This is probably the main reason Qualcomm is against the NVIDIA purchase of ARM while Mediatek is for it. Qualcomm doesn't want more money pumped into ARM to make more competitive CPU cores, as that would allow Qualcomm's competitors (which are Mediatek and soon Unisoc) to build SoCs that are more competitive with Qualcomm's Nuvia-based SoCs. A less active ARM makes Qualcomm more dominant.
A bit off topic. Lot of politicians are complaining the big tech is too big but the last 20 years of my life i see them buying up companies none stop. For me this is a other example were I am just please stop. But then promises are made , hands are shaken and few years down the line people go "How did we let them get so powerfull and market dominating" And yes I am aware halting Nvidia would feel very unfair since everyone else got away with it but I personally find you have to start at some point.
I don't see a very strong case for the US government to deny Nvidia's acquisition of ARM. I do think it's probably bad for most of ARM's customers, in the long run, and therefore the entire ARM ecosystem. But, that's not really the government's purview.
It's not as if they're buying out a direct competitor, which is usually the grounds for rejecting mergers & acquisitions. There are really just a couple small overlaps between them.
Arm's Mali hardly competes with Nvidia's graphics. Who exactly has a choice of implementing Mali graphics or Nvidia graphics and will be losing choice if Nvidia buys Arm? Practically speaking, Arm's competitive offerings will be made better because Nvidia's graphics IP will eventually be offered to customers as well and it will be preferable to Mali in many instances. But that won't come as a result of loss of choice in the market. And it's not like there aren't other graphics providers besides Arm or Nvidia in the business sectors where each of them operate, anyway.
> Mali graphics covers over half of the graphics solutions on the market at this point.
Nvidia only slightly overlaps with their top-tier offering, and hasn't traditionally offered it as standalone IP (although they've announced a deal with Mediatek that amounts to that). It's not obvious whether Nvidia can or has any desire to try and scale down its solutions to compete with Mali's lower tiers.
Buying a competitor from the same field (graphics) but not the same market (mobile vs. datacenter, or highend vs. lowend) is even worse as it gets you closer to having the entire market.
Don't confuse the Arm ecosystem with Arm's customers. Firstly, I think NVIDIA buying Arm is good for most of Arm's customers compared to any reasonable alternative. But it's definitely good for the ecosystem. To say one thinks it's bad for the ecosystem is to say that one thinks NVIDIA is going to make a mistake. NVIDIA certainly has no intention of damaging the ecosystem of their purchase. They don't have any reason to shut it down (sometimes companies do buy properties to shut them down). It's certainly good for the expansion of Arm into new areas such as Arm in the datacenter and Arm in OpenRan. And actually, it doesn't matter so much if it's "good" for Arm's current customers or not. A concern might be if it were particularly disruptive for a significant segment. But the basic question is whether it causes a problem for free and fair competition in the market as a whole.
> Don't confuse the Arm ecosystem with Arm's customers.
I don't, but if the deal drives away enough customers, it's very likely bad for the ecosystem, also.
> They don't have any reason to shut it down
There are other ways they could damage the market, such as by offering products competing with what a lot of ARM's customers are building, and that could scare those customers into feeling like they might end up at a disadvantage. Everyone knows how ruthlessly competitive Nvidia is.
> the basic question is whether it causes a problem for free and fair competition
As I said, I don't really see the case that it inherently damages competition, as long as ARM's customers have other options (which I think they do).
It doesn't drive away customers. Customers are going to continue to use Arm like they've been and they'll switch to RISC-V in parts of the embedded space just like they've been.
NVIDIA releasing more products to the market is not going to damage the market.
Don't worry, companies don't need safe spaces.
The bottom line is, if no one buys Arm, Arm is in trouble. If Arm is controlled by a consortium of its customers, Arm is in trouble. People forget that the money pumped into Arm by Softbank over the past 5 years was an investment by Softbank. And since Softbank wants to unload Arm for about what they paid for it it was an investment that didn't pay off. All these great cores that enabled Graviton and enabled so many companies to forgo designing their own Arm cores are a result of that investment. Perhaps everyone would like an entity like Softbank to own Arm but that's like asking for a sugar will not exist any more. Under NVIDIA, the Arm ecosystem is likely to increase more than it did under Softbank. Softbank set the stage for the Arm ecosystem to grow, but NVIDIA is going to do the yoeman's work of building the foundation for the ecosystem. NVIDIA is not going to start making smartphone SoCs. Right now Arm in the data center is controlled by Amazon, and that would be a horrible thing if it were to go that way, to be controlled by the hyperscalers. It would be like being controlled by Apple. Each cloud company creates their own locked-in ecosystem and has a domineering relationship with any small supplier.
I think most people on message boards who are against the merger have an idealized view of the world and of Arm. Open source software did nothing to prevent public cloud lock in. An independent Arm will similarly do nothing in that regard. An independent Arm is unlikely to build an independent server ecosystem outside the clouds.
If you mean your post is off topic to mine and to the article, then yes, I agree. What I said is on topic to the article:
"In statements to Reuters, Amon had made comments regarding the company’s future CPU roadmap, which come to further contextualise the company’s completed acquisition of NUVIA last March.
"We needed to have the leading performance for a battery-powered device," Amon said. "If Arm, which we've had a relationship with for years, eventually develops a CPU that's better than what we can build ourselves, then we always have the option to license from Arm.""
