or would be violated should this merger succeed?
One thing the TFA says: "Because Armās technology is a critical input that enables competition between Nvidia and its competitors in several markets, the complaint alleges that the proposed merger would give Nvidia the ability and incentive to use its control of this technology to undermine its competitors, reducing competition and ultimately resulting in reduced product quality, reduced innovation, higher prices, and less choice, harming the millions of Americans who benefit from Arm-based products, the complaint alleges."
When a downstream firm merges with an upstream supplier that is really important to the downstream firm and its competitors, the merged firm can competitively hobble the downstream competitors. They can refuse to sell the upstream good to competitors, or raise the cost for competitors. Plus it may force the competitors (or potential new entrants) to vertically integrate themselves and enter both the downstream and upstream markets, which chills competition in the downstream market. The government also credits business justifications for the merger, and in the end they balance those with the potential harms to competition.
The article (I can't find the complaint) also says, and this is a typical vertical merger concern, that this will give Nvidia (the downstream firm) access to sensitive information of Nvidia's rivals that they had previously shared with Arm.
EDIT: There is also a general antitrust push in the Biden administration, notably in the appointments of Lina Khan and Tim Wu. Interesting to see its fruits.
Between Intel and AMD there are two competitors here. Strange that the government didn't raise a similar concern when AMD acquired ATI or anytime Intel acquired just about anything but you know how these things work and I guess Apple isn't happy with this.
Somehow reminds me of when the SEC decided that Nvidia was the next Enron back in the 00s because a few of their employees did some piddly insider trading and they spent the next two to three years trying to destroy the company. In the end, earnings were adjusted up 10 million and they got rid of their CFO as a sacrificial lamb. All these government bureaucrats need to look busy after all.
>Somehow reminds me of when the SEC decided that Nvidia was the next Enron...
There was significant insider trading occurring due to an internal email about the Xbox deal. 10 employees and 15 people total.
_"The Securities and Exchange Commission has sued 15 people, including the 10 suspended nVidia employees, accusing them of insider trading in shares of the graphics chipmaker based on advance information that it would win a lucrative contract from Microsoft Corp."_ [1]
In a separate incident, nVidia wanted to show better quarterly results and tried to pressure their supplier to reduce costs, with the promise of paying more in the future. Their supplier wanted it in writing. The CFO knew they could not have such an explicit agreement in writing, as it would not allow them to write down the cost savings for the quarter, so they directed an employee to author two separate agreements to obfuscate their mutual nature.
_"We can not sign this or have this in print. Will wipe out the credit in Q1. Need to arrange this separately and trust us to abide by it."_ [2]
Seems kind of in the purview of the SEC to look into these kinds of things.
1.
https://money.cnn.com/2001/11/20/technology/nvidia/
2.
https://www.sec.gov/litigation/admin/34-48480.htm
An inappropriate release of market influencing news that was exploited improperly by 10 employees of a 500 employee organization is your definition of significant insider trading on par with Enron? $1.7M in profits total. Those employees were terminated and they paid fines and I believe one was banned from working in tech going forward. The hyperbole here is ridiculous. That's about 1/1000th of an Enron.
But thanks for pointing out why they got rid of the CFO. I didn't know about that part. It makes more sense now.
The SEC didn't compare nVidia to Enron, the media did. That was because nVidia was the best performing S&P 500 stock for 2001, a title it took over from Enron. The unwinding of Enron's massive accounting scandal was still hot in the news at the time. It was an obvious comparison to make about another "top performer" having an accounting scandal. However, the "overreaction" by the media and market was not the fault of the SEC.
