Nvidia has approached Softbank and is considering buying ARM Holdings

Page 11 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

uzzi38

Platinum Member
Oct 16, 2019
2,565
5,574
146
Will update with articles as they come out (they haven't yet but news is public). For now just a couple of Tweets:



This is a real significant move for Nvidia, having a huge role in ARM core designs could have major implementations down the line. Most importantly, it secures them a platform and ecosystem down the line.

Oh and I guess Nvidia becomes the defacto standard for GPU IP for mobile instead of Mali. That too.

EDIT: Bloomberg article here: https://www.bloomberg.com/news/arti...ompany-arm-is-said-to-attract-nvidia-interest
 
Last edited:

moinmoin

Diamond Member
Jun 1, 2017
4,933
7,619
136
In other words, is ARM IP and owning its in-house engineer team worth $40B? Nvidia must have thought so, and I guess the price is easier to bear since Nvidia stock is pretty much at an all-time high.
It's not $40B, it's $35B of which only $12B is actual hard cash, paid in several instances (the first being only $2B). $23B are in stocks, $21.5B for SoftBank, and $1.5B for Arm employees. One could argues spreading stock ownership is not the worst thing to happen to a public company. The remaining $5B depend on how Arm fares, if it fares well it costs some more, but likely that additional cost is worth a shrug when it actually comes down to it and may be paid out in stocks as well.

What Nvidia gets is future proofing its business and being able to offer fully vertically integrated solutions.
 

amrnuke

Golden Member
Apr 24, 2019
1,181
1,772
136
It's not $40B, it's $35B of which only $12B is actual hard cash, paid in several instances (the first being only $2B). $23B are in stocks, $21.5B for SoftBank, and $1.5B for Arm employees. One could argues spreading stock ownership is not the worst thing to happen to a public company. The remaining $5B depend on how Arm fares, if it fares well it costs some more, but likely that additional cost is worth a shrug when it actually comes down to it and may be paid out in stocks as well.

What Nvidia gets is future proofing its business and being able to offer fully vertically integrated solutions.
This is a great deal for Nvidia and a great deal for SoftBank, IMO.

Have to imagine Nvidia is looking for a bigger chunk of exascale/compute and Arm IP is a great way to get there.
 

coercitiv

Diamond Member
Jan 24, 2014
6,151
11,685
136
Are you saying that you're in favor of opportunistic price gouging then? Why not have Nvidia jack up the licensing fee by 10x because it's all about making money and screwing the customers, right? Martin Shkreli would be proud.
Huh? I'm of the opinion that Nvidia's acquiring ARM can be toxic for the industry because there's a way through all of this in which Nvidia gets to have the cookie and eat it too: keep ARM as a successful business while neutering the competition.
 

Cogman

Lifer
Sep 19, 2000
10,277
125
106
Huh? I'm of the opinion that Nvidia's acquiring ARM can be toxic for the industry because there's a way through all of this in which Nvidia gets to have the cookie and eat it too: keep ARM as a successful business while neutering the competition.

IMO, the likely route for nVidia is the Microsoft route. Embrace, Extend, extinguish.

What we are likely to see, IMO, is nVidia adding a bunch of new ARM extensions, implementing them first, and then selling their SoC as "See, nVidia Tegra is cutting edge arm, you should all use it in the future!". They probably will still let others get those extensions through licensing, however, they'll be first to the floor.

I also see them doing similar things with Cuda to combat OpenCL/Vulkan stuff. They'll put cuda into the spec for all future ARM products so they can slowly squeeze out other hardware manufactures "See, write it in Cuda, use it across platforms!"

Their goal will be to push other hardware manufactures out of the business while trying to win over device manufactures.
 
  • Like
Reactions: Tlh97 and KompuKare

Roland00Address

Platinum Member
Dec 17, 2008
2,196
260
126
Qualcomm is already trying to create an 'ARM PC' market, so if Nvidia sold chips targeted for such a market they would be directly competing with one of their largest licensees who is hoping to expand into a new market.

