ARM deal nears closure with Nvidia mulling $40B purchase from SoftBank
Softbank and Nvidia are reportedly close to making a deal over the sale of ARM Holdings, one that could be finalized in the next week and could see the British chip design firm handed over for more than $40 billion.
Nvidia and Softbank have been in talks for several weeks over the matter, with initial murmurs of a sale or IPO of ARM by Softbank in July followed later in the month by the two entering "advanced talks." Over a month later, the two sides are apparently getting very close to closing a deal.
According to people familiar with the talks speaking to the Wall Street Journal, a deal could be made early next week, so long as it doesn't hit a sudden roadblock. As to the value of the deal, it is expected to be a cash and stock transaction that would be for more than $40 billion.
For Softbank, the deal would be beneficial as it would make a considerable profit. The Japanese communications giant acquired ARM in 2016 for $32 billion, which would mean Softbank will have earned at least $8 billion in profit over four years from the acquisition.
It is claimed Softbank CEO Masayoshi Son has been working on the deal with a small team of executives, including ARM CEO Simon Segars, CFO Yoshimitsu Goto, Vision Fund CEO Rajeev Misra, and Vision Fund executive Akshay Naheta.
Initial reports on a possible sale of ARM alleged Apple was approached for potential bid, which apparently reached preliminary talks but stalled. As Apple licenses the ARM chip architecture used in its A-series SoCs, it seemed plausible for Apple to have an interest, but report sources suggested it would be a poor fit with the rest of the company's business structure.
A purchase of ARM by Nvidia would give the graphics chip producer access to more patents and intellectual property to enhance its own offerings, as well as giving itself more of an opening to move deeper into processor sales. However, the deal may hit regulatory hurdles, as ARM licenses its technology to many other companies, including Apple, AMD, Intel, and Qualcomm, and Nvidia's control over a vital license that its competitors need would raise questions by critics if the deal goes through.
Nvidia and Softbank have been in talks for several weeks over the matter, with initial murmurs of a sale or IPO of ARM by Softbank in July followed later in the month by the two entering "advanced talks." Over a month later, the two sides are apparently getting very close to closing a deal.
According to people familiar with the talks speaking to the Wall Street Journal, a deal could be made early next week, so long as it doesn't hit a sudden roadblock. As to the value of the deal, it is expected to be a cash and stock transaction that would be for more than $40 billion.
For Softbank, the deal would be beneficial as it would make a considerable profit. The Japanese communications giant acquired ARM in 2016 for $32 billion, which would mean Softbank will have earned at least $8 billion in profit over four years from the acquisition.
It is claimed Softbank CEO Masayoshi Son has been working on the deal with a small team of executives, including ARM CEO Simon Segars, CFO Yoshimitsu Goto, Vision Fund CEO Rajeev Misra, and Vision Fund executive Akshay Naheta.
Initial reports on a possible sale of ARM alleged Apple was approached for potential bid, which apparently reached preliminary talks but stalled. As Apple licenses the ARM chip architecture used in its A-series SoCs, it seemed plausible for Apple to have an interest, but report sources suggested it would be a poor fit with the rest of the company's business structure.
A purchase of ARM by Nvidia would give the graphics chip producer access to more patents and intellectual property to enhance its own offerings, as well as giving itself more of an opening to move deeper into processor sales. However, the deal may hit regulatory hurdles, as ARM licenses its technology to many other companies, including Apple, AMD, Intel, and Qualcomm, and Nvidia's control over a vital license that its competitors need would raise questions by critics if the deal goes through.
Comments
I would prefer that ARM reside in Japan or the UK, and not Taiwan, simply for National Security reasons.
Another thing: basic R&D like this isn't Apple's deal. It is amazing that so many people are convinced that it is. In fact, Apple doesn't do originality. Instead they take existing technology - stuff that has been around for awhile and has been proven - and incorporate them into their existing design language. At most, one could say that they excel at taking parts innovated or improved by others and using them to make new great products. But the truth is that nothing in Apple's present existence or their previous history indicates that they are capable of coming up with a "new" CPU design, or even a major advance on an existing design. Even their own CPUs, in addition to being based on the existing ARM design, were the result of acqui-hiring PA Semiconductor. Even something MUCH SIMPLER such as a fingerprint scanner, they had to buy a company that already had the tech, where Qualcomm and Samsung created their own using their own R&D departments (which is why they were able to make under-the-screen fingerprint scanners so quickly).
Apple doesn't do this because Apple - see above - Apple isn't into basic R&D. They do R&D for their own products for which it isn't in their interests to license.
As for the IPO thing ... Softbank investigated that, largely because it was the stated preference of ARM Holdings' current employees, previous owners/stakeholders as well as many of the licensees. The problem is that there is absolutely no way that ARM Holdings is going to generate anywhere near $32 billion in an IPO. ARM Holdings doesn't make/sell products. They are what a lot of folks on this board would not hesitate to call a patent troll were ARM to ever sue Apple or vice versa. Their only value is to an existing company that wants to bring their R&D in house.
