Qualcomm, Google, and Microsoft aren't happy about Nvidia buying Arm

[nanoguy]

In brief: Arm chips are eating the world thanks to the company's unique business model that allows others to license its designs. Some of Arm's biggest customers hate the idea of Nvidia gatekeeping the tech through an acquisition, so they're urging regulators to take a careful look at the potential antitrust implications.

When Nvidia announced it was buying Arm for a whopping $40 billion, it understandably shocked a number of tech giants, as it's bound to change the semiconductor landscape as we know it and extend Arm's influence beyond smartphones and the Internet of Things. Even the company's co-founder, Hermann Hauser, believes it to be a bad idea, since Arm will most likely be unable to retain its neutrality as most of its licensees are competitors of Nvidia.

Nvidia told regulators in the UK, US, EU, and China that it is committed to maintaining Arm's open-licensing model and customer neutrality, but many of the companies that depend on that model aren't convinced. This includes Qualcomm, Microsoft, and Google, who have expressed their concerns to government authorities overseeing the acquisition. And others like Apple, Samsung, Intel, AMD, Amazon, NXP, MediaTek, and Huawei are also dependent on access to Arm's chip designs and software, which could get more expensive under Nvidia ownership.

According to a CNBC report, Qualcomm is the most adamant to see the acquisition blocked by regulators, as it believes it could lead to Nvidia gatekeeping important parts of Arm's intellectual property.

For years, Arm has been described as an "enabler of competition" thanks to its business model, which is unique in the hardware industry. The company's customers -- both big and small -- create a feedback loop that leads to continuous improvements of its chip designs and a healthy ecosystem of both standard and custom processors for a variety of applications.

Theoretically, Nvidia plans to build on that open-licensing model and add some of its own intellectual property and research and development capabilities into the mix. But more importantly, Nvidia is eyeing the "combined company's target addressable market," which it estimates at $250 billion by 2023. Essentially, it wants to dominate in pretty much every area of computing -- consumer, enterprise, and industrial.

The impact to Arm's 530 licensees could be huge, although Nvidia insists it would only help them grow their businesses. In its latest quarterly report, Arm revealed that its designs were materializing into around 842 chips per second by the end of last year, which should offer some perspective on the scale of its success. Regulators will need to go through a lengthy review process to ensure the acquisition won't affect competition, and opposition from several large tech companies spells trouble for Nvidia's ability to close the deal.

In the end, tech giants like Microsoft, Amazon, Google, and AMD are currently developing custom silicon based on Arm designs, and they don't want to see a repeat of the x86 situation where the market is closed to new entrants. This is even more important for startups, which are using free access to 75 percent of Arm's chip portfolio to spearhead hardware innovation.

Permalink to story.

https://www.techspot.com/news/88646-qualcomm-google-microsoft-arent-happy-about-nvidia-buying.html

 

[scavengerspc]

Maybe Qualcomm, Google, and Microsoft should mind their own damn business and stay out of others.

What?

I said maybe.

 

[terzaerian]

I'm pretty sure the likes of Microsoft are less upset that Nvidia might gatekeep ARM as the fact they didn't get to do it themselves.

Einstein was wrong, only three things are infinite: the universe, human stupidity, and corporate greed.

 

[Uncle Al]

You know, any one of them had the resources to purchase that company so why didn't they? Sounds more like a bunch of cry babies that need to get a new pair of big boy pants and get to work!

 

[umbala]

Maybe Qualcomm, Google, and Microsoft should mind their own damn business and stay out of others.

What?

I said maybe.

Maybe nVidia should stay in their lane and stick to making video cards.

What?

I said maybe.

 

[Wereweeb]

Maybe they should invest in RISC-V ; )

 

[punctualturn]

Besides my hate for Nvidia, I wonder one thing, what is their end game?

Nvidia doesnt need to spend 40 Billions, just to be able to make "better" ARM socs.

Second, ARM itself produces around 2 billions a year in profit, so it will take Nvidia around 20 plus years to just break even.

The math is simply not there.