I was not talking about the regulatory situation concerning NVIDIA and Arm. I was talking about the reason Qualcomm is likely against the merger. But I will reply to what you said anyway. In fact what is going on now with the new administration is the bureaucracy flexing its muscle to show the corporations who is the boss. It's a sort of saber rattling. Not that they wouldn't think it's heroic to pull off some big breakup. It has nothing to do with innovation or consumer protection. If government regulators and politicians really were concerned about companies being too powerful in the marketplace then they wouldn't let subscribestar be attacked the way it was. They wouldn't let gab be attacked the way it was. As far as decision on M&A, it's definitely not up to regulators to choose winners or losers. A regulator should look to see if a merger causes a trust or otherwise interferes in a free market. It's not about fear, or about predicting the future. The minute they think that way is the minute there is no free market to protect. NVIDIA buying ARM in no way creates a trust. Neither does it create a company that is particularly large overall. Perhaps regulators shouldn't be allowing facebook to buy whatsapp or instagram. If NVIDIA were buying up all the AI chip startups that would be a problem. The current tech giants (Google, Facebook, Apple, Amazon) buying up all the little companies that could turn into real competition for them is a problem. For example, if Zoom didn't explode like it did because of the pandemic perhaps someone would have bought it. It had a market cap of $20 billion at the start of 2020 and not it's about $120 billion. Such a purchase would have gone through no problem. If NVIDIA down the road is acting anti-competitively towards the market then the regulators will do what they love and punish them for it, as they should. But as is now, it's just the equivalent of making the market to block such a merger, not protecting it. The deal should go through and the deal with go through. Because regulators 1) understand the situation that no real trust is being created, 2) would rather get involved in a heroic punishment that they can defend than a controversial block that gives them no glory and which they could be forced to impossibly defend themselves against in the future (suppose Arm fades away after the block the deal, and Intel, for example, takes a bigger bite of the market, there would be little way for regulators to defend themselves). There are political reasons non-US regulators might have for wanting to block the deal, but they will probably think better of them.
> In fact what is going on now with the new administration is the bureaucracy flexing > its muscle to show the corporations who is the boss.
In the lawsuit against Facebook just thrown out by SCotUS, 48 States joined the Federal Government, filed ** last December **. So, no. The backlash against Big Tech is not something new or specific to Biden.
The fact is that big platform companies have more control over our lives than ever, and the people need to have more say over that, since the market has failed to provide the necessary levers of control. That's what it boils down to, and where people of most political persuasions would agree.
> If NVIDIA down the road is acting anti-competitively towards the market then > the regulators will do what they love and punish them for it
Huh? I don't think there's anything they *can* do. I certainly can't think of any examples, in recent history. After the acquisition goes through, the only recourse is for someone to sue for damages, due to unfair practices.
Firstly, the bureaucracy has a will of its own. Secondly, just because there was already a lawsuit says nothing about growing calls for regulation. If you pay attention to the way Biden's administration or Congressional Democrats are talking you will see that clearly. Thirdly, regulation is not in general a bad thing, it is a necessary thing. It's just that regulation can become its own reward if the system gets corrupted.
Of course there's something they can do. They can break up a company, they can prevent a company from doing certain things. It happens to Microsoft. It almost happened to Qualcomm but the Trump admin stepped in and prevented it because Qualcomm was viewed being of strategic national importance. As far as for Google, Amazon, Facebook, these companies are too newly dominant and people don't understand the ways in which they are dominant. Things take time. There almost certainly will be regulation of them in the future. If you start trying to prevent the situation from occurring and guessing winning or losers you are then by necessity picking winners or losers and you no longer have a free market. Regulators can't be stock pickers. Not only that but there is nothing illegal about a monopoly. What is illegal is using a monopoly for anti-competitive practices. Now regulators will step in to prevent a monopoly from forming when the combined companies would actually create a monopoly immediately as the companies are combined. But to say "oh this company is going to be too successful and then later they will become a monopoly and then they are going to abuse that monopoly" is not, and should never be, the job of regulators.
Question: You imply that the main benefit of the ARM acquisition by NVDA is that more money will be pumped to ARM and the products will become better.
But ARM is already valued at let’s say $40B. Are you saying a company with 7000 employees and 40B valuation needs “money” to make the product better?
Ok, how about this. Why don’t they do an IPO themselves? or just go ask investors for private money?
And what stops their competitors or pouring themselves more money to their own products to make them better. Does INTC lack money ? Do you think the stagnation of INTC is due to lack of resources?
I am sorry but I cannot really follow the reasoning of why the money that will be pumped by NVDA are the only money on the planet available for that or why just money will produce better products.
I really hope the NVDA lawyers come up with better arguments :)
Arm is valued at $12 billion plus $28 billion of NVIDIA stock by NVIDIA. That is the amount NVIDIA is paying for the money in stock and cash. It's a bet they are making, not anyone else. That $12 billion in cash and the $28 billion in stock goes to Softbank, not to Arm. That is not cash that Arm has on hand or would otherwise have on hand for its operations. It's like the difference between a transfer fee and a player salary in soccer. If Man Utd buys Jadon Sancho for a $100 million transfer fee it's not like Jadon Sancho is getting $100 million or otherwise would have $100 million on hand if he stayed at Borussia Dortmund.
Company valuation and company operations are two different things and it makes no sense to talk about dollar values from one and apply it to another. Company operations rely on credit worthiness, revenue, and expenditures. An independent Arm would have short-term financial demands placed on them. Company must display profitability in the open market or their stock will tank. If their stock tanks then investors lose money and the company management team will at best receive much less compensation and at worst be replaced.
Private investors similarly want to have their investment paid off. They could be doing a million other things with their money. It's not easy to find a private investor to outbid $40 billion, anyway.
Additionally, NVIDIA has the experience, expertise, and drive to expand Arm into the data center and OpenRan. It takes more than just throwing money at something to do something. It takes vision and skill, and it takes the willingness to try.
NVIDIA doesn't need lawyers to come up with such arguments. That's not what regulation is about. The regulators' business is not to try to make the strongest Arm. Various UK entities might be worried about such a thing. But regardless, they already know that money doesn't come from thin air. They already know about company operations. They also know that Arm has a small revenue and a small profitability. So it's not a hard argument to make to them.