The punitive remediation for the insider traders was done due to the SEC investigating, which then lead to discovery of the accounting issue. Your characterization that the SEC had a vendetta against nVidia is completely wrong. Maybe nVidia should have had tighter controls on privileged information and better insider trading education for their employees. Maybe they also shouldn't have tried to cook their books to deceive the market. If it really was only a puny $1.7M and didn't really matter, why'd they do it? _CORRECTION: $1.7M was profits by the insider traders, nVidia misstated $3.3 million in cost savings._
Ironically, Nvidia restated earnings slightly upward as opposed to Enron overreporting by $600M. Not so sure whether that was "cooking the books" or just playing fast and loose with ambiguous accounting laws which is in my experience what accountants do, but they did get caught, no argument there and the CFO was fined. But I wonder who else would get caught doing similar things of similar magnitude if their annual statements were scrutinized in as much detail here specifically looking for trouble.
As for the inside trading employees, they were fools. Good luck keeping out fools like that once you have a hundred or more employees or why do some googlers stalk their ex partner's search histories? Why do some Amazon employees snoop on Alexa recordings? Why do some Facebook employees look at the private friends list of their ex partners? Etc. There was a second insider trading investigation in 2014 that was handled quietly and efficiently unlike this fiasco. That one seemed a bit more nefarious and systematic IMO and yet no one compared it to Enron. Funny that.
https://www.marketwatch.com/story/nvidia-sec-accounting-prob...
https://www.sec.gov/news/press-release/2014-82
You just seem to want to try to come up with a bunch of excuses to downplay or redirect attention away from nVidias culpability in what they did. It was not the SEC's fault they had insider traders and their CFO knowingly and deceptively hid quarterly costs on purpose. Accountants should not actually be playing fast and loose, unless they want SEC attention.
Did 2014 involve accounting discrepancies? Did the SEC still release a public press release like they did in the 2000s? Their handling wasn't much different here. Enron was long dead by 2014 and it doesn't look like there was wrongdoing by nVidia directly, so obviously the media wouldn't make such comparisons.
Digital privacy concerns don't directly involve the SEC, though maybe the FTC. That's another topic, but one worthy of competent regulatory oversight.
So you are stating that the release of the "X is ours" email that started all this was deliberate corporate misconduct by the CEO? I don't even think the SEC insinuated that.
And that the SEC finding a small accounting discrepancy so as to justify the time they put into this investigation is actually an Enron level event?
Got it. We see things differently. You think I'm downplaying what happened. I think you're overstating it just like the media compared them to Enron.
No, it was the CFO explicitly calling out that they can't do something and then doing it anyways to deceive shareholders about quarterly costs. That should be obvious.
Just because Enron had massive fraud doesn't make $3.3 million "a small discrepancy", nor is purposefully doing something merely a discrepancy or oversight.
No one compared the insider trading to Enron. The actual accounting issue got them those comparisons.
Neither Intel nor AMD supply SoC cores in the same way that ARM does--they don't really operate in the same market segment.
NVidia's competition in the hardware acceleration space often includes ARM cores for management (e.g., the Xilinx Zynq
https://www.xilinx.com/products/silicon-devices/soc/zynq-700...
), or are extension of the ARM instruction set (e.g. AWS's Graviton
https://aws.amazon.com/ec2/graviton/
).
RISC-V suppliers (e.g, SiFive) is the closest thing that ARM to has to a competitor here, and they're nowhere near the same scale.
That seems awfully esoteric to me. Compare and contrast with Intel taking Altera off the map and AMD taking Xilinx off the map. At the time, FPGAs were erroneously seen as a credible competitor to GPUs in artificial intelligence. They absolutely smash GPUs at extreme low latency, but it was a poor technological take to assume that meant they could beat them where GPUs were strong but that didn't stop lots of money being thrown at the attempt.
Which is to say it seems odd that the FTC would get its hackles up over a case as nichey as you describe. it would seem the elephant in the room is Nvidia's dominance of AI. There's nothing stopping AMD or Intel developing equivalent SoCs. In fact one of the cases you mentioned comes from Xilinx ergo AMD.