So as they say, the simplest answer, must be the correct one and in this case, Nvidia, being known to be very antiplayer and anti-consumer, must have found some really nefarious loophole that will allows them to grab the whole industry by the short ones and that might be why these companies are fighting back.

So there we have it. I hope this is definitely blocked.

 

[neeyik]

ARM itself produces around 2 billions a year in profit, so it will take Nvidia around 20 plus years to just break even.

The math is simply not there.

In FY 2020, Arm generated $1.898 billion of revenue - net income will be lower than this. But the estimate of 20 years to break even assumes that this revenue doesn't change - Arm technology can be found in almost every single mobile chip phone around the world, and this market is likely to continue to expand.

As will the server/datacentre and AI/machine learning markets, and Arm already has a reasonable presence in these areas. Nvidia has heavily invested in this sector, spending $7b purchasing Mellanox. The estimate also assumes that Nvidia, if their bid is successful, doesn't alter licencing and royalty fees, and/or alter staffing levels to cut costs.

$40b might seem ridiculous, but Softbank paid $31b for Arm nearly fives years ago - which suggests that Arm's market value is fairly near the 40 mark. If the bid doesn't go through, and Softbank make the Arm sector public, then it wouldn't surprise me if the IPO is around the same figure.

 

[poohbear]

.

Second, ARM itself produces around 2 billions a year in profit, so it will take Nvidia around 20 plus years to just break even.

The math is simply not there.

so...Nvidia has it's math wrong with this acquisition huh? you don't think they looked at the numbers closely before trying to make this acquisition? lol silly Nvidia!

 

[Lionvibez]

Maybe Qualcomm, Google, and Microsoft should mind their own damn business and stay out of others.

What?

I said maybe.

Ya no ARM needs to be not owned by any of the companies it supplies tech too and to be kept vendor neutral. Far to many companies rely on their IP for one company to own them. You can bet your retirement savings NV will gate keep and increase licensing fees if they own them which will hurt the market.

 

[terzaerian]

Ya no ARM needs to be not owned by any of the companies it supplies tech too and to be kept vendor neutral. Far to many companies rely on their IP for one company to own them. You can bet your retirement savings NV will gate keep and increasing licensing fee if they own them which will hurt the market.

On the one hand I do acknowledge that it would be wise to make sure Nvidia does not become a monopoly; on the other hand, I would enjoy watching them lead the likes of Apple and Samsung around by the short hairs. Alas....

 

[Bulllee]

At first I thought the price was too low but it seems there's no bidding war but plenty of disinterested business parties standing in the wings with the reps.of quite a few governments noticeably in attendance.

 

[hahahanoobs]

This is a joke.
Wake me when Big Tech and those startups have better reasons than what they THINK will happen if it goes through. I see nothing here and I assume the source material is and has been the same since this acquisition was announced months ago. Speculation is laziness. And when it comes from the internet, the ideas are not usually unique. Parrots come to mind.

God bless the internet.

 

[hahahanoobs]

Second, ARM itself produces around 2 billions a year in profit.

With Softbank's name attached to it, yes.

 

[punctualturn]

In FY 2020, Arm generated $1.898 billion of revenue - net income will be lower than this. But the estimate of 20 years to break even assumes that this revenue doesn't change - Arm technology can be found in almost every single mobile chip phone around the world, and this market is likely to continue to expand.

As will the server/datacentre and AI/machine learning markets, and Arm already has a reasonable presence in these areas. Nvidia has heavily invested in this sector, spending $7b purchasing Mellanox. The estimate also assumes that Nvidia, if their bid is successful, doesn't alter licencing and royalty fees, and/or alter staffing levels to cut costs.

$40b might seem ridiculous, but Softbank paid $31b for Arm nearly fives years ago - which suggests that Arm's market value is fairly near the 40 mark. If the bid doesn't go through, and Softbank make the Arm sector public, then it wouldn't surprise me if the IPO is around the same figure.

With all that said, please enlighten me in how quickly AND reasonable nvidia can recoup those 40 billions without resorting to ways that will affect the whole industry?

and about softbank, they have been making some really bad investments and are “lucky” enough with ARM that nvidia is willing to offer that much.

so...Nvidia has it's math wrong with this acquisition huh? you don't think they looked at the numbers closely before trying to make this acquisition? lol silly Nvidia!