Thank you for the lesson in finances and company valuations. I appreciate it, even though I still do not feel I need it.
Why do I feel you didn’t really answer my question though? You certainly went out of your way to verbiase on details but I do not think you touched the gist... I will leave it at that....
I was at Qualcomm back in the days during Krait development. I remember this guy. This was before Android seriously took off and the iPhone was both the gold standard and also something everyone took pot-shots at.
He was touting an HTC phone that ran Windows Mobile (not Windows Phone yet) and had HTC's skin on it. It used one of the very first Snapdragons (Scorpion CPU based back then). He talked up how open it was to side-loading apps.
Someone mentioned how laggy the UI was -- and it was very laggy, definitely not an optimized for touch device. At the time anything non-iPhone had a very stuttery, delayed touch UI.
His answer was "who gives a shit?" in a half-joking, half-dismissive manner.
I can’t (obviously) backup your employment claims, But, Amon has 100% been dismissive of claims QC failed to innovate on our CPU performance.
I mean, they ran tail between their legs to Arm Ltd for cores! The delusion!
Amon in 2018: “So as the hardware has evolved we realize that for single threaded performance on the CPU, we are in a place that is good enough for all the applications that are needed, and the vectors of innovation needed to come from other areas.”
Foolish CEO. Really? You thought people *liked* the ancient Windows on Arm device performance? Nope. We liked WoA in spite of the terrible performance to any modern x86 laptop.
You can’t fix these kinds of delusions. When / If NUVIA falters in a few years, he’ll be back to claiming, “See, nobody needs better CPUs.”
A walking contradiction. NUVIA saved Qualcomm, not the other way around. Samsung, MT, NVIDIA, AMD or Intel even, could’ve bought NUVIA out. They were fine.
You should distinguish between what he says and what he really thinks/does. Publicly, they're always going to put a positive spin on whatever happened. They have to, unless it's so bad (like Intel's 10 nm delays) that there no way to hide what a debacle it is.
Honest leadership can be positive, too. Look at Dr. Lisa Su when they launched Zen. Look at Arm when they launch new cores. Do any of them sulk, “Yeah, this is good enough. Don’t expect we need to much faster for Windows / Android” ?
Amon comes from a long line of arrogance at Qualcomm. See their Court battles, mass royalties, monopolization, etc.
Stop lying to your customers and investors: turns out, this option has no negative consequences.
They devoted CPU design resources to Centriq rather than mobile which was canned during the Broadcom hostile takeover bid for Qualcomm to appease shareholder demand for lower R&D costs. The part was very good in terms of efficiency:
but those gains at the SoC level would get diluted by other components and by accelerator card power budgets (which is why a pivot to AI100 PCI-e cards which go dozens into a rack and better multiplies their expertise in power efficiency makes sense.)
So it was probably the least promising of their initiatives in terms of long term value at that time since even though margins were incredibly high, competition from AMD in the x86 space with sticky software stacks would quickly erode margins and hyperscalers were starting to do their own in-house SoC designs based on ARM IP.
Nuvia started their business specifically in server CPU parts to attract venture capital because Intel had such artificially high margins and with typical SoCs sporting GPUs, ISP/DSP, modems etc. they could not realistically do anything else with their specific team expertise, so I honestly don't think they much care that it's servers in particular, just something that could leverage their bread and butter CPU core IP design prowess. However, with the trouble companies are having booking leading edge foundry capacity, hyperscalers doing in-house SoC designs and newer servers integrating ever more disciplines outside of the CPU core IP itself, they would find it hard to distinguish themselves if they had to do a part at 10nm, without an integrated FPGA, on package cache tiles, or specialty high bandwidth interconnects. A sale to someone like Qualcomm seems win-win given this.
So, you think he's all hype? Or was his point that it was pre-release and its UI issues are probably fixable via software? Or maybe he was more interested in just being in the game and on a *trajectory* for competing with iPhone?
To some extent, a CEO does need to deliver the hype. However, what counts is that they have a good management team behind them, that can deliver the goods. I've heard a CEO is more an external-facing role. They have a lot of influence on the company, but the most important thing they do is present a good face and message to customers and investors. And history has shown that investor relations is not something to be taken lightly, at Qualcomm.
I just wonder what innovations Nuvia can possibly have that make them so much better than everyone else. I know it's staffed by some ex-Apple engineers, but they supposedly left that company with some idea(s) that even it isn't doing.
I see two patents from Nuvia:
US20210151549A1: Integrated System with Power Management Integrated Circuit Having On-Chip Thin Film Inductors
US10911048B1: Dynamically adjustable CMOS circuit
Of these, the second sounds much more intriguing. I'm not too sure what to make of it, though. I wonder if their whole reason for founding the company could be based on such a low-level tweak, or if it's just one of many tricks they had stored up.
I think the latest Jim Keller interview noted if you start with an old / bad design and try to iterate (aka “scale” in marketing), you’ll hit constraints sooner and can’t make large enough leaps per generation.
He definitely seems to be a clean sheet proponent and I think that’s what NUVIA is trying as well. Not new as in “unimaginably new architectures”, but new as in “let’s imagine we can start from scratch and build it for high-perf.”
So, the assumption is that Arm Ltd. is very much simply iterating its cores and not starting fresh enough often. A76, A77, A78, etc.
And, if you know you’re designing for high-performance (via PPA), then maybe your fresh architecture begins differently. IIRC, X1 is just a larger + wider A78. It’s not a new design.
> At a high level, the design could be summed up as being a ultra-charged A78 – maintaining the same functional principles, but increasing the structures of the core significantly in order to maximize performance.
This brings up a great point: where are NUVIA’s small cores? Are they pushing a big-core-only laptop?
Most, but not all, in the Arm space advocate for big.LITTLE approaches to save battery life / improve efficiency. Like, I already have the experience of a 50W PL2 on my lap. I don’t like it.
