Arm Co-Founder: Nvidia Sale Is Because Softbank Over-Invested In Firm (newstatesman.com) 65
Co-founder and ex-president of Arm Holdings Tudor Brown says that SoftBank's projected sale of the chip company to Nvidia is the result of a bungled business strategy that saw the fund throw too much money at Arm and prioritize the wrong business areas. NS Tech reports: "In my opinion, they put too much money into it, spent money on things that clearly -- in my opinion -- weren't going make money in the short term, and now, suddenly, they're saying, 'Oh, dear me, this company isn't performing very well,'" says Brown. He added that "[SoftBank] invested too heavily [...], threw too much money at it and haven't got a good return as a result."
Brown laments the potential purchase, saying that it would fundamentally clash with Arm's underlying business model. Because the firm designs technology that is sold or licensed to a great number of companies, Arm's business model requires it remaining on good terms with "an unholy clan of competitors," according to Brown. Brown says Nvidia or any other semiconductor company owning Arm is "immediately going to upset that balance and make it very, very difficult for other companies to feel that they have equal access to the technology." Brown says he can't imagine why a company would want to buy Arm if it wasn't seeking to give itself an unfair advantage such as "early access" or to "deprive the other guys from having whatever innovations were to take place." Another Arm co-founder, Hermann Hauser, told the BBC that the UK government should intervene to help the firm go public and remain an independent British company.
Brown laments the potential purchase, saying that it would fundamentally clash with Arm's underlying business model. Because the firm designs technology that is sold or licensed to a great number of companies, Arm's business model requires it remaining on good terms with "an unholy clan of competitors," according to Brown. Brown says Nvidia or any other semiconductor company owning Arm is "immediately going to upset that balance and make it very, very difficult for other companies to feel that they have equal access to the technology." Brown says he can't imagine why a company would want to buy Arm if it wasn't seeking to give itself an unfair advantage such as "early access" or to "deprive the other guys from having whatever innovations were to take place." Another Arm co-founder, Hermann Hauser, told the BBC that the UK government should intervene to help the firm go public and remain an independent British company.
It's a good thing. (Score:1)
Nvidia is the only chipmaker left who still demonstrates innovation and improvement. ARM will massively improve under Nvidia.
Re:It's a good thing. (Score:4, Insightful)
Care to elaborate? It seems to me we're currently experiencing the first big leap in consumer processors since Intel came out with the Core2Duo. AMD is doing great stuff and the Apple SoC stuff looks neat. I am not a hardware expert by any means, so I'm genuinely curious why you feel that way (I don't play video games or do anything GPU intensive so I may have overlooked whatever innovative things Nvidia has been doing).
Re: It's a good thing. (Score:5, Informative)
In fact the GeForce was the very first GPU. All of the others have basically just copied ideas from Nvidia.
Nope. Voodoo was 1994 [wikipedia.org], GeForce was 1999 [wikipedia.org].
Re: (Score:3)
And arcade systems started already in the 1970's with dedicated graphic logic.
Atari ANTIC and NEC PD7220A were among the first.
Not that they were very smart, but they did exist and can be seen as leading the way of intelligent graphics.
Re: It's a good thing. (Score:1)
Re: (Score:3)
The Sony Playstation coined the termed GPU 4 years before.
To mean "geometry", not graphics, but that's why nVidia jumped on it years after to say they had the first "geometry"PU, because their competitors weren't "geometry" but "graphics".
nVidia also bought up 3DFX in 2000, so it's unfair to say all their innovation came from in-house.
And they'd pretty much shot themselves in the foot with RTX which is a desperate grab back to the days when Quake mods could do the same, rather than anything forward-thinking
Re: (Score:2)
3Dfx didn't accelerate polygons in hardware (aka 'transform & lighting')
What nonsense. It certainly did accelerate polygons. It didn't have the programability of the pipeline like the later geforce architecture, but it sure did accelerate basic 3D stuff like triangles, gouraud shading and texture mapping.
Re: (Score:2)
Re: (Score:2)
Re: (Score:2)
Re: (Score:2)
What nonsense. It certainly did accelerate polygons.
It didn't accelerate polygon math- transforms and lighting.
It could rasterize polygons, but as far as computing what a scene looked like in after rotating/moving a camera? all done on the CPU.
Re: (Score:2)
Re: (Score:2)
Though I recall that at the time, you could actually get CPUs faster than the hardware T&L capabilities, so if you were rich sauce with a really nice CPU, you could get a Voodoo2 to outperform a Geforce.
Re: It's a good thing. (Score:2)
3dfx Interactive was a company headquartered in San Jose, California, founded in 1994, that specialized in the manufacturing of 3D graphics processing units, and later, graphics cards.
They didn't do 2d, at first, only 3d. As someone else pointed out Nvidia also bought them in 2000, for their intellectual property.