So I have to think this has to be something as basic as slinging mud at their dominance of AI. Something as simple as trying to exclude competing browsers from your operating system's desktop. Something as understandable as trying to exclude third parties from collecting money on your mobile platform. And since both other cases are larger more general instances of dominance, it's curious they aren't being investigated as well. Well not really, there's probably a lot of grift here.
But if we're going to worry about a single party having dominance of AI then we have to start asking questions about Google and Facebook controlling the major interfaces to AI. Sure, they are open source, you can fork them if you like. But they get to control all the pull requests into the master branch. That lets them control how well any one platform runs their framework. That seems a bit anti-competitive as well when at least one of the parties has their own AI hardware.
The people enforcing the laws change between administrations, so it is not unexpected that past actions look inconsistent. Biden's people were not in power then (in fact the current FTC chair was seventeen).
I'm not sure if you're saying that Intel and AMD existing means that competition is doing fine, but three actors is a highly concentrated market.
4/5 of the FTC commissioners were appointed in 2018
We seem to be OK with only two players in mobile and on the desktop for now. Interesting the battles we choose to fight.
If thereās an antitrust push, how have they not looked at Google or Amazon yet? Both businesses engage in anti-competitive behaviors in markets they dominate which possibly (likely) cross the line.
Who says they're not looking at Google or Amazon? From
https://fortune.com/2020/02/11/ftc-antitrust-probe-google-fa...
:
"U.S. technology giants face a new wave of scrutiny from antitrust officials, as the Federal Trade Commission demanded information about their acquisitions of startups that may have eliminated emerging competitors. The FTC issued orders to Alphabetās Google, Apple, Facebook, Amazon.com and Microsoft for information on the terms and purposes of transactions they closed from the beginning of 2010 through 2019, the agency said Tuesday."
Here's Matt Stoller's explanation:
https://mattstoller.substack.com/p/ftc-sues-to-block-nvidia-...
I'm guessing "the law" is probably the Sherman Antitrust Act, which is still on the books, though enforcement went out of fashion in the 1980s. Seems to be making a comeback, though.
I follow Matt and often agree with his takes. This short piece just seems to be a non-opinion announcement to keep his followers appraised.
> This is Lina Khanās first major merger challenge. It is also a unanimous vote, and ironically, not all that bad for some of the key players in big tech.
I think this key point is tellingā¦ Big tech has a big hold on our political-think. And even Matt seems to have conflicting thoughts.
Either Sherman or Clayton Act.
From the Clayton Act:
* mergers and acquisitions where the effect may substantially lessen competition
This is my bone. I don't understand how this merger substantially lessens competition. Let's all concede the merger is obviously bad for every single business in the US and there is nothing redeemable about Nvidia and no reason to believe it should own ARM IP. They're still not competitors and no competition in the market has been lessened by such a merger. In fact, access to other people's sensitive business details probably makes things more competitive and forces participants to innovate in other areas. Only day to day contract negotiation has become shittier because now you have deal with "shitty" Nvidia. Where's the law that says companies can't vertically integrate? Is there a precedent for blocking these type of vertical mergers solely because it might be good business for the acquirer and unfortunate news for other participants? Why couldn't other participants put a bid out on ARM? Why can't Apple and Google just throw 100 billion at SoftBank and say we're buying and freeing ARM? Idk maybe I was simply born into an age of spineless non-enforcement of anti-trust, but I'm not seeing how this scenario warrants more scrutiny than "normal".
_I don't understand how this merger substantially lessens competition._
Step 1: acquire ARM
Step 2: make life more difficult for other ARM license holders and/or easier for NVidia
Step 3: be the only viable supplier of ARM chips
_I'm not seeing how this scenario warrants more scrutiny than "normal"._
Speculating: maybe the old "normal" was too low, this was just the easiest first move, and other more difficult antitrust moves are in the pipeline.