Or back to Arm Ltd. for efficiency? IIRC, Phoenix (NUVIA’s core) is Armv8, so the Arm v9 a510 isn’t a drop-in little core. Back to A55? I don’t even know how Windows uses A55 cores. Are they useful enough?
Would be hilarious if Intel laptops are big.LITTLE in 2023 and Qualcomm releases a big-only Arm CPU.
And the final question: which OEMs will use the NUVIA/QC SoC? Microsoft, Samsung, and HP seem to be the biggest WoA proponents. And yes, current Samsung WOA devices use QC SoCs!
> I think the latest Jim Keller interview noted if you start with an old / bad design and try to > iterate, you’ll hit constraints sooner and can’t make large enough leaps per generation.
I thought that was more in reference to ISA, where you simply can't shake the legacy stuff. Sure, it can apply to other cases, but Apple or ARM can do a new design at least as easily as Nuvia.
> Arm Ltd. is very much simply iterating its cores and not starting fresh enough often.
But isn't the core after A710 supposed to be a new design?
>> At a high level, the design could be summed up as being a ultra-charged A78
The A76 is considered extremely close to clean-sheet, which invalidates your entire post, as it is still an extremely recent launch.
And no, there are plenty of cases where feeding the execution units better might as well be a fresh design. If we're just gonna cite Keller like it's gospel, he discusses how the execution units are just a functional block, and not where the innovation happens anymore.
That’s exactly what I wrote. A76 is a clean sheet design from 2018 for mid cores. What is the X-1 and X-2 lineage? We’re 4 years later and a single design is now being tweaked beyond its limits for large enough generational improvement.
The X line should’ve been clean designs half a decade ago. Arm moved slow, Qualcomm flopped, Samsung gave up, NVIDIA lost hope. Then comes NUVIA and people are like, “Oh, actually, we could’ve done that too, but like, we just didn’t want to.”
Huh? Nobody said each year needs a clean design. But, by Zeus, you’re going to make your high-performance cores just…faster middle cores? Anyone could’ve told Qualcomm (and many did): “Yeah, this is stupid. Let’s not put this into a laptop.”
It’s been clear for ages: Arm was never going to stay in phones. As soon as the 1T perf gap closed in on mainstream x86, OEMs were interested in jumping ship from Wintel.
Why do people always wildly misinterpret posts with the least charitable view and expect the authors to defend their outlandish claims?
The issue was that the single A76 core design had to pull triple duty as a phone, laptop, and server core in various SoC configurations. If they were doing something targeted to just consumer laptops, the pipeline could be a bit longer, decode and ROB wider, and the cache hierarchy could reasonably end at L3 with bigger banks at L1 and L2.
The first gen. Nuvia design will likely be a repurposed server design without any corresponding little cores, and possibly more uncore complexity than it needs. I doubt we'll see Nuvia cores in phones for a while until they get low power small cores with an uncore going, and I doubt they'll need more than just one of those big cores a phone or VR headset for benchmarketing purposes as the other blocks, ISP/DSP & GPU do most of the heavy lifting in low power use cases:
It looks like Qualcomm is slowly adopting the Intel behavior from 10 years ago, that lead them in their current ridiculous position where the fall is only slowed down thanks to inertia.
Qualcomm better have some amazing upcoming updates very soon to justify Nuvia's acquisition because if the non-apple competitivity takes off, the switch will be WAY faster than what happens in the x86 static world.
Outside of the CPU complex, Qualcomm achieves as good if not better PPA than any other vendor on the same process with their custom ISP/DSP, GPU and uncore all while integrating as many or more disciplines like modem (and WiFi + Bluetooth now) onto a single chip. AX SoCs not a sporting a modem (a very very hard block to do at low power) allows them to use this extra die area for bigger CPU/GPU and caches; that the CPU core complex moved from a 3 level cache hierarchy to 2 level with much bigger sizes starting with the A11 helped them tremendously as well.
If Qualcomm wants to have the best ARM SoC's, they also need a completely closed ecossystem with proprietary everything so that they make up for the lower hardware sale margins that come with using the most expensive silicon and completely ignoring PPA in favour of ultimate performance. Oh, and a brainless cult willing to finance all of that and surrender technological diversity and competition in favour of a monopoly by a Friendly Evil Megacorporation.
Doubt that will work in any market other than America tho.
We’ve updated our terms. By continuing to use the site and/or by logging into your account, you agree to the Site’s updated Terms of Use and Privacy Policy.
Facebook
Twitter
RSS
36 Comments
Back to Article
defaultluser - Friday, July 2, 2021 - link
Yeah, until it gets beaten consistently for a few quarters, and you end-up back in the dust bin with Kry o 820. and later the Centriq.Yojimbo - Friday, July 2, 2021 - link
This is probably the main reason Qualcomm is against the NVIDIA purchase of ARM while Mediatek is for it. Qualcomm doesn't want more money pumped into ARM to make more competitive CPU cores, as that would allow Qualcomm's competitors (which are Mediatek and soon Unisoc) to build SoCs that are more competitive with Qualcomm's Nuvia-based SoCs. A less active ARM makes Qualcomm more dominant.plopke - Friday, July 2, 2021 - link
A bit off topic.Lot of politicians are complaining the big tech is too big but the last 20 years of my life i see them buying up companies none stop. For me this is a other example were I am just please stop. But then promises are made , hands are shaken and few years down the line people go "How did we let them get so powerfull and market dominating"
And yes I am aware halting Nvidia would feel very unfair since everyone else got away with it but I personally find you have to start at some point.
mode_13h - Friday, July 2, 2021 - link
I don't see a very strong case for the US government to deny Nvidia's acquisition of ARM. I do think it's probably bad for most of ARM's customers, in the long run, and therefore the entire ARM ecosystem. But, that's not really the government's purview.It's not as if they're buying out a direct competitor, which is usually the grounds for rejecting mergers & acquisitions. There are really just a couple small overlaps between them.