Re: (Score:2)
Too early. The Voodoo chipset was released in 1996. 3DFX as a company was founded in 1994.
To me, 1994 was the year Commodore went bankrupt, and in 1993, they were working on a GPU based on the PA-RISC processor, which was designed to use OpenGL right from the start. I can only imagine how different the world would have been if there were a true OpenGL-based GPU on the market in 1995. The Atari Jaguar also had a similar, fully programmable graphics chip, though it didn't use a graphics API, and its perfo
Re: (Score:2)
In fact the GeForce was the very first GPU.
Didn't you kind of forget Jim Clark's Geometry Engine (1981)?
Re: (Score:2)
Re: (Score:2)
In fact the GeForce was the very first GPU. All of the others have basically just copied ideas from Nvidia.
How baseless... Geforce wasn't the first GPU, it wasn't even nvidia's first GPU. See: https://en.wikipedia.org/wiki/... [wikipedia.org]
Moreover, the geforce architecture was built to support the DirectX HAL model. So it was largely the DirectX consortium created by microsoft that drove innovation. But nvidia was on that consortium, just like the other pc-market graphical chip manufacturers in that time (somewhere mid-90s i guess?).
Re: (Score:2)
The GeForce was the first GPU because nVidia coined the term to market that specific product. Graphics chips were not called a "GPU" before then. ATI at the time did not call their products GPUs, and tried to coin the competing term "VPU". This didn't catch on, and they eventually also adopted the GPU term.
Re: (Score:2)
The GeForce was the first GPU because nVidia coined the term to market that specific product. Graphics chips were not called a "GPU" before then.
"The term [GPU] was first used by Sony in 1994 with the launch of the PS1. That system had a 32-bit Sony GPU (designed by Toshiba)." — Jon Peddie, "Is it Time to Rename the GPU?" [computer.org].
Re: (Score:2)
The GeForce was the first GPU because nVidia coined the term
I guess you missed the Sony Playstation which came out with an embedded "Sony GPU" in 1994, 5 years before the Geforce did. https://www.computer.org/publi... [computer.org]
Seriously WTF is it in this thread with people misattributing half of computing history to NVIDIA. NVIDIA were innovative sure but they weren't the first to do anything. They may have been the best at it, but not the first.
Re: (Score:2)
Apple SoC stuff looks neat
Not if arm goes NVIDIA lol. Honestly though Apples SOC thing will be another iPhone moment I think.
Re: (Score:2)
Not if arm goes NVIDIA lol
Are you seriously betting against Apple? They are not _known_ to the general public as a chip developer, but they are shipping about the same number of processors as Intel today, and in a year or two will ship significantly more.
Re: (Score:2)
Apple has a lifetime architecture license, and their CPU designs are in-house from scratch. They don't care what happens to ARM because they're not impacted by it. If they can't get access to newer ARM instruction sets, no skin off their back, they can just do their own thing.
Re: (Score:2)
and will require closed source drivers (Score:2)
i think we can learn how nvidia will behave by looking at how it has behaved for its entire existence.
Re: (Score:1)
Nvidia takeover is bad for everyone. Including UK and obviously China. For China they must guarantee that they will not apply US sanctions or face losing their IP.
Re: (Score:2)
Paid commercial endorsement from a bot
Wework, Uber (Score:5, Interesting)
Re: Wework, Uber (Score:1)
Indeed. Even the super rich and super smart suffer from the Dunning-Kruger effect.
Re: (Score:1)
Obligatory XKCD [xkcd.com].
Re: (Score:3)
Arm is not like those two, in that it produces actual value, Softbank just paid too much and have no clue about the asset they now have as the article says.
In my view this a good thing, as the money Softbank gets from the Saudis will wind up in the hands of people who will do something useful with it,
Re: (Score:2)
Right now they're pumping petroleum out of the ground as fast as their infrastructure can, I rather suspect the royals want to drain as much out of the reserves as they can before the commoners toss them out on their collective asses. When that happens a big group of people are suddenly going to have to learn what a budget is.
Re: (Score:2)
The commoners of Saudi Arabia will be welcome to what's left presumably.
Re: (Score:1)
They were looking for next Jack Ma and found only Ali Baba and his 40 friends.
Re: (Score:2)
Investing in the wrong thing seems to be becoming a pattern over there at Softbank.
What's frustrating to me is that much of the financial media writing on this is about "mis-steps", whereas it's probably more likely that Softbank only made a few _good_ decisions.
Have to agree (Score:5, Interesting)
The company really needs to be independent. It is a mistake for Nvidia to buy it.
Re:Have to agree (Score:5, Funny)
I was previously modded down for saying the same thing.
Re: (Score:3, Informative)
The British government won't help ARM. Their policy is to asset strip the entire country.
