Four years after swallowing Arm Holdings, SoftBank said to be mulling Brit chip biz sale Troubled Japanese conglomerate SoftBank is reportedly exploring the possibility of offloading Arm Holdings, the iconic British chip designer acquired in 2016 for £23.4bn (or $32bn at historic exchange rates). Reports citing internal sources claimed SoftBank is considering a full or partial sale. Goldman Sachs is said to be …
I'm not sure ARM was ever "patriated" in the first place, so can it be repatriated? Just following the prior conviction that ARM was never nationalized, so it can't be re-nationalized.
(Just trying to see whether we can peg the pedantry meter in this thread.)
This post has been deleted by its author
Arm should never have been sold to a non UK peep in the first place. If the uk gov has any sense at all (*) would repatriate it. And you could extend that sentiment to a boat-load odd other tech industry, starting with the now sadly long lost Inmos.
(*) Oh hang on - I’m seeing a flaw in my plan
ICL is still there in spirit, as well as in the person of many people who worked for ICL then and are still doing much the same thing all these years later, despite being bought by Fujitsu from the 80s through the 90s and finally eradicated as a brand in 2002.
The Putney building is now flats and a Wetherspoons.
Is that the East Putney Building or the Management Centre (Lon03-05 I think they were - it's a long time ago!). My first real job was with ICL at East Putney in 1987-88 on a work placement for my HND. After I 'diplomated' I worked for them at STE08 until Fujitsu took them over completely in 1991
No idea I'm afraid, I only haunted MAN05 and BRA01, with the occasional trip to STE10. Putney was land far off of which we knew little. This site had picture of the current building on the Putney site which added a suitable note of pathos to the tale of ICL.
That little history of ICL does lead to an interesting line of thinking about all the decent and successful ( often) companies that get bought out and then traded into oblivion. If you were to describe such a system to the objective observer from another planet I can't help thinking they'd laugh themselves silly.
The chance to do that was a couple of years ago. SoftBank now owns ARM and has already made sure that it controls all the relevant IP.
This kind of flip was always going to happen. The SNAFU of the "Vision Fund" has only brought forward the date. A bidding war is probably preferable to an IPO for SoftBank but difficult to think who could really buy the company without being ruled out over competition grounds. Maybe do a deal with a collection of the sovereign wealth funds?
The monopolies people would have a heart attack
Generally IP deals for something like ARM already have clauses for this. So if Apple buys ARM, Samsung get to keep their current licenses for ever - for free. Otherwise nobody would ever buy into any platform with the risk that it could be bought by a competitor
Everyone benefits from ARM being something of an industry standard, you get cheaper tools and cheaper more experienced people than if ARM was an internal product of company X.
Ironically no. One of the problems with ARM's business model is that you have to keep the IP license cost at 0.00001$ below what it would cost customers to just invent their own instruction set or switch to something like RISC-V. Most of Apple's secret sauce is in making the SOC and it's own based-on-ARM cores. Buying ARM (even if it could) to save on the license fee wouldn't be worth it.
Apple was an early investor in ARM and has since invested even more heavily in it. But if it wanted to buy the company, its best chance was a couple of years ago. Nowadays it's difficult imagining how the purchase could get past the competition authorities and Qualcomm would probably get the the DoD to nix it anyway.
Everyone benefits from the current structure: ARM sells designs but not the chips themselves and it sells to everybody. There are competitors both of the fabless and fab support to keep it on its toes and the market keeps developing.
They design their own ARM cores, better than the ones ARM designs. The only thing ARM does they care about is introduce updates to the ISA in the form of new instructions. They'd get nothing out of owning them aside from not paying the rather modest license fee, which would take take hundreds of years (if not more) to get payback on at the price they'd have to pay for ARM.
The ARM people are not stupid, they have priced the royalty rates where they are for a reason...
If they start cranking them up, a lot of customers would leave and move to other architectures - MIPS, RISC-V or POWER etc. Most embedded devices are not tied to any particular architecture, Linux runs on everything and the firmware is generally rebuilt for each new device anyway so their customers are generally not locked in.
The Chinese government will buy it
Taking the company over would be both a military and business strategic advance for the Chinese. We have to hope our government recognise this and prevent any overseas purchase. The company should be nationalised NOW before any more secrets are lost
While it was sad that ARM was sold a few years ago to Softbank, I don't think that the UK government buying it back if it is up for sale is actually of much use for national security grounds. ARM don't make any of the chips they design, and even if it were fully UK owned and ran that wouldn't stop the Chinese, US or anyone else producing ARM SOC if they had paid for a license.
It does appear that Softbank are not great at doing due diligence before investing, as they don't seem to understand how ARM operates and look at the fsck up that was WeWork.
Europe is every bit as desperate as China to get out from under US technological hegemony. It even has an explicit program to foster "local" tech. Oddly, that program includes ARM (see the EuroHPC projects) although that hardly fits the definition of "European" these days - nor would it even if ARM were returned to UK ownership.
And the big difference between Europe and China is this: if China decides to go its own way, it can readily do so. Its centralized political structure, its huge, homogeneous internal market, and the buying power of government, army, and other centrally-directed entities mean that if it wants to make an alternative to Intel happen, it can simply dictate it. Whether that is a "perpetual and irrevocable" license for ARM, adoption of RISC-V, or some other direction is anybody's guess.
Europe does not have those advantages. For Europe, the choices are either endorse something that already exists; or spend decades squabbling over whether the fab plants will be in Germany or France and the design headquarters in Belgium or Italy. The fast solution is for a number of European tech firms to take large stakes in ARM such that the ownership is distributed across the EU and no one has majority control. For example, Siemens, Thales, Groupe Bull would be a decent core, plus some of the larger telecoms companies. Of course, in order to be successful, ARM would have to be allowed to operate commercially rather than politically, the killer of all state-owned enterprises, but it could be done.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