lmcd - Friday, July 2, 2021 - link
Honestly what are you talking about? Mali graphics covers over half of the graphics solutions on the market at this point.Yojimbo - Friday, July 2, 2021 - link
Arm's Mali hardly competes with Nvidia's graphics. Who exactly has a choice of implementing Mali graphics or Nvidia graphics and will be losing choice if Nvidia buys Arm? Practically speaking, Arm's competitive offerings will be made better because Nvidia's graphics IP will eventually be offered to customers as well and it will be preferable to Mali in many instances. But that won't come as a result of loss of choice in the market. And it's not like there aren't other graphics providers besides Arm or Nvidia in the business sectors where each of them operate, anyway.mode_13h - Saturday, July 3, 2021 - link
> Mali graphics covers over half of the graphics solutions on the market at this point.Nvidia only slightly overlaps with their top-tier offering, and hasn't traditionally offered it as standalone IP (although they've announced a deal with Mediatek that amounts to that). It's not obvious whether Nvidia can or has any desire to try and scale down its solutions to compete with Mali's lower tiers.
mode_13h - Saturday, July 3, 2021 - link
Also, I've read that the top teir Mali isn't terribly popular. I think Mediatek was its main customer.at_clucks - Saturday, July 3, 2021 - link
Buying a competitor from the same field (graphics) but not the same market (mobile vs. datacenter, or highend vs. lowend) is even worse as it gets you closer to having the entire market.Yojimbo - Friday, July 2, 2021 - link
Don't confuse the Arm ecosystem with Arm's customers. Firstly, I think NVIDIA buying Arm is good for most of Arm's customers compared to any reasonable alternative. But it's definitely good for the ecosystem. To say one thinks it's bad for the ecosystem is to say that one thinks NVIDIA is going to make a mistake. NVIDIA certainly has no intention of damaging the ecosystem of their purchase. They don't have any reason to shut it down (sometimes companies do buy properties to shut them down). It's certainly good for the expansion of Arm into new areas such as Arm in the datacenter and Arm in OpenRan. And actually, it doesn't matter so much if it's "good" for Arm's current customers or not. A concern might be if it were particularly disruptive for a significant segment. But the basic question is whether it causes a problem for free and fair competition in the market as a whole.mode_13h - Saturday, July 3, 2021 - link
> Don't confuse the Arm ecosystem with Arm's customers.I don't, but if the deal drives away enough customers, it's very likely bad for the ecosystem, also.
> They don't have any reason to shut it down
There are other ways they could damage the market, such as by offering products competing with what a lot of ARM's customers are building, and that could scare those customers into feeling like they might end up at a disadvantage. Everyone knows how ruthlessly competitive Nvidia is.
> the basic question is whether it causes a problem for free and fair competition
As I said, I don't really see the case that it inherently damages competition, as long as ARM's customers have other options (which I think they do).
Yojimbo - Saturday, July 3, 2021 - link
It doesn't drive away customers. Customers are going to continue to use Arm like they've been and they'll switch to RISC-V in parts of the embedded space just like they've been.NVIDIA releasing more products to the market is not going to damage the market.
Don't worry, companies don't need safe spaces.
The bottom line is, if no one buys Arm, Arm is in trouble. If Arm is controlled by a consortium of its customers, Arm is in trouble. People forget that the money pumped into Arm by Softbank over the past 5 years was an investment by Softbank. And since Softbank wants to unload Arm for about what they paid for it it was an investment that didn't pay off. All these great cores that enabled Graviton and enabled so many companies to forgo designing their own Arm cores are a result of that investment. Perhaps everyone would like an entity like Softbank to own Arm but that's like asking for a sugar will not exist any more. Under NVIDIA, the Arm ecosystem is likely to increase more than it did under Softbank. Softbank set the stage for the Arm ecosystem to grow, but NVIDIA is going to do the yoeman's work of building the foundation for the ecosystem. NVIDIA is not going to start making smartphone SoCs. Right now Arm in the data center is controlled by Amazon, and that would be a horrible thing if it were to go that way, to be controlled by the hyperscalers. It would be like being controlled by Apple. Each cloud company creates their own locked-in ecosystem and has a domineering relationship with any small supplier.
I think most people on message boards who are against the merger have an idealized view of the world and of Arm. Open source software did nothing to prevent public cloud lock in. An independent Arm will similarly do nothing in that regard. An independent Arm is unlikely to build an independent server ecosystem outside the clouds.
Yojimbo - Friday, July 2, 2021 - link
If you mean your post is off topic to mine and to the article, then yes, I agree. What I said is on topic to the article:"In statements to Reuters, Amon had made comments regarding the company’s future CPU roadmap, which come to further contextualise the company’s completed acquisition of NUVIA last March.
"We needed to have the leading performance for a battery-powered device," Amon said. "If Arm, which we've had a relationship with for years, eventually develops a CPU that's better than what we can build ourselves, then we always have the option to license from Arm.""