If on the other hand Nvidia cares primarily about servers for the market they hope to enter, they wouldn't be stepping on the toes of any major licensees. There are a few tiny players with big hopes and dreams but not much scale to speak of, and Amazon who cares only about highly customizing chips for their own use so isn't a customer Nvidia could win anyway.
Nvidia is not going to out compete Qualcomm on quality or the cost of manufacturing the logic board, the chip, and so on with Qualcomm vs Nvidia ARM.

What might happen is they lower Qualcomm's profit margin on laptop sold and that may create bad blood.

Right now the problem with windows on arm is not the hardware but the software. Likewise the same for Arm on Server, you need everyone to recompile, make drivers, and so on.

(Server this is going to be so much easier than WoA for there is a chicken vs. egg scenario where good hardware will cause the software people to recompile which will attract more software and more arm which will create better hardware. WoA has a problem it is not just software but drivers and there is little incentive to fix drivers.)
 

Heartbreaker

Diamond Member
Apr 3, 2006
4,222
5,224
136
Qualcomm is already trying to create an 'ARM PC' market, so if Nvidia sold chips targeted for such a market they would be directly competing with one of their largest licensees who is hoping to expand into a new market.

I really can't see NVidia avoid a fat plumb like ARM PCs just because Qualcomm is dabbling at it. They compete strongly against their AIB GPU partners by selling full "Founders Edition" GPU Cards , while selling AIB partners chips.

If Windows on ARM takes off, NVidia is going to be in their with their own APUs. Why settle for 1-2% royalty when they can make 40%+ margins on Windows SoCs.

This was the one Weakness NVidia had against AMD and Intel in the Windows GPU market. AMD and Intel APU/iGPUs kill sales of low end discrete GPUs and over time those APUs are getting better, and killing more low end discrete GPUs.

With no capability to build an APU, there is nothing NVidia can really do about it.

But if ARM Windows starts to have success, then they tables start turning.
 

ThatBuzzkiller

Golden Member
Nov 14, 2014
1,120
260
136
I don't see how that could happen at this point. The reason China could stop that Qualcomm/NXP merger was because Qualcomm had most of its revenue coming from China. That's not true for Nvidia and there is a separate ARM China subsidiary (where ARM only owns 49%) which has pretty much already gone rogue at this point. That along with the U.S. bans for tech IP for China, Nvidia has to be prepared to just cut out China from their business and move on leaving China no power to stop the merger.

It is much more likely that U.K. regulators may step in but seeing as ARM was already owned by a Japanese company, I don't know if they actually will or not even though Nvidia could be seen as a competitive takeover compared to Softbank's acquisition just being a portfolio expansion. We'll find out soon enough.

I don't think you understand how damaging sanctions from China would be for Nvidia ...

A quarter of Nvidia's revenue comes from China alone which is still a growing market with much more potential in the future ...

Losing potentially as much as a third of the total revenue in the future would wreck their future expansion prospects. Nvidia would have to eat a massive stock devaluation making it harder for them to acquire ARM Ltd ...
 

Hitman928

Diamond Member
Apr 15, 2012
5,177
7,629
136
I don't think you understand how damaging sanctions from China would be for Nvidia ...

A quarter of Nvidia's revenue comes from China alone which is still a growing market with much more potential in the future ...

Losing potentially as much as a third of the total revenue in the future would wreck their future expansion prospects. Nvidia would have to eat a massive stock devaluation making it harder for them to acquire ARM Ltd ...

That's not accurate. Nvidia's regional revenue numbers aren't saying where the end product is sold but rather is based upon the headquarters location of the partner their chips are sold to. So the 25% (and Taiwan is even higher by the same accounting) comes from Nvidia selling to AIB partners based in China who then sell cards all over the world. If Nvidia had to exclude Chinese AIB partners there would be some short term pain but they could fill that hole medium and long term and have already started to shift away from being so reliant on Chinese partners as about 5 years ago that 25% number was over 50%.

