FTC Sues To Stop Blockbuster Chip Deal Between Nvidia and Arm (nytimes.com) 46
The Federal Trade Commission on Thursday sued to block Nvidia's $40 billion acquisition of a fellow chip company, Arm, halting one of the biggest semiconductor industry deals in history. From a report: The F.T.C. said the deal between Nvidia, which is based in California and makes chips, and Arm, a British company that designs chips, would stifle competition and harm consumers. The proposed deal would give Nvidia control over computing technology and designs that rival firms rely on to develop competing chips. "Tomorrow's technologies depend on preserving today's competitive, cutting-edge chip markets," said Holly Vedova, the director of the F.T.C.'s competition bureau. "This proposed deal would distort Arm's incentives in chip markets and allow the combined firm to unfairly undermine Nvidia's rivals." The companies announced the merger in September 2020 and said the merger would position the companies as leaders in semiconductors for artificial intelligence.
It begins (Score:4, Insightful)
This is just the first move. UK regulator seems to be going in the direction of blocking the deal as well, and there are similar problems with competition bureaucracy in other nations. Pretty much all major regulators in nations relevant to this deal seem to be semi-publicly talking about how this deal is unlikely to be approved. This is just the first genuinely open shot at it. It's very likely others will follow.
Looks like Nvidia was far too optimistic in its estimated ability to get the deal through. It's going to be interesting to see how SoftBank goes about this as well, because they specifically needed to offload ARM to offset losses they incurred to the rest of their investment portfolio before the deal to sell ARM to Nvidia was struck.
Re: (Score:3)
Not too optimistic, rather, they probably haven't had the campaign contributions that Intel and AMD make. There's actually a lot of competition out there, and RiscV designs will mature, and someone else will mature, and this happens for decades and decades.
Yes, ARM is in the vast majority of phones, but this isn't like there's a royalty path, just a licensing path. Will it disrupt other companies? Good.
NVIDIA isn't very good at politics and making empty promises to politicians. That's largely what this is a
Re: It begins (Score:2)
Re: It begins (Score:2)
Re: (Score:3)
ARM is huge in IoT, etc.
Intel and AMD can't go there because of their leaden ancient designs. Tough.
RiscV the dregs of MIPS, and plentiful others are waiting for the sheer oxidation of Intel and AMD. Custom fab is the current progression. SoC, hardening, multi-radio management, there's so much more to semiconductors than the ancient hallowed ways.
But this isn't about ARM Holdings, this is about politicians wanting new fabs and tech HQ in Cornwall, or perhaps Manchester, or to revive some office park in Dulu
Re: (Score:2)
Re: (Score:2)
ARM is huge in IoT, etc.
Intel and AMD can't go there because of their leaden ancient designs. Tough.
Intel abandoned that market not long after their 486; I have one of their embedded 486 chips somewhere. AMD never did and had some presence until recently with their Geode chips which they inherited from National Semiconductor who inherited from Cyrix.
Re: (Score:2)
Uh, no.
Intel tried the 896, the Atom (in many versions) and others to try to get to something that smells like IoT. They are out of tricks. The hat is empty. They're not doomed, simply, not Andy Grove's Intel.
Re: (Score:2)
Yup, for every ARM chip in a phone there might be 10 used in higher volume markets. Of course, smartphone makers can make up the difference by demanding that customers upgrade to new hardware every year whereas many other ARM applications are designed to last for many years.
Re: (Score:3)
Yes, ARM is in the vast majority of phones
And tablets, and smartwatches, and TVs, and embedded... and ARM has a non-trivial and increasing share in the data center, and a growing share in laptops (Chromebooks and now Macbooks). ARM is much, much bigger than phones.
There's actually a lot of competition out there, and RiscV designs will mature, and someone else will mature, and this happens for decades and decades.
Eventually, sure, but for the next 5-10 years, ARM is an important competitor to Intel and AMD in a lot of areas, and RISC V isn't. (I'm hoping to see RISC V chips grow -- I'm especially excited about OpenTitan, along with lots of other potential open source hardware -- but it's going to
Re: (Score:2)
I would much rather see NVIDIA take the lead than Softbank.
That Google can't control NVIDIA because of null sets from a long list of Venn diagrams is Google's problem. They can cow Softbank. Yeah, I understand Google has a lot invested into ARM, but welcome to open source.
RiscV is going to be awesome. But every Xeon and Ryzen suffers from insane architectural problem, brushed under the rug, that researchers monthly find new problems with. ARM fuels lots of stuff, let's agree on that. Give it to someone resp
Re: (Score:2)
Nvidia does not have an interest in "taking the lead". As far as they are concerned ARM is a competitor, buying it out is a way to remove it from the market.
Re: (Score:2)
Nvidia does not have an interest in "taking the lead". As far as they are concerned ARM is a competitor, buying it out is a way to remove it from the market.
How is ARM a competitor to Nvidia when Nvidia uses ARM in their Tegras and upcoming Grace? In fact, x86 is the competitor, and Nvidia has been scratching for a way to compete against Intel for a long time.