I was not talking about the regulatory situation concerning NVIDIA and Arm. I was talking about the reason Qualcomm is likely against the merger. But I will reply to what you said anyway. In fact what is going on now with the new administration is the bureaucracy flexing its muscle to show the corporations who is the boss. It's a sort of saber rattling. Not that they wouldn't think it's heroic to pull off some big breakup. It has nothing to do with innovation or consumer protection. If government regulators and politicians really were concerned about companies being too powerful in the marketplace then they wouldn't let subscribestar be attacked the way it was. They wouldn't let gab be attacked the way it was. As far as decision on M&A, it's definitely not up to regulators to choose winners or losers. A regulator should look to see if a merger causes a trust or otherwise interferes in a free market. It's not about fear, or about predicting the future. The minute they think that way is the minute there is no free market to protect. NVIDIA buying ARM in no way creates a trust. Neither does it create a company that is particularly large overall. Perhaps regulators shouldn't be allowing facebook to buy whatsapp or instagram. If NVIDIA were buying up all the AI chip startups that would be a problem. The current tech giants (Google, Facebook, Apple, Amazon) buying up all the little companies that could turn into real competition for them is a problem. For example, if Zoom didn't explode like it did because of the pandemic perhaps someone would have bought it. It had a market cap of $20 billion at the start of 2020 and not it's about $120 billion. Such a purchase would have gone through no problem. If NVIDIA down the road is acting anti-competitively towards the market then the regulators will do what they love and punish them for it, as they should. But as is now, it's just the equivalent of making the market to block such a merger, not protecting it. The deal should go through and the deal with go through. Because regulators 1) understand the situation that no real trust is being created, 2) would rather get involved in a heroic punishment that they can defend than a controversial block that gives them no glory and which they could be forced to impossibly defend themselves against in the future (suppose Arm fades away after the block the deal, and Intel, for example, takes a bigger bite of the market, there would be little way for regulators to defend themselves). There are political reasons non-US regulators might have for wanting to block the deal, but they will probably think better of them.
mode_13h - Saturday, July 3, 2021 - link
> In fact what is going on now with the new administration is the bureaucracy flexing> its muscle to show the corporations who is the boss.
In the lawsuit against Facebook just thrown out by SCotUS, 48 States joined the Federal Government, filed ** last December **. So, no. The backlash against Big Tech is not something new or specific to Biden.
The fact is that big platform companies have more control over our lives than ever, and the people need to have more say over that, since the market has failed to provide the necessary levers of control. That's what it boils down to, and where people of most political persuasions would agree.
> If NVIDIA down the road is acting anti-competitively towards the market then
> the regulators will do what they love and punish them for it
Huh? I don't think there's anything they *can* do. I certainly can't think of any examples, in recent history. After the acquisition goes through, the only recourse is for someone to sue for damages, due to unfair practices.
Yojimbo - Saturday, July 3, 2021 - link
Firstly, the bureaucracy has a will of its own. Secondly, just because there was already a lawsuit says nothing about growing calls for regulation. If you pay attention to the way Biden's administration or Congressional Democrats are talking you will see that clearly. Thirdly, regulation is not in general a bad thing, it is a necessary thing. It's just that regulation can become its own reward if the system gets corrupted.Of course there's something they can do. They can break up a company, they can prevent a company from doing certain things. It happens to Microsoft. It almost happened to Qualcomm but the Trump admin stepped in and prevented it because Qualcomm was viewed being of strategic national importance. As far as for Google, Amazon, Facebook, these companies are too newly dominant and people don't understand the ways in which they are dominant. Things take time. There almost certainly will be regulation of them in the future. If you start trying to prevent the situation from occurring and guessing winning or losers you are then by necessity picking winners or losers and you no longer have a free market. Regulators can't be stock pickers. Not only that but there is nothing illegal about a monopoly. What is illegal is using a monopoly for anti-competitive practices. Now regulators will step in to prevent a monopoly from forming when the combined companies would actually create a monopoly immediately as the companies are combined. But to say "oh this company is going to be too successful and then later they will become a monopoly and then they are going to abuse that monopoly" is not, and should never be, the job of regulators.
demian_thorne - Saturday, July 3, 2021 - link
Question: You imply that the main benefit of the ARM acquisition by NVDA is that more money will be pumped to ARM and the products will become better.But ARM is already valued at let’s say $40B. Are you saying a company with 7000 employees and 40B valuation needs “money” to make the product better?
Ok, how about this. Why don’t they do an IPO themselves? or just go ask investors for private money?
And what stops their competitors or pouring themselves more money to their own products to make them better. Does INTC lack money ? Do you think the stagnation of INTC is due to lack of resources?
I am sorry but I cannot really follow the reasoning of why the money that will be pumped by NVDA are the only money on the planet available for that or why just money will produce better products.
I really hope the NVDA lawyers come up with better arguments :)
Care to explain what am I missing?
Yojimbo - Saturday, July 3, 2021 - link
Arm is valued at $12 billion plus $28 billion of NVIDIA stock by NVIDIA. That is the amount NVIDIA is paying for the money in stock and cash. It's a bet they are making, not anyone else. That $12 billion in cash and the $28 billion in stock goes to Softbank, not to Arm. That is not cash that Arm has on hand or would otherwise have on hand for its operations. It's like the difference between a transfer fee and a player salary in soccer. If Man Utd buys Jadon Sancho for a $100 million transfer fee it's not like Jadon Sancho is getting $100 million or otherwise would have $100 million on hand if he stayed at Borussia Dortmund.Company valuation and company operations are two different things and it makes no sense to talk about dollar values from one and apply it to another. Company operations rely on credit worthiness, revenue, and expenditures. An independent Arm would have short-term financial demands placed on them. Company must display profitability in the open market or their stock will tank. If their stock tanks then investors lose money and the company management team will at best receive much less compensation and at worst be replaced.
Private investors similarly want to have their investment paid off. They could be doing a million other things with their money. It's not easy to find a private investor to outbid $40 billion, anyway.
Additionally, NVIDIA has the experience, expertise, and drive to expand Arm into the data center and OpenRan. It takes more than just throwing money at something to do something. It takes vision and skill, and it takes the willingness to try.