PC gaming in China is big and Chinese customers do make up a lot of the PC gaming market, but hardware wise they tend to skew heavily to the budget options and software wise they tend to skew heavily to MOBA and MMO type games. I don't know how much Nvidia would lose from Chinese sanctions but it is not the same situation as Qualcomm and knowing the current relationship between the US and Chinese governments, if they aren't prepared for having to accept this as part of their takeover bid then they haven't done their due diligence. Either that or they are very confident in getting it approved.
 
  • Like
Reactions: Tlh97

chrisjames61

Senior member
Dec 31, 2013
721
446
136
From ars Tecnica- "
if Nvidia does end up ruining ARM, there aren't a lot of alternative options currently out there. The great hope would be in the RISC-V instruction set, an open source ARM alternative that doesn't require any licensing. Spinning up a quality ecosystem of hardware, compilers, operating systems, and software is such a daunting task, though, that most vendors were happy paying ARM licensing fees.

Maybe Nvidia's purchase will be the push companies needed, though. The earlier Reuters report we mentioned already claims there is a "backlash" forming against Nvidia, and it says that "a source at one US company using Arm designs said the move would likely accelerate an industry shift already under way from Arm designs to RISC-V."

 

SarahKerrigan

Senior member
Oct 12, 2014
339
468
136
Ah, interesting. I'm only familiar with the HPC side of things- thanks! I was under the impression that they weren't making a successor to the SPARC64 XII, and that's a 20nm part from 2017- felt like the end of the line for SPARC, given the ARM transition in their HPC line.

It's getting some kind of rev next year (may just be system, not silicon), but yes, you're right that it's on its way out. But it's at least still an active product line, unlike the XIfx HPC component, which is deader than disco.

This leaves GS21 mainframes and A64fx long term, which are closely related due to common ancestry from the SPARC64 V family.
 
  • Like
Reactions: NTMBK

ThatBuzzkiller

Golden Member
Nov 14, 2014
1,120
260
136
That's not accurate. Nvidia's regional revenue numbers aren't saying where the end product is sold but rather is based upon the headquarters location of the partner their chips are sold to. So the 25% (and Taiwan is even higher by the same accounting) comes from Nvidia selling to AIB partners based in China who then sell cards all over the world. If Nvidia had to exclude Chinese AIB partners there would be some short term pain but they could fill that hole medium and long term and have already started to shift away from being so reliant on Chinese partners as about 5 years ago that 25% number was over 50%.

PC gaming in China is big and Chinese customers do make up a lot of the PC gaming market, but hardware wise they tend to skew heavily to the budget options and software wise they tend to skew heavily to MOBA and MMO type games. I don't know how much Nvidia would lose from Chinese sanctions but it is not the same situation as Qualcomm and knowing the current relationship between the US and Chinese governments, if they aren't prepared for having to accept this as part of their takeover bid then they haven't done their due diligence. Either that or they are very confident in getting it approved.

Considering how seriously Nvidia was willing to take the CCP's blessing with their acquisition of Mellanox, I'm willing to bet that China is easily their second biggest market going by their unreleased telemetry data that they collect from their drivers. It's obvious that the CCP's words holds lot's of value for Nvidia otherwise they would've went ahead with the acquisition instead of waiting for the ruling ...

And Qualcomm is in a similar situation since they supply their SoCs to Chinese ODMs who then integrates it into the handsets to sell them at the international market so I highly doubt that it is solely Chinese consumers driving most of Qualcomm's value ...
 

Hitman928

Diamond Member
Apr 15, 2012
5,177
7,629
136
Considering how seriously Nvidia was willing to take the CCP's blessing with their acquisition of Mellanox, I'm willing to bet that China is easily their second biggest market going by their unreleased telemetry data that they collect from their drivers. It's obvious that the CCP's words holds lot's of value for Nvidia otherwise they would've went ahead with the acquisition instead of waiting for the ruling ...