Intel could do the exact same thing to any X86 vendor today, couldn't it? And AMD to X86_64 vendors, no? I'm not saying your logic is wrong, but this problem seems more pervasive than Nvidia and ARM. In reality if we stop this merger what we're saying is that it's a problem for _any_ company to own IP that they license to competitors while at the same time producing and selling an in-house product that leverages the same IP. I'm not ignoring the conflicting interest here, but we've never said "a company can't be vertically integrated and also license technology to competitors". Maybe times are a changing, but if this becomes the new norm the only thing it seems to protect are incumbents in the space who are already vertically integrated...
> Intel could do the exact same thing to any X86 vendor today, couldn't it? And AMD to X86_64 vendors, no?
They already have killed off any effective competition, and that's a bad thing. If x86 was more diverse and was moving toward the current set of two, blocking it would also be good.
At least the important parts are falling out of patent...
> In reality if we stop this merger what we're saying is that it's a problem for any company to own IP that they license to competitors while at the same time producing and selling an in-house product that leverages the same IP.
Well it is. How much of a problem depends on how much competition the different parts of the market have.
Take note that the FTC has not actually done anything, they have only initiated a lawsuit. A court may disagree with the FTC's assessment and permit the merger. The government loses antitrust suits. But antitrust suits are long and expensive, so a challenge like this may cause Nvidia and Arm to back off.
Your intuition is right in that vertical mergers are viewed less suspiciously than mergers between competitors. But courts still assess the impact to competition in vertical mergers. As you did, assume the worst case: if Arm is really crucial to Nvidia's competitors, and the merged firm keeps Arm's designs to themselves, competition in chip-supply is harmed. Yes, Nvidia was savvy etc. and is just doing what another actor could do, but compare this method of Nvidia beating their competitors versus the "ideal" way where Nvidia makes a better product, does it cheaper, and is generally more efficient. That's what antitrust law wants. It wants the merits of the product and the org to decide the winner in a market, not things like buying Arm and keeping them to yourself.
It looks like your question was answered in another reply to you more than a half hour before this comment of yours.
https://news.ycombinator.com/item?id=29423870
I'm by no means an attorney but the argument made seems sound.
Well if the original creator of this thread had replied in-thread to my other comment, then I wouldn't have felt the need to engage in two different spots. I agree it's annoying.
Anyway I can't help but feel the argument is pretty loose on demonstrating that Nvidia would be in a position to harm the market in a _substantial_ way. Usually the law is enforced reactively to punish bad/unwanted behavior. We preemptively prevent mergers that would result in no real competition existing in a market, in other words: a monopoly. We do that because it's bad for consumers. You can echo the FTC's statements about how having leverage over other market participants might be harmful all you want, but the reality is that they don't explain how Nvidia owning ARM creates a monopoly (is akin to something like Intel owning ARM) and then how that is inherently bad for fundamentally bad for consumers. Nvidia owning arm looks like savvy business at best and at worst annoying and disruptive to some people who put most of their eggs in the ARM basket. I see how it could in theory have an effect on some competition to have this type of vertical integration happen, but is that _substantial_ to the point of Nvidia being a monopoly on microprocessors and consumers left abused and holding the bag? That's quite the claim.
AMD competes with Intel and Intel is similarly vertically integrated. AMD even makes graphics cards after merging with ATI and that didn't kill any markets or harm consumers, if anything AMD graphics has become more competitive. Intel is entering the graphics card space. Personally I'd love to see another player in the processor space. Right now it's Intel and AMD and now Apple. Why wouldn't an Nvidia N2 ARM SOC that competes with the Apple M1 be a good thing? IDK I see potential consumer benefits to Nvidia being able to run with ARM. I can't help but feel like we're straining here under the guise of "big tech is big and bad let's punish them all".
Anyway probably at the end of the utility of going back an forth on wether the merger _substantially_ lessens competition to the point of causing consumer harm. We'll see what the courts decide.
I mean, ARM is basically the only game in town when it comes to mobile chips. Combine that with nvidia's graphics technologies and now you've got basically the entire market for XR devices (I reject the use of 'metaverse' in this context) cornered.