Re: (Score:2)
That Google can't control NVIDIA because of null sets from a long list of Venn diagrams is Google's problem. They can cow Softbank.
You completely misunderstood my comment. Though I can't speak for Google as a company, I don't see that Google has any interest in controlling either nVidia or Softbank, nor do I see that Google has any ability to control either.
Re: (Score:2)
I think I did indeed understand it. And yes, Whatever Happens To ARM is of enormous interest to Google. They dominate the world's client operating systems, oh yeah, on ARM.
Re: (Score:2)
Am sure Nvidia will have to pay a break up fee, probably in the billion or billions of dollars range.
After that Softbank can always spin it off again, with an IPO, if they really need the money.
Re: (Score:2)
I just don't think that this is going to be enough to cover the needs of SoftBank considering the hammering it took to much of its portfolio in last five years or so. SoftBank really needed a buyer for ARM to cover those losses.
And spinning it off is actually hard because of what ARM does. There aren't that many potential buyers that would hit the same snag of "conflict of interest" that actually can use it in any meaningful way. As for IPO, that's a whole different can of worms, since ARM is actually a hol
Re: (Score:2)
Well, it was a listed company (was listed in the LSE), and it can always be relisted again. May just take some time.
Intel/AMD (Score:2)
It may be anticompetitive on the embedded SoC front but if they are serious about applying their GPU/ML tech to server chips, it could be very competitive against amd64.
And, yeah, I expect their standards compliance and open source support to suck because they're nVidia but it's not like AMD has lived up to their promises whether on Radeon, Zen, or ML. Always a year late and a dollar short, which is mystifying given the RoI potential.
Re: (Score:2)
They could simply get an ARM license if that's what they wanted it for, and they still can after the merger falls apart.
The price they were going to pay only made sense if they were going to use control of ARM in an anti-competitive way.
Re: (Score:2)
They could simply get an ARM license if that's what they wanted it for, and they still can after the merger falls apart.
The price they were going to pay only made sense if they were going to use control of ARM in an anti-competitive way.
Exactly.
Default Regulator Rule (Score:1)
Once your business gets large enough to dominate it's industry, then any merger plans involving that business earn increased regulator scrutiny.
After all ... SIZE MATTERS.
Ok.... (Score:2)
Re: (Score:3)
That means, Intel, AMD, Qualcomm, Apple, Samsung are also certainly out of the question for taking over ARM.
So that might lead to Microsoft being able to take over as they aren't producing any (own) chips.
I would feel it would be best owned by a holding company or a consortium. Softbank is a good owner for now, but I'd be curious to see what happens.
Re: (Score:3)
Why should ARM be owned?
Why can't ARM be an independent company?
Re: (Score:1)
Re: (Score:2)
Are independent companies owned?
Oh, right, not in China.
But in countries with private property, are companies owned?
Re: (Score:2)
Are independent companies owned?
Independent companies are owned by shareholders, but not by other companies.
ARM cores are used everywhere. There is a conflict of interest for almost any potential parent company.
Re: (Score:2)
So your belief is that companies don't own shares in other companies?
Of course it is a conflict of interest, but that has nothing to do with your gobbledygook.
Re: (Score:2)
That means, Intel, AMD, Qualcomm, Apple, Samsung are also certainly out of the question for taking over ARM.
So that might lead to Microsoft being able to take over as they aren't producing any (own) chips.
I would feel it would be best owned by a holding company or a consortium. Softbank is a good owner for now, but I'd be curious to see what happens.
Softbank reportedly revealed that they offered ARM to Apple first, before announcing their desire to sell it off. Apple declined. They likely figured their Perpetual ARM "Architecture"-Class License would protect them sufficiently to not be worth all the likely Regulatory and/or Judicial Pushback (like what nVidia is now enjoying).
But Apple also reportedly started to investigate RISC-V about the same time as the nVidia/ARM deal was being first talked about; so. . . ?
But I agree, ARM wants to be Free (as in
Nice windfall for ARM. (Score:2)
A free 1.25 billion buckaroos.
Is this 1990? (Score:1)
I thought Blockbuster was out of business.
Re: (Score:2)
I thought Blockbuster was out of business.
That's the point, Blockbuster rentals is so irrelevant that we can use the word again.
Unfortunately there's no one stop companies from stealing words we use a lot like META.
FB wasn't content to steal just one word, those motherfuckers had to steal a whole Modifier!
I think it's fair for them to buy ARM (Score:2)
YES!! (Score:2)
Finally, senses over money!
Thank god... (Score:2)
Jensen Huang doesn't know how to share, so I'm very happy to hear this. The Mellanox acquisition was check due to their stranglehold on HPC interconnects, and the ARM acquisition would have been checkmate as that's the magic sauce that would have let them corner the emerging compostable infrastructure market. Thankfully AMD was able to acquire Xilinx, that gives them RoCE v2 chipsets, FPGAs, PCIe switches, and embedded processors. I do hope AMD's ROCm takes off, as that's the only open source platform we ha