And Qualcomm is in a similar situation since they supply their SoCs to Chinese ODMs who then integrates it into the handsets to sell them at the international market so I highly doubt that it is solely Chinese consumers driving most of Qualcomm's value ...

For sure, Nvidia doesn't want to go against China and would much prefer to keep that bridge intact. Things are getting pretty crazy though and I just think Nvidia needs (and is prepared for) the crap really hitting the fan here. Who knows, in 2 months the U.S. could have a complete administration swap and all of this could be moot.
 

Hitman928

Diamond Member
Apr 15, 2012
5,177
7,629
136
Will it ? Even if it does will they lift the sanctions ? This will be critical to improving the CCP's mood ...

Obviously I can't say what will happen in the election and without wanting to get too into politics, my personal opinion is that if the current administration is replaced, then the incoming administration will have a very different attitude and I would expect very large policy changes in regards to China - US relations and they would be seen as much more favorable by the CCP.
 

CluelessOne

Member
Jun 19, 2015
76
49
91
From ars Tecnica- "
if Nvidia does end up ruining ARM, there aren't a lot of alternative options currently out there. The great hope would be in the RISC-V instruction set, an open source ARM alternative that doesn't require any licensing. Spinning up a quality ecosystem of hardware, compilers, operating systems, and software is such a daunting task, though, that most vendors were happy paying ARM licensing fees.

Maybe Nvidia's purchase will be the push companies needed, though. The earlier Reuters report we mentioned already claims there is a "backlash" forming against Nvidia, and it says that "a source at one US company using Arm designs said the move would likely accelerate an industry shift already under way from Arm designs to RISC-V."


Open Source is not always an answer to all bad things. Especially on an ISA that will be used widely. What is the licensing term? Is there something like GPL v3 or GPL v2 where all proposed extensions must be mainlined?
Observe OpenGL where extensions were not mandated to be supported by all users. See how fragmented the support for extensions were?
If similar things happen to RISC V ISA, all it will do is create multiple walled garden where each manufacturer has their own sets of extensions. Don't get me wrong, manufacturers will love that. Imagine them saying supported on WD silicon only, NXP silicon only or Renesas exclusive. But consumers and developers will be in trouble for support. Ultimately it will create a bad taste for the consumers driving away from the products and give RISC V a bad name.
 

Hitman928

Diamond Member
Apr 15, 2012
5,177
7,629
136
Incoming administration would be in January if the administration changes.

Yes, actual change wouldn't happen until Jan. but once the new president is elected the sitting president becomes a "lame duck" and foreign affairs essentially get put on pause, especially when the new president is expected to have a very different foreign policy.
 

A///

Diamond Member
Feb 24, 2017
4,352
3,154
136
Yes, actual change wouldn't happen until Jan. but once the new president is elected the sitting president becomes a "lame duck" and foreign affairs essentially get put on pause, especially when the new president is expected to have a very different foreign policy.
This is true. But given what's happened, I can't imagine they wouldn't mess things up further.
 

ThatBuzzkiller

Golden Member
Nov 14, 2014
1,120
260
136
Obviously I can't say what will happen in the election and without wanting to get too into politics, my personal opinion is that if the current administration is replaced, then the incoming administration will have a very different attitude and I would expect very large policy changes in regards to China - US relations and they would be seen as much more favorable by the CCP.

Even if there is a new administration, is there a promise that they'll lift sanctions against China and it's domestic firms ? That's the only policy that matters. Anything less will be viewed as belligerence to them ...

If not I'd imagine that the CCP would very much prefer to kill two birds (Intel & Nvidia) with one stone (via anti-trust regulators) at this point. Intel is pretty much finished and their manufacturing days are numbered so when the top party leadership heard the news they were very ecstatic that Intel was failing. Now that Intel is totally out of the picture they don't want another American megacorporation like Nvidia filling the vacuum either so nixing their expansion plans would be an effective way to sabotage the American semiconductor industry altogether ...
 

