Nvidia Announces ARM Acquisition for $40 Billion

Nvidia ARM Press Deck
(Image credit: Nvidia)

After months of speculation and reports yesterday that a deal was imminent, Nvidia announced today that it is acquiring ARM Limited from SoftBank in a transaction valued at $40 billion. SoftBank "will remain committed to Arm's long-term success through its ownership stake in NVIDIA, expected to be under 10 percent," signaling that Nvidia will gain full control of ARM. However, the transaction does not include ARM's IoT Services Group, which remains with SoftBank.

The deal ranks as the largest semiconductor acquisition in history and marks the second-largest tech acquisition in history behind Dell's $64 billion purchase of EMC. As expected, the ARM/Nvidia deal will be subject to regulatory approval and could take up to 18 months to finalize. ARM's current licensees include industry leaders like Apple, Qualcomm, Amazon, and Samsung, among many others, and it remains to be seen how those companies will react to the acquisition. 

Nvidia stated that it would remain committed to ARM's open licensing model and customer neutrality, all while expanding ARM's IP licensing portfolio with Nvidia's GPU and AI technology. According to the press release, ARM licensees have already shipped over 180 billion chips based on the ARM architecture (22 billion in the last year), making the addition of Nvidia IP to that portfolio an enticing proposition. 

ARM's intellectual property will also remain registered in the U.K., and Nvidia will "retain the name and strong brand identity of ARM." ARM will remain headquartered in Cambridge, UK, and Nvidia announced that it would expand ARM's R&D presence there with a new AI research and education center that will house an ARM/Nvidia-powered AI supercomputer for research.

Pending regulatory approval, Nvidia will pay SoftBank a total of $21.5 billion in Nvidia common stock and $12 billion in cash, which includes $2 billion payable at signing. SoftBank could receive an additional $5 billion in cash or common stock under an earn-out construct, meaning the payment hinges on specific financial performance targets. Additionally, Nvidia will issue $1.5 billion in equity to ARM employees, which will certainly help with employee retention in the wake of the acquisition. 

That works out to $33.5 billion in guaranteed money for SoftBank, which purchased ARM for $31.4 billion in 2016.

Nvidia will gain access to a treasure trove of IP and engineering talent that could enable the company to quickly develop custom CPU architectures for its own use, which would then further the company's broadening push into the profit-rich data center market. Nvidia has already paved the way for ARM integration with its GPUs through the recent introduction of CUDA support for ARM architectures, and now the company will have control of the underlying ARM ISA, too, enabling extremely tight integration into its solutions.  

Nvidia recently purchased Mellanox for $6.9 billion and Cumulus Networks, bringing leading networking capabilities into its portfolio. The firm also recently purchased SwiftStack, a company focused on object storage software for AI and HPC computing. 

As such, Nvidia's own line of custom ARM-based data center chips could be the final piece of the puzzle that allows it to create vertically-integrated data center-scale architectures that would deliver on Jensen Huang's vision for the future of computing

Nvidia also announced a webcast on September 14 at 5:30am PT that will cover the details of the acquisition. The webcast will be available on Nvidia's Investor Relations website.

Paul Alcorn
Managing Editor: News and Emerging Tech

Paul Alcorn is the Managing Editor: News and Emerging Tech for Tom's Hardware US. He also writes news and reviews on CPUs, storage, and enterprise hardware.

  • Shadowclash10
    first :p

    OH MY. WHAT. okay then, just let me get this out of my system.

    Well. I wonder if this means x86 is gonna die. Apple moving to x86, and Nvidia acquired ARM. When do you think we'll be putting ARM components inside out systems?

    EDIT: I was thinking: we all know that Nvidia would love to enter the CPU market but can't, because of x86 rights etc etc. This is their chance. I just wonder if Jensen will have to pull CPUs out of his microwave instead cause they're smaller :P :P :P
    Reply
  • spongiemaster
    Kind of odd that Softbank sold all its Nvidia shares worth $3.3 billion less than 2 years ago (great move, stock is worth 3x as much now), only to become the largest Nvidia shareholder after this deal goes through with $21.5 billion in common stock. Seller's remorse.
    Reply
  • hotaru251
    ARM's current licensees include industry leaders like Apple, Qualcomm, Amazon, and Samsung, among many others, and it remains to be seen how those companies will react to the acquisition.

    Apple does not have a good friendship with Nvidia.

    This ought to be funny.


    TBH i expected them to get it.

    They have the $ and ppl (and the brand) to make ARM grow.

    doubt Nvidia will go into CPU like they do gpu's.
    Reply
  • gg83
    Will Nvidia get approval from all the regulatory bodies?
    Reply
  • bit_user
    I wonder if this has anything to do with SoftBank needing to somehow cover the losses on their WeWork investments.
    Reply
  • bit_user
    gg83 said:
    Will Nvidia get approval from all the regulatory bodies?
    I expect so, though I wonder if there will be any strings attached by Europeans or other foreign regulators.

    In the US, a rubber stamp would've been a foregone conclusion, until Trump decided to start wielding merger approvals as a political tool.
    Reply
  • bit_user
    Shadowclash10 said:
    I wonder if this means x86 is gonna die.
    I think it possibly worsens ARM's long-term chances, a little. When ARM was independent, there was basically no conflict of interest in licensing from them. However, since Nvidia is in competition with some of ARM's customers, it could give some ARM licensees reason to reconsider.

    Case in point: before Jim Keller left AMD, he reported designed an ARM CPU core for them. However, they mothballed it in order to focus their limited resources on finishing Zen. Even so, they already brought one ARM-based server CPU to market and could've been planning to return to that market with more, once it matured.

    I could even believe Intel has been eyeing the ARM server market, if you believe what Nuvia has been touting about ARM's inherent efficiency advantages.

    (source: https://www.anandtech.com/show/15967/nuvia-phoenix-targets-50-st-performance-over-zen-2-for-only-33-power )
    Reply
  • Zizo007
    Nvidia Tegra? I had an Nvidia Tegra tablet.

    Edit: It was a Google Nexus tablet with Nvidia Tegra SoC.
    Reply
  • hannibal
    ...
    This can be problem of companies that use arm...
    Nvdia allready did say that They will make their own custom arm.
    how long it takes Until open license arm cpus Are weak and nvidias own custom arm Are Great. It could kill the competition if nvdia only lisense old arm designs and keep the best themselves.
    well Hopefully I am wrong in this case.
    Reply
  • bit_user
    hannibal said:
    Nvdia allready did say that They will make their own custom arm.
    They've already made several generations of custom ARM cores. See: Project Denver (and Carmel)

    hannibal said:
    It could kill the competition if nvdia only lisense old arm designs and keep the best themselves.
    I don't see that. I think they'll still license highly-competitive cores, but the question I have is: at what price?
    Reply