NVIDIA officially acquires Arm for $40 billion in bid for AI dominance

NVIDIA GeForce RTX 2080 and RTX 2080 Ti
NVIDIA GeForce RTX 2080 and RTX 2080 Ti (Image credit: NVIDIA)

What you need to know

  • NVIDIA has officially started to acquire Arm Limited.
  • NVIDIA plans to use the technology to continue its lead in AI.
  • The deal is valued at $40 billion.

As expected, popular GPU and server provider NVIDIA is taking over SoftBank Group Corp. (SBG) 's share of Arm Limited. The deal's price is valued at $40 billion and was officially announced by NVIDIA (opens in new tab) on late Sunday night. In turn, Softbank gets a 10 percent stake in NVIDIA and is "committed to Arm's long-term success."

In the full press statement, NVIDIA cited multiple ways for the company to leverage Arm's research and development (R&D) and chip designs to accelerate its plans for artificial intelligence (A.I.) and modern computing solutions for enterprise and consumers.

Details in the announcement were highlighted as follows:

  • Unites NVIDIA's leadership in artificial intelligence with Arm's vast computing ecosystem to drive innovation for all customers
  • NVIDIA will expand Arm's R&D presence in Cambridge, UK, by establishing a world-class AI research and education center, and building an Arm/NVIDIA-powered AI supercomputer for groundbreaking research
  • NVIDIA will continue Arm's open-licensing model and customer neutrality and expand Arm's IP licensing portfolio with NVIDIA technology
  • Immediately accretive to NVIDIA's non-GAAP gross margin and EPS
  • Consideration of $40 billion to be met through a combination of NVIDIA shares and cash

Arm will also remain in Cambridge, UK, for the foreseeable future:

Arm will remain headquartered in Cambridge. We will expand on this great site and build a world-class AI research facility, supporting developments in healthcare, life sciences, robotics, self-driving cars and other fields. And, to attract researchers and scientists from the U.K. and around the world to conduct groundbreaking work, NVIDIA will build a state-of-the-art AI supercomputer, powered by Arm CPUs. Arm Cambridge will be a world-class technology center.

Arm Holdings is separate from companies like Qualcomm, Apple, Texas Instruments, Broadcom, and NVIDIA. The former company designs and licenses chips, while the latter companies build off those platforms through customization and addons, while also manufacturing them (called ARM with caps). It will be curious to see if this deal with NVIDIA affects the long-term market for the various companies, but the whole point of Arm Holdings is licensing of the IP as the business model. There could be some downsides too, however.

NVIDIA has quite a year with stock rallying more than 100 percent, and up over 40 percent from just May. While many know NVIDIA for its gaming GPUs, the company is very diversified with leading performance in various segments, including AI, machine learning, data centers, and self-driving cars.

Daniel Rubino

Daniel Rubino is the Editor-in-chief of Windows Central, head reviewer, podcast co-host, and analyst. He has been covering Microsoft since 2007 when this site was called WMExperts (and later Windows Phone Central). His interests include Windows, laptops, next-gen computing, and for some reason, watches. Before all this tech stuff, he worked on a Ph.D. in linguistics, watched people sleep (for medical purposes!), and ran the projectors at movie theaters because it was fun.