Did you read my whole post or simply grasped at it until something gave you simply enough to try to ridicule me?

 

[punctualturn]

On the one hand I do acknowledge that it would be wise to make sure Nvidia does not become a monopoly; on the other hand, I would enjoy watching them lead the likes of Apple and Samsung around by the short hairs. Alas....

Given the history of apple and Arm, looks like no matter what nvidia tries to pull, they wont affect apple. Samsung and the rest are another matter.

 

[jpuroila]

Far to many companies rely on their IP for one company to own them.

Then maybe those companies shouldn't have chosen a closed-source IP.

 

[Irata]

so...Nvidia has it's math wrong with this acquisition huh? you don't think they looked at the numbers closely before trying to make this acquisition? lol silly Nvidia!

Oh, I think they did look at the numbers closely, just that ARM‘s business model going forward most likely won‘t be what it is now.

 

[meric]

I don't like the idea of nvidia buying arm because I think in the end we, consumers, will have to deal with its consequences. It would be naive thinking to say nvidia will stay neutral and objective to their competitors. Tech should be accessible and innovation should not be held captive by giant corps. Not only for Nvidia but every other giant included.

 

[SixTymes]

Why did no one mind when AMD bought XILINX ? Not a word on that deal, in fact it got pushed through quickly. This wreaks of corruption.

 

[neeyik]

With all that said, please enlighten me in how quickly AND reasonable nvidia can recoup those 40 billions without resorting to ways that will affect the whole industry?

By the methods I mentioned before: through growth in developing sectors and reduction in costs. Here's how Nvidia's revenue has looked over the 4 financial years (Q4 FY2021 hasn't been released yet):



It's clear that datacenters are the market they're going to focus on. A key difference between Nvidia and Arm is that the latter sells just licences, whereas the former earns the bulk of its income through hardware. This sector is estimated to have potential for significant growth, especially in AI, where Nvidia's GPUs already hold the market share - now they want a slice of the CPU side of that (which they've already indicated as being a key target for them).



Naturally, the likes of Qualcomm, Samsung, et al are all concerned by this - they want to still have the same deals that they already have with Arm-Softbank. Nvidia have claimed that this operating model for Arm wouldn't change and it probably wouldn't, or not by a huge amount, as that market is too profitable for Nvidia to watch it wholesale switch to a different system.

But that's not going to placate those chip vendors, even though Nvidia isn't a market competitor for them, despite their claims. They're going to be far more concerned that the direction that Arm is forced to follow pushes it increasingly towards energy efficient AI systems, leaving the growth in mobile CPU IP to become increasingly stagnate. This means they would need to increase their own expenditure on R&D for new architectures, such as RISC-V, or worse still, develop their own from scratch.

Nvidia has a large R&D division, that's seen a steady growth in its budget for the past 5 years, and was $3.5b last year. In comparison, Softbank put $0.6b into Arm's R&D budget in 2019. So they can probably cut some costs here, though not much, by utilising their own resources where there's overlap. More important is the fact that Nvidia can afford to have Arm focusing on AI.

As to the question of timeframe, that's much harder to judge. Analysts are expecting datacenter and AI growth to be the largest of any IT sector over the next 5 years or so. Whether that's sustained or not is the hardest thing to forecast, and I don't think anyone has a really good insight into this.

All IP & business purchases come with a degree of risk (as do all investments), and as pointed out, Softbank has lost a fair bit of money with some of theirs. So far, Nvidia seems to have made sound purchases and Arm could be another one. But as Intel has shown in recent years, even pulling in huge levels of revenue doesn't guarantee long term certainty, and Arm could well be back on the market, if datacenter & AI growth isn't to the level that Nvidia hope it will be.

 

[Peter Farkas]

Maybe nVidia should stay in their lane and stick to making video cards.

What?

I said maybe.

why should they? they have been making chips and now they are making even more chips. Following your logic AMD shouldn't have been making both CPUs and GPUs... or qualcomm, or Intel...
Yours is a weird thought.