The short sighted fools who voted for them have done irreparable damage to the country. We can never get those things back.
Re: (Score:2)
It has fuck all to do with the government. This was a private company, owned by private shareholders, who sold up to another private company. The government never got involved - which is possibly why it became so successful.
Re: (Score:2)
The British government won't help ARM. Their policy is to asset strip the entire country.
The short sighted fools who voted for them have done irreparable damage to the country. We can never get those things back.
So basically you're still butthurt that Brits voted for independence from the EU.
Re: (Score:2)
Yes
Re: (Score:3)
The company really needs to be independent. It is a mistake for Nvidia to buy it.
I work with ARM and many of its customers on a regular basis and I have to agree as well. ARM's neutrality is a crucial factor in its broad success. Almost every major and minor chipmaker in the world licenses its IP under roughly the same terms, enabling them all to compete aggressively with each other while still creating an ecosystem of devices using the same ISA which can run nearly the same software. This structure has dominated the world because it works very well. Giving one chipmaker an inside track
Re: (Score:2)
The company really needs to be independent. It is a mistake for Nvidia to buy it.
If the sale goes through, it's the beginning of the end for ARM. It'll take years, but the multitudes of smart phone vendors that depended on that vendor neutrality will start looking for a replacement. And Apple would be in a curious position as well, since their whole future map now depends on ARM.
This will accelerate RISC-V (Score:1)
Maybe SoftBank will now invest in a couple of risc-v startups.
RISC-V - Not likely: Legacy (Score:2)
And it won't work, for the same reason we got stuck with x86 for so many years in the desktop/server space (despite attempts to shake things up even with Intel's own Itanium):
to much binary legacy (at least backthen). Even switching from 32bits x86 to 64bits x86_64 took an eternity on Windows (as opposed to opensource ecosystems like Linux).
In the Win-PC world, to many users were relying on binary proprietary blobs that were exclusively made for the older architecture. They wouldn't want to rebuy a new vers
Re: RISC-V - Not likely: Legacy (Score:1)
despite attempts to shake things up even with Intel's own Itanium
What attempts?? Intel crippled Itanium.
Nothing to do with anything but blowing it on.... (Score:4, Insightful)
Nothing in that statement is true. They are selling ARM because it's their most valuable asset to try to recover some of the nearly $47 billion they blew on WeWork. Softbank is in serious shit after WeWork collapsed. SoftBank lost more than half it's assets in WeWork.
Re: (Score:2)
WTH is WeWork?
(I've never heard of it. So no wonder. And I don't need to know, but it's just funny that someone would spend that amount of money on something most people have never heard of).
Re:Nothing to do with anything but blowing it on.. (Score:5, Funny)
Re: (Score:2)
Yep. That sounds about right. :-)
Re: (Score:2)
Maybe if nVidia buys ARM, we'd see more support for nVidia products in MacOS again.
What makes you think that?
Re: Upside for Apple users? (Score:1)
How to stop terrorism? (Score:5, Insightful)
Take the funds from the sponsors of Wahabbism and burn it in a bunch of bad investements.
Someone give Masa a Nobel Peace prize.
His own fault (Score:3)
If he cares so much, he shouldn't have sold the company back when he had the chance. Instead he took the cash, undoubtedly enough that he doesn't need to bother with working for the rest of his life, and now other people own his company and decide how to dispose of it.
Re: (Score:2)
He's probably still got some shares. And I think you're right that he doesn't need to bother working (and he's certainly at retirement age), but most successful business people try to find new successes: he was certainly an "angel investor" twenty years ago, when he gave a lecture at my university to explain what that meant to a bunch of compsci students.
Re: (Score:2)
He's probably still got some shares.
Er, no. SoftBank bought out ARM Holdings for cash. They aren't listed any more.
Re: (Score:3)
If he cares so much, he shouldn't have sold the company back when he had the chance.
If you built a company, and you really really love it, and you feel it's worth $10 billion and someone offers you $32 billion for it, what do you do? You sell it.
Re: (Score:2)
I can only speak for myself, but if I had properties worth 10 billion that I truly cared about, I wouldn't sell. I wouldn't give a fuck about money anymore and I'd concentrate on doing cooler stuff. As long as everyone gets a paycheck, there's no need to focus on net profit.
That's the reason why companies want tons of billions of dollars, right? To do interesting and meaningful things? Meh. By the time you have $32 billion, you're officially a stage 4 company and the only reason why you exist is to kee
Softbank should just rename itself. (Score:2)
To "Overinvested".
It seems to be their modus operandi anyway.
Interesting situation for RISC V (Score:2)
If NVidia succeeds in buying ARM, RISC V will be really interesting to a lot of companies that depend on ARM processors.
There's a lesson here about the risks of depending on privately owned intellectual property.