It's one case out of many where suddenly a defacto market leader emerges with almost no room for other entrants, whereas at least without the merger device makers can source silicon from arm licensors as well as from AMD, NVDA, and INTC (soon-ish) for solid graphics components.
I agree with you that it's _obviously_ dumb to let Nvidia buy ARM... BUT we're already in a situation where softbank own ARM. I think the only argument that Nvidia is more dangerous is that they're more competent.
But Softbank and ARM are not in remotely similar markets and ARM can't meaningfully favor Softbank as a customer even if they wanted to.
So yeah, I guess I'd argue that NVidia is more dangerous because they're more competent. In the sense that they are competent because they are in a close enough market for there to be a conflict of interest.
With SoftBank Group as a holding company, ARM pretty much runs the same as before.
Should Nvidia own ARM, the avenues for exploitation aren't just obvious, but anticipated.
I donāt think there is any charter to protect competition outside the US. The reason the FTC would block the merger is to protect US companies and agencies doing business with ARM and/or Nvidia. It maintains competitiveness in their purchasing options.
> But more than that, even if you do take the position that they only care about realpolitik, the US government's desire to reign in big tech is about the sole thing that has bipartisan support. This action is well in line with keeping big tech from usurping the power of government.
It's kind of a team project, isn't it? The links between the two go pretty deep.
I guess, the government is friends with Apple. The corruption in the US is beyond imagination, just on a totally different level than anywhere else in the world! When iPhone was a monopoly, it was fine. When Facebook is a monopoly, it's fine. When Google is a monopoly, it's fine, but when it comes to more traditional businesses - no, it's not fine as chips can't be used as political tools unlike Apple's, Google's, and Facebook's platforms!
I think you need to recognize that the FTC is governed by its commissioners and chairperson, and those people have their own priorities that are going to be different than their predecessors. Lina Khan, current chair, is much more focused[1] on controlling tech monopolies than Trumpās appointees, who had what you might charitably call more āmarket orientedā priorities[2] (to me, a euphemism for ābought by big businessā).
[1]:
https://www.newyorker.com/magazine/2021/12/06/lina-khans-bat...
[2]:
https://www.politico.com/story/2019/05/08/ftc-chairman-congr...
Is there are more a blatant monopoly than Meta? Or Google for search?
It's shocking how just four years of Trumpism have made it difficult for Americans to even _imagine_ a US government agency actually doing its job, as it was intended to do.
We now just automatically jump to assuming "fuck everything, except making more money and Mercuh".
Sad.
Would you please stop taking HN threads further into flamewar? Discussion quality takes a sharp step down with posts like this. Please make your substantive points without doing that.
We detached this subthread from
https://news.ycombinator.com/item?id=29423448
.
> It's shocking how just four years of Trumpism have made it difficult for Americans to even imagine a US government agency actually doing its job, as it was intended to do.
It wasn't just four years of Trumpism. Trumpism was simply the fruition of _decades_ of successful neoliberal anti-government propaganda, primarily from the Republicans, going back at least as far as Reagan saying "Government is not the solution to our problem, government is the problem." What Trumpism added was mainstreaming conspiracy theory on top of the general disposition that government should be starved until it was weak enough to be drowned in the bathtub. So now the Libertarians are talking about limited government _and_ how Fauci created COVID for the Illuminati.
Its not
Trump that jailed chelsea manning and wage war against assange. Wake up
So what I'm reading is Intel ordered the FTC to dismantle any potential competition.
Please don't post unsubstantive/flamebait comments. We're trying for something else on this site.
https://news.ycombinator.com/newsguidelines.html
Sorry, I'm expressing my sincere belief that there's real corruption at the upper echelons that's benefitting Intel. I can see how my terse comment would come across as flamebait, it wasn't intentional. Intel's consistently rich margins don't really make sense if we actually have a fair and competitive market.
This might be the geopolitical blunder of the decade. If this merger doesn't happen China is going to win the chip race.