  • On the surface, I'm fine with this. It seems as though ARM under Nvidia would be business as usual. If anyone knows why this could potentially be bad for the market, I'd be interested to know.
  • This has a high risk to be catastrophic as Nvidia can choose to jack up the prices on SoC designs for their competitors. Whereas they can “license” the designs to themselves for next to nothing using fiscal loop holes. As a result costs get passed to down the chain and Nvidia would be able to leverage their increased market cap to assert more dominance. Which then has a domino effect on other segments and depending on various varibles there could be cascading effect of with Nvidia holding all the chips. Take the GPU market for example, 1) If Intel is not going to pursue GPUs after Dg2 as rumored and 2) AMD is not able to offer meaningful competition. Nvidia's cap will grow again and Nvidia would be free again to hike up the prices. or 1) if Intel leaves gpu space after dg2. 2) Even if AMD offers better performance in the GPU space at lower cost and yet most people further delve into Gold fish syndrone by Nvidia cards anyway thus increasing Nvidia's cap and so on. or the best case scenario happens, Intel stays in the GPU space, AMD offers better performance GPUs and people don't engage in Goldfish syndrome. Nvidia is forced to remain competitive. In that scenario we're still dependent on competition keeping Nvidia in line. In the cpu space, the ARM engineers give Nvidia more than just a foot in the door to creating their own X86/X64 space in the short term. Sure, it would be good thing to have one more player in the CPU space. In the long term, not so much especially if it ends up putting the squeeze on AMD as we need AMD to provide competition in the GPU space and for that to happen AMD has to grow so that they can increase wafer allocations for their products and so on. Additionally, long term for Nvidia it bolsters their tegra line up as well as any other ARM based offerings. and so on. There is one other factor at play here, Nvidia is directly competing with GPU partners even more so with Ampere with custom cards. Whereas previously it was just tuned “reference” cards. Throw in the fact Nvidia is both the Chip supplier and manufacturer via partner fabs. Everyone should be looking at Nvidia's actions with a fine tooth comb or at the very least be very concerned of Nvidia's ARM acquisition.
  • @TechFreak1, those are all valid assessments, but missing the long-term benefits that accrue from the quest for profit. These mitigate the potential downsides in the long run and may even turn them into positives. Specifically, let's say that Nvidia behaves badly with managing ARM and gets away with it. By "badly" I mean they work to give themselves a significant cost advantage for using ARM's designs. If that happens, they have effectively raised the price of RISC chips to us. Bad for consumers in the short run, for sure. However, this higher price also makes it a more attractive for prospective competitors (the higher price means either a competitor can come in and sell at that same higher price, making it more profitable to enter the market or it's easier to gain market share by undercutting, because it's easier to undercut a higher price without losing money than a lower one). You cite Intel for this, which is a good example here. The higher prices and margins earned by Nvidia creates an incentive for Intel to enter the market, because higher prices mean it's more profitable to compete, even without a market dominant position. So the very behavior that would be a worst-case scenario is also effectively an invitation to competitors. As Peter Drucker wrote, when in a strongly dominant position, companies feel safe to raise prices, but those high prices effectively serve as an umbrella to protect competitors who are considering entering the storm that is a competitive market, drawing them in with near certain profits (it's easy to make a profit when market prices are high).
  • @TechFreak1 Thanks for the write-up! @GraniteStateColin - When you say competitors, which chips did you have in mind? Off-hand, I can't think of any non-ARM based chips that are suitable for phones. Didn't Intel used to have one but it ran poorly?
  • @HeyCori You're welcome.
  • @GraniteStateColin
    In theory yes, in reality? Not so much. There are crucial elements here a competitor cannot easily enter this market - due to patents and licensing. Another is brand familiarity and ability to leverage such brand familiarity through marketing. ARM doesn't make SOCs they make designs which in turn are leveraged by companies to create their own product stacks. A competitor would need to have significant disposable market capital
    to stay relevant. Every company aspires to remain dominant forever. Through such aspirations forward looking patents are always filed - colloquially known as research and development . This is where disposable market capital counts. Higher profit margins become an illusion of security once a company ceases to be innovate. Especially, when security is forsaken in the pursuit of performance. As that will always come back and bite you in the rear down the line and depending on the company size the result could be extremely damaging or fatal. Additionally profits are not an umbrella that shields the competition. That presumes competitors are able to control both able to control and move that umbrella - they can't. Nor is an umbrella able to grow. Profits are in laymens terms - is a tree prone to lightning strikes. A company can either lease or buy land that the tree grows on. But they do not have complete control how the tree grows. Distilled, it's a complex ecosystem that can cause a tree to grow or wither. Ergo, why you have branches. Going back to the topic at hand, there may very well be new competitors to ARM as Apple is trying to become self reliant. Qualcomm is has also become a massive company with significant disposable market capital. However, there is at present only a handful of ways a SoC can be designed. Given most companies file broad legal actions against any sort of perceived infringement it would be extremely difficult for any new competitor to challenge ARMs dominance. Especially with a company like Nvidia at the helm. Thus in reality, not so much.
  • https://www.extremetech.com/computing/313471-arm-co-founder-sale-to-nvid...
  • This can be good or bad. It can be good because Nvidea will probably push ARM performance further and Nvidea will try to push ARM further to PC. It can be bad because Nvidea can decide to change ARMs business model. Nvidea needed to this, not because of AI, but because APUs are the future and they will kill alot of GPU segments, just look at the budget GPU market, there is only the GT1030 instead of a 710, 720, 730 and 740, and the 1030 is barelly even worth it, today there isn't even a low end 16 series GPU.
    In the future you will see not only qualcoom as the only player making ARM SoCs for PCs, Nvidea will be there too. For the people doubting the potential of APUs, just look at consoles APUs, the Series S's APU is smaller than Ryzen 3000 mobile CPUs for example.
  • Looks like Apple is getting nvidia cards again :)
  • My short take on this deal is Nvidia just became one of the premier technology companies on the planet. The momentum for ARM is already massive. The potential SOCs with native AI built in just went up by a bunch. I can see this extensively impacting edge computing over the next decade. Toss in some 5G and what is going on today begins look like old news down the line.
  • Well, this could be quite catastrophic in the long term. Time will tell.
  • The UK government tried to block it, they have raised concerns about their UK base of operations
  • I don't like consolidation like this. I would prefer they remained separate, because competition is good for customers and industry innovation. Having said that, I think some of the concerns (mostly expressed in other places, not so much here on Windows Central) don't understand competitive forces in the long run (at least if we go out 5+ years). In the very short run, this is probably a good thing, in that Nvidia will be on good behavior and keep the licensing going while also trying to get into the market itself, possible competing with Qualcomm for mobile chips or maybe just furthering its position in the AI space. That would be good. In the medium run, if they do choose to get into making mobile chips, they will probably find a way to give themselves an advantage in the ARM ownership to have a competitive cost advantage over Qualcomm. But in the long run, if they do that, the only way it's possible would mean that they are effectively raising the market price (higher external licensing costs than internal), which in turn makes it more attractive for entirely different technologies, like those from Intel. Or maybe this gives Apple a greater advantage. Or maybe some other company we've never heard of steps into the fray. Point is, that in the long run, if this move leads to higher prices (better margins at the same cost) or slower innovation (able to be leapfrogged with a less expensive R&D budget), that equates to opportunity for competitors. That's the beauty of how capitalism drives technology growth.
  • BBC News - ARM: Can 'crown jewel' of UK technology be protected? https://www.bbc.co.uk/news/business-54144434
  • It's not the UK's tech crown jewel. It's the only jewel, with a large ownership share coming from Australia. And with mini-Trumps running policy over there (and their next most likely alternative is British Francois Mitterand), who wants to HQ in the UK anymore? One wonders why Europe (the UK included) is so behind in so many kinds of tech innovation.