NVIDIA doesn't need lawyers to come up with such arguments. That's not what regulation is about. The regulators' business is not to try to make the strongest Arm. Various UK entities might be worried about such a thing. But regardless, they already know that money doesn't come from thin air. They already know about company operations. They also know that Arm has a small revenue and a small profitability. So it's not a hard argument to make to them.
demian_thorne - Saturday, July 3, 2021 - link
Thank you for the lesson in finances and company valuations. I appreciate it, even though I still do not feel I need it.Why do I feel you didn’t really answer my question though? You certainly went out of your way to verbiase on details but I do not think you touched the gist... I will leave it at that....
metafor - Friday, July 2, 2021 - link
I was at Qualcomm back in the days during Krait development. I remember this guy. This was before Android seriously took off and the iPhone was both the gold standard and also something everyone took pot-shots at.He was touting an HTC phone that ran Windows Mobile (not Windows Phone yet) and had HTC's skin on it. It used one of the very first Snapdragons (Scorpion CPU based back then). He talked up how open it was to side-loading apps.
Someone mentioned how laggy the UI was -- and it was very laggy, definitely not an optimized for touch device. At the time anything non-iPhone had a very stuttery, delayed touch UI.
His answer was "who gives a shit?" in a half-joking, half-dismissive manner.
ikjadoon - Friday, July 2, 2021 - link
I can’t (obviously) backup your employment claims, But, Amon has 100% been dismissive of claims QC failed to innovate on our CPU performance.I mean, they ran tail between their legs to Arm Ltd for cores! The delusion!
Amon in 2018: “So as the hardware has evolved we realize that for single threaded performance on the CPU, we are in a place that is good enough for all the applications that are needed, and the vectors of innovation needed to come from other areas.”
https://www.anandtech.com/show/13703/talking-snapd...
Foolish CEO. Really? You thought people *liked* the ancient Windows on Arm device performance? Nope. We liked WoA in spite of the terrible performance to any modern x86 laptop.
You can’t fix these kinds of delusions. When / If NUVIA falters in a few years, he’ll be back to claiming, “See, nobody needs better CPUs.”
A walking contradiction. NUVIA saved Qualcomm, not the other way around. Samsung, MT, NVIDIA, AMD or Intel even, could’ve bought NUVIA out. They were fine.
mode_13h - Friday, July 2, 2021 - link
You should distinguish between what he says and what he really thinks/does. Publicly, they're always going to put a positive spin on whatever happened. They have to, unless it's so bad (like Intel's 10 nm delays) that there no way to hide what a debacle it is.ikjadoon - Friday, July 2, 2021 - link
Honest leadership can be positive, too. Look at Dr. Lisa Su when they launched Zen. Look at Arm when they launch new cores. Do any of them sulk, “Yeah, this is good enough. Don’t expect we need to much faster for Windows / Android” ?Amon comes from a long line of arrogance at Qualcomm. See their Court battles, mass royalties, monopolization, etc.
Stop lying to your customers and investors: turns out, this option has no negative consequences.
Raqia - Friday, July 2, 2021 - link
They devoted CPU design resources to Centriq rather than mobile which was canned during the Broadcom hostile takeover bid for Qualcomm to appease shareholder demand for lower R&D costs. The part was very good in terms of efficiency:https://blog.cloudflare.com/arm-takes-wing/
but those gains at the SoC level would get diluted by other components and by accelerator card power budgets (which is why a pivot to AI100 PCI-e cards which go dozens into a rack and better multiplies their expertise in power efficiency makes sense.)
So it was probably the least promising of their initiatives in terms of long term value at that time since even though margins were incredibly high, competition from AMD in the x86 space with sticky software stacks would quickly erode margins and hyperscalers were starting to do their own in-house SoC designs based on ARM IP.
Nuvia started their business specifically in server CPU parts to attract venture capital because Intel had such artificially high margins and with typical SoCs sporting GPUs, ISP/DSP, modems etc. they could not realistically do anything else with their specific team expertise, so I honestly don't think they much care that it's servers in particular, just something that could leverage their bread and butter CPU core IP design prowess. However, with the trouble companies are having booking leading edge foundry capacity, hyperscalers doing in-house SoC designs and newer servers integrating ever more disciplines outside of the CPU core IP itself, they would find it hard to distinguish themselves if they had to do a part at 10nm, without an integrated FPGA, on package cache tiles, or specialty high bandwidth interconnects. A sale to someone like Qualcomm seems win-win given this.
mode_13h - Friday, July 2, 2021 - link
So, you think he's all hype? Or was his point that it was pre-release and its UI issues are probably fixable via software? Or maybe he was more interested in just being in the game and on a *trajectory* for competing with iPhone?To some extent, a CEO does need to deliver the hype. However, what counts is that they have a good management team behind them, that can deliver the goods. I've heard a CEO is more an external-facing role. They have a lot of influence on the company, but the most important thing they do is present a good face and message to customers and investors. And history has shown that investor relations is not something to be taken lightly, at Qualcomm.
sonicmerlin - Monday, July 5, 2021 - link
And it still lags to this day. They just do a better job of hiding it behind visual flair.mode_13h - Friday, July 2, 2021 - link
I just wonder what innovations Nuvia can possibly have that make them so much better than everyone else. I know it's staffed by some ex-Apple engineers, but they supposedly left that company with some idea(s) that even it isn't doing.I see two patents from Nuvia:
US20210151549A1: Integrated System with Power Management Integrated Circuit Having On-Chip Thin Film Inductors
US10911048B1: Dynamically adjustable CMOS circuit
Of these, the second sounds much more intriguing. I'm not too sure what to make of it, though. I wonder if their whole reason for founding the company could be based on such a low-level tweak, or if it's just one of many tricks they had stored up.
ikjadoon - Friday, July 2, 2021 - link
IIRC, they left Apple to build server cores.I think the latest Jim Keller interview noted if you start with an old / bad design and try to iterate (aka “scale” in marketing), you’ll hit constraints sooner and can’t make large enough leaps per generation.
He definitely seems to be a clean sheet proponent and I think that’s what NUVIA is trying as well. Not new as in “unimaginably new architectures”, but new as in “let’s imagine we can start from scratch and build it for high-perf.”