As an European I'm increasingly bullish on China and see the US and EU as self-sabotaging neurotic entities.
> If this merger doesn't happen China is going to win the chip race.
Wut? How is NVIDIA keeping new ARM designs from competitors like Intel, AMD and phone manufacturers helping to win the "chip race"?
Promoting fair competition (not crony captalism) is how the US wins the chip race
What is the "chip race" and why is a combined NVIDIA/ARM necessary to "win" it? Aren't AMD/TSMC, Apple/TSMC, and Intel enough to be competitive?
TSMC is Taiwanese, China will takeover Taiwan in the next 5-10 years, and Intel's manufacturing processes are ancient. The future when US/EU can't produce our own chips is rapidly approaching.
TSMC is building new plant in the US and elsewhere outside of Taiwan.
There's no guarantee that Chine will "takeover" Taiwan any time soon, the geo-politics of that are way more enormous than just chip manufacture.
Most of the development in manufacturing of chips is by the companies that supply the fab developers, the integration of those skills and components is where the different fab owners have their competitive advantages.
The "future when the US/EU can't produce our own chips" was also in the 80s when there was a push to develop memory manufacturers and back then, Japan was the evil empire that needed to be stopped.
TSMC would continue post a China takeover of Taiwan.
And... What do ARM or NVIDIA have anything to do with fabrication? If anything a merger will hurt Intel and make it harder for them to update their processes.
You do realize that neither Nvidia nor Arm, have their own foundries and have shown no signs of wanting to build one ?
TSMC and Nvidia are not in the same business.
The idea that you can possibly have a bunch of small companies competing in the modern semiconductor space is risible. If a big private concern makes you feel bad you could think about bringing the government into it but I see little other way forward.
This is not material in my view in the sense that it was going to get blocked by EU. And to be perfectly honest, itās really not Nvidiaās fault that ARM sold out to Softbank, a known player with hard investment return mandates. I am sure Apple gave this a big push along with other players mentioned in comments. As for the validity/veracity of the suit, itās 50/50.
It's likely to be blocked by the UK first, then potentially the EU, then the FTC.
There are plenty of people lining up against NVidia on this.
Our government continues to show how corrupt and bad it really is. It'll be super funny when Qualcomm buys ARM instead because that's a-okay and all that. Edit: you know Qualcomm has already said that's exactly what they're going to do right? Maybe the Nvidia CEO needs to kill a goat with a stun gun and feed it to the FTC to send the right message? That seemed to work really well for Zuck.
https://www.cnbc.com/2021/06/14/qualcomm-offers-to-invest-in...
You realize that its not only the FTC involved here.
The UK is already investigating as well, see
https://www.theverge.com/2021/11/16/22785832/nvidias-arm-acq...
The EU is also investigating:
https://www.theverge.com/2021/10/27/22266504/nvidia-arm-deal...
Any one of these regulators can both delay and/or stop and/or change the conditions of the proposed merger.
It's highly likely that any other direct acquisition of ARM would get the same investigation, no matter who the acquirer is.
And I get why the UK is concerned. Their regulators caught flack for allowing foreign companies to buy Deepmind and ARM. But I don't get the United States government's motivation here yet. The story will come out in time.
https://www.cnbc.com/2020/11/13/britain-to-protect-tech-firm...
>Edit: you know Qualcomm has already said that's exactly what they're going to do right?
Well that's not what your link says
Okay if you trust that buying an unspecified stake doesn't lead to an acquisition once the smoke has cleared you have a very different worldview than I do. We will have to agree to disagree.
It does seem like the next 100 years is just more of what's been going on the past decade and it seems pretty grim for the customer and great for the corporations as they duke it out with each other much like what this seems to really be about.
I don't need to trust it, the rules for acquisition are much more strict than the rules for part ownership.
Fascinating... So many strange acquisitions and this is the one that triggers all the red flags. The story will come out eventually and I think it's going to be a doozy.