Nvidia is pushing an to create an ARM-based edge computing platform (something about parallel processing on GPUs to take advantage of architectural capabilities that do not exist in CPUs because of the way that CPUs are designed to handle instructions) to sell to data centers and cloud companies. They are satisfied with the software but right now are basically running it on commodity hardware. They want to create their own custom ARM-based data center GPU hardware that is designed specifically for and optimized for their platform software and the data center workloads. If they are able to buy ARM and dedicate their R&D resources towards this design issue, they are going to dominate this market - which is very lucrative and on the verge of exploding but is also niche because it requires specialized hardware and software that is very difficult and expensive and not many companies have the expertise or capability to provide or any real way of getting it anytime soon - and this $40 billion will pay for itself many times over. But if they are not able to, then that will give the competition - which does exist - time to catch up by coming up with a similar platform but with better hardware (or software) or another approach to edge computing altogether.
But the bottom line, Nvidia's reasons for buying ARM have nothing to do with consumer devices like phones, tablets, PCs and smartwatches. Nvidia's competitors in the ARM-based enterprise hardware space have some reasons to be concerned but that is about it.
https://appleinsider.com/articles/19/12/02/a7-how-apples-custom-64-bit-silicon-embarrassed-the-industry
Apple has never been challenged in SOC's since.
"Apple doesn't do originality" is a bullshit meme.
"In the late 1980s, Apple Computer and VLSI Technology started working with Acorn on newer versions of the Arm core. In 1990, Acorn spun off the design team into a new company named Advanced RISC Machines Ltd.,[30][31][32] which became Arm Ltd when its parent company, Arm Holdings plc, floated on the London Stock Exchange and NASDAQ in 1998.[33] The new Apple-Arm work would eventually evolve into the Arm6, first released in early 1992. Apple used the Arm6-based Arm610 as the basis for their Apple Newton PDA." [ref: ] wikipedia
@cloudguy Apple has been involved with Arm for over 30 years so they do have experience and probably have been adding their own technology back into ARM licenses. Never sell their technical abilities short. Apple released revolutionary products well ahead of their time even though some of them, like the Newton, took awhile before technology and consumers were ready for them.
You don't know what a patent troll is.
ARM was originally founded November, 1990 as a joint venture between Apple, Acorn, and VLSI Technology to develop a chip for the Apple Newton, now widely regarded as the world’s first decent mobile device. Apple held a share in the company until it was sold to SoftBank in 2016.
Apple now holds a perpetual multi-use architectural ARM license, which basically means it can build whatever it wants our of it, modify and extend it, which is exactly what they've been doing. What it comes down to is that ARM is just an instruction set, not a processor design, which is why they bought PA Semiconductor with the patents and expertise to produce RISC processors, in addition to the patents Apple already holds from previous RISC ventures with IBM, Motorola, etc,.
As for originality, check out this list of innovators...
Companies with a 64-bit ARMv8-A architectural licence include Applied Micro, Broadcom, Cavium, Huawei (HiSilicon), Nvidia, AMD, Qualcomm, Samsung, and Apple.
Very good info. But at the same time I’m hoping you haven’t violated any proprietary information agreements.
Apple was one of the inventors of ARM. But I thought they sold their shares pre-iPhone because they weren't expecting the barrage of knockoffs to follow. 2016 seems very recent.
From Wikipedia
“Intrinsity’s main selling point was its Fast14 technology, a set of design tools implemented in custom EDAsoftware, for using dynamic logic and novel signal encodings to permit greater processor speeds in a given process than naive static design can offer.”
https://www.youtube.com/watch?v=ozaf1-EIafk
1. That is still an iteration of the original ARM Holdings design. It is just a better one than Qualcomm and Samsung's.
2. It is ALSO the result of Apple BUYING the PA Semiconductor company. Not Apple's own unique design. So your rejoinder only confirmed my original comment. I repeat ... there is NOTHING that indicates Apple having the ability to create its own original ARM CPU design, especially a design that maintains the power and performance advantages that the ARM Holdings and PA Semiconductor designs have.
If someone makes opinion to talk jibber-jabber about Apple's innovation ability than not worth wasting time to argue against. Over $2T valuation doesn't fall from the sky and that is Apple's ability to create products people love to buy,willing to pay. Innovation is a form of evolution and in process many smaller steps or one large step..
P.A. Semi was founded in 2003 and Apple acquired them in 2008. That processor, the A7, debuted in 2013. They existed for five years separately, then five years within Apple by the time the A7 launched. Furthermore, their work when independent was on a POWER ISA core, not ARM. It's fair to call the A7 an Apple-originated design.
As for Nvidia potentially acquiring ARM Holdings, I suspect their interest is mostly in the high-performance computing world. Right now, Nvidia GPUs are peripherals of Intel or IBM processors. They're the main compute in six of the top ten supercomputers in the world as of June. Nvidia wants an Apple-like degree of control over their own fate (a big reason Apple doesn't use Nvidia cards anymore is Apple doesn't want other companies writing kernel-space code, and Nvidia doesn't want another company writing their low-level drivers). Earlier this year, they purchased Mellanox (which does InfiniBand and other high-performance interconnect chips), then a week later they also purchased Cumulus Networks (which does datacenter interconnect operating systems). Give a GPU a Mellanox transceiver instead of a PCIe transceiver, hook it up to a Mellanox switch managed by an ARM processor running Cumulus software, and Intel and IBM are no longer in the picture.