 

[Lionvibez]

Then maybe those companies shouldn't have chosen a closed-source IP.

Sure but that we are here now and cannot go back in the past to change that. So they have to work with the current situation and right now most of these companies are pushing back and for a good reason.

 

[punctualturn]

By the methods I mentioned before: through growth in developing sectors and reduction in costs. Here's how Nvidia's revenue has looked over the 4 financial years (Q4 FY2021 hasn't been released yet):

View attachment 87501

It's clear that datacenters are the market they're going to focus on. A key difference between Nvidia and Arm is that the latter sells just licences, whereas the former earns the bulk of its income through hardware. This sector is estimated to have potential for significant growth, especially in AI, where Nvidia's GPUs already hold the market share - now they want a slice of the CPU side of that (which they've already indicated as being a key target for them).

View attachment 87502

Naturally, the likes of Qualcomm, Samsung, et al are all concerned by this - they want to still have the same deals that they already have with Arm-Softbank. Nvidia have claimed that this operating model for Arm wouldn't change and it probably wouldn't, or not by a huge amount, as that market is too profitable for Nvidia to watch it wholesale switch to a different system.

But that's not going to placate those chip vendors, even though Nvidia isn't a market competitor for them, despite their claims. They're going to be far more concerned that the direction that Arm is forced to follow pushes it increasingly towards energy efficient AI systems, leaving the growth in mobile CPU IP to become increasingly stagnate. This means they would need to increase their own expenditure on R&D for new architectures, such as RISC-V, or worse still, develop their own from scratch.

Nvidia has a large R&D division, that's seen a steady growth in its budget for the past 5 years, and was $3.5b last year. In comparison, Softbank put $0.6b into Arm's R&D budget in 2019. So they can probably cut some costs here, though not much, by utilising their own resources where there's overlap. More important is the fact that Nvidia can afford to have Arm focusing on AI.

As to the question of timeframe, that's much harder to judge. Analysts are expecting datacenter and AI growth to be the largest of any IT sector over the next 5 years or so. Whether that's sustained or not is the hardest thing to forecast, and I don't think anyone has a really good insight into this.

All IP & business purchases come with a degree of risk (as do all investments), and as pointed out, Softbank has lost a fair bit of money with some of theirs. So far, Nvidia seems to have made sound purchases and Arm could be another one. But as Intel has shown in recent years, even pulling in huge levels of revenue doesn't guarantee long term certainty, and Arm could well be back on the market, if datacenter & AI growth isn't to the level that Nvidia hope it will be.

Once again, impressive wall of text, but you are missing the whole point of my question and post.

ARM wont produce that much money, that quickly, just based on their current way of operations.

As it is, they are selling gazillions of SOC's, yet they are earning "pennies" on a yearly basis.

Nvidia is making their killing there due to their proprietary anti-consumer maneuvers and will continue for the foreseeable future, but again, ARM itself due to their licensing fees are not simply making that kind of money and Nvidia cannot simply raise those fees without seriously disrupting the whole industry that depends on ARM.

Also, as usual, ignored by you and others, Nvidia doesnt need to purchase ARM to continue making SOC's, period. They already can and if they need access to more, they can pay a license that will definitely be less than 40 billions and not having to deal with other companies and the current deals they have with ARM.

So in the end, my post stands, Nvidia can only recoup that money if they do their typical anti-competitive and anti-consumer maneuvers and that's why all these other companies are against them owning ARM.

 

[mbrowne5061]

With all that said, please enlighten me in how quickly AND reasonable nvidia can recoup those 40 billions without resorting to ways that will affect the whole industry?

Well, the most immediate thought is they will get to start licensing ARM's IP to themselves for free. That right there will cut their own costs, boosting their profits. You may also see additional products from nVidia in market segments they previously haven't really touched - like mobile phone processors with really beefy GPUs.

They don't necessarily need to put the screws to their competition when they go to ARM for licensing, but nVidia likely can take advantage of their position to make themselves really competitive in any computing sector they enter.