So, the assumption is that Arm Ltd. is very much simply iterating its cores and not starting fresh enough often. A76, A77, A78, etc.
And, if you know you’re designing for high-performance (via PPA), then maybe your fresh architecture begins differently. IIRC, X1 is just a larger + wider A78. It’s not a new design.
> At a high level, the design could be summed up as being a ultra-charged A78 – maintaining the same functional principles, but increasing the structures of the core significantly in order to maximize performance.
https://www.anandtech.com/show/15813/arm-cortex-a7...
ikjadoon - Friday, July 2, 2021 - link
This brings up a great point: where are NUVIA’s small cores? Are they pushing a big-core-only laptop?Most, but not all, in the Arm space advocate for big.LITTLE approaches to save battery life / improve efficiency. Like, I already have the experience of a 50W PL2 on my lap. I don’t like it.
Or back to Arm Ltd. for efficiency? IIRC, Phoenix (NUVIA’s core) is Armv8, so the Arm v9 a510 isn’t a drop-in little core. Back to A55? I don’t even know how Windows uses A55 cores. Are they useful enough?
Would be hilarious if Intel laptops are big.LITTLE in 2023 and Qualcomm releases a big-only Arm CPU.
And the final question: which OEMs will use the NUVIA/QC SoC? Microsoft, Samsung, and HP seem to be the biggest WoA proponents. And yes, current Samsung WOA devices use QC SoCs!
mode_13h - Friday, July 2, 2021 - link
> I think the latest Jim Keller interview noted if you start with an old / bad design and try to> iterate, you’ll hit constraints sooner and can’t make large enough leaps per generation.
I thought that was more in reference to ISA, where you simply can't shake the legacy stuff. Sure, it can apply to other cases, but Apple or ARM can do a new design at least as easily as Nuvia.
> Arm Ltd. is very much simply iterating its cores and not starting fresh enough often.
But isn't the core after A710 supposed to be a new design?
>> At a high level, the design could be summed up as being a ultra-charged A78
You're talking about the X1? Bad example, IMO.
lmcd - Friday, July 2, 2021 - link
The A76 is considered extremely close to clean-sheet, which invalidates your entire post, as it is still an extremely recent launch.And no, there are plenty of cases where feeding the execution units better might as well be a fresh design. If we're just gonna cite Keller like it's gospel, he discusses how the execution units are just a functional block, and not where the innovation happens anymore.
ikjadoon - Friday, July 2, 2021 - link
That’s exactly what I wrote. A76 is a clean sheet design from 2018 for mid cores. What is the X-1 and X-2 lineage? We’re 4 years later and a single design is now being tweaked beyond its limits for large enough generational improvement.The X line should’ve been clean designs half a decade ago. Arm moved slow, Qualcomm flopped, Samsung gave up, NVIDIA lost hope. Then comes NUVIA and people are like, “Oh, actually, we could’ve done that too, but like, we just didn’t want to.”
Huh? Nobody said each year needs a clean design. But, by Zeus, you’re going to make your high-performance cores just…faster middle cores? Anyone could’ve told Qualcomm (and many did): “Yeah, this is stupid. Let’s not put this into a laptop.”
It’s been clear for ages: Arm was never going to stay in phones. As soon as the 1T perf gap closed in on mainstream x86, OEMs were interested in jumping ship from Wintel.
Why do people always wildly misinterpret posts with the least charitable view and expect the authors to defend their outlandish claims?
Raqia - Friday, July 2, 2021 - link
The issue was that the single A76 core design had to pull triple duty as a phone, laptop, and server core in various SoC configurations. If they were doing something targeted to just consumer laptops, the pipeline could be a bit longer, decode and ROB wider, and the cache hierarchy could reasonably end at L3 with bigger banks at L1 and L2.The first gen. Nuvia design will likely be a repurposed server design without any corresponding little cores, and possibly more uncore complexity than it needs. I doubt we'll see Nuvia cores in phones for a while until they get low power small cores with an uncore going, and I doubt they'll need more than just one of those big cores a phone or VR headset for benchmarketing purposes as the other blocks, ISP/DSP & GPU do most of the heavy lifting in low power use cases:
https://twitter.com/ID_AA_Carmack/status/131867574...
rmfx - Friday, July 2, 2021 - link
It looks like Qualcomm is slowly adopting the Intel behavior from 10 years ago, that lead them in their current ridiculous position where the fall is only slowed down thanks to inertia.Qualcomm better have some amazing upcoming updates very soon to justify Nuvia's acquisition because if the non-apple competitivity takes off, the switch will be WAY faster than what happens in the x86 static world.
lmcd - Friday, July 2, 2021 - link
Slowly? Qualcomm has looked awful basically since the Snapdragon 800.Raqia - Friday, July 2, 2021 - link
Outside of the CPU complex, Qualcomm achieves as good if not better PPA than any other vendor on the same process with their custom ISP/DSP, GPU and uncore all while integrating as many or more disciplines like modem (and WiFi + Bluetooth now) onto a single chip. AX SoCs not a sporting a modem (a very very hard block to do at low power) allows them to use this extra die area for bigger CPU/GPU and caches; that the CPU core complex moved from a 3 level cache hierarchy to 2 level with much bigger sizes starting with the A11 helped them tremendously as well.Wereweeb - Saturday, July 3, 2021 - link
If Qualcomm wants to have the best ARM SoC's, they also need a completely closed ecossystem with proprietary everything so that they make up for the lower hardware sale margins that come with using the most expensive silicon and completely ignoring PPA in favour of ultimate performance. Oh, and a brainless cult willing to finance all of that and surrender technological diversity and competition in favour of a monopoly by a Friendly Evil Megacorporation.Doubt that will work in any market other than America tho.