DrMrLordX

Lifer
Apr 27, 2000
21,582
10,785
136
. If Nvidia had to exclude Chinese AIB partners there would be some short term pain but they could fill that hole medium and long term and have already started to shift away from being so reliant on Chinese partners as about 5 years ago that 25% number was over 50%.

Several AIBs have been moving actual manufacturing to other countries like Vietnam. My ASRock x370 Taichi was produced there. It's not trivial to pack up and move an entire operation (or create a subsidiary) to a place like 'Nam such that the long arm of the PRC can't effectively embargo it, but it can be done.

If similar things happen to RISC V ISA, all it will do is create multiple walled garden where each manufacturer has their own sets of extensions. Don't get me wrong, manufacturers will love that. Imagine them saying supported on WD silicon only, NXP silicon only or Renesas exclusive. But consumers and developers will be in trouble for support. Ultimately it will create a bad taste for the consumers driving away from the products and give RISC V a bad name.

Too right. If RISC V wants to be taken seriously as an alternative to ARM, the first thing they have to do is roll the ISA extensions together into tiers of support based on the target market (embedded microcontrollers, mobile SoCs, desktop SoCs, workstation, server). At least that way, binary interoperability will be increased between different vendors of RISC V SoCs within the same market. And if it's really going to be an open standard, there needs to be a way to encourage/force manufacturers to open up proprietary ISA extensions.
 

teejee

Senior member
Jul 4, 2013
361
199
116
Several AIBs have been moving actual manufacturing to other countries like Vietnam. My ASRock x370 Taichi was produced there. It's not trivial to pack up and move an entire operation (or create a subsidiary) to a place like 'Nam such that the long arm of the PRC can't effectively embargo it, but it can be done.



Too right. If RISC V wants to be taken seriously as an alternative to ARM, the first thing they have to do is roll the ISA extensions together into tiers of support based on the target market (embedded microcontrollers, mobile SoCs, desktop SoCs, workstation, server). At least that way, binary interoperability will be increased between different vendors of RISC V SoCs within the same market. And if it's really going to be an open standard, there needs to be a way to encourage/force manufacturers to open up proprietary ISA extensions.

No, I don't agree at all.
This is a made up issue.

First binary compatibility is not a thing in embedded/iot etc at all even today. Firmware are always compiled directly for specific targets.

If we talk about future high end socs for smart phones, PCs, server etc, then de facto standards for what official extensions are required will be established by the market themselves. For example if someone plans to do a Risc-V smart phone soc, then they will of course collaborate with an owner of a ecosystem/OS (e g Google/Android), otherwise their product won't have a market.
And there are very few reasonable Risc-V extension combinations for the high-end at a given time anyway.
 
  • Like
Reactions: Vattila

DrMrLordX

Lifer
Apr 27, 2000
21,582
10,785
136
First binary compatibility is not a thing in embedded/iot etc at all even today. Firmware are always compiled directly for specific targets.

That's the only market currently embracing RISC V. They know exactly what software will run on those SoCs, and they can churn out an SoC that has only bare-minimum support for that software/firmware. Hence the appeal of RISC-V.

If we talk about future high end socs for smart phones, PCs, server etc, then de facto standards for what official extensions are required will be established by the market themselves. For example if someone plans to do a Risc-V smart phone soc, then they will of course collaborate with an owner of a ecosystem/OS (e g Google/Android), otherwise their product won't have a market.
And there are very few reasonable Risc-V extension combinations for the high-end at a given time anyway.

Let's all take a trip down memory lane by remembering about an Intel SoC called Baytrail. Baytrail had several variants, and Intel practically paid people to use these things (contrarevenue, for the tablet variant specifically). Because Baytrail was cheap-as-in-free, enterprising low-end market OEMs began using the tablet-oriented Baytrail in desktop applications, which wasn't "supposed" to happen. Baytrail-T found its way into set top boxes and mini PCs. Fortunately for people who bought those . . . things, Baytrail-T supported all the same instructions (I think) as other Baytrail products, so it's not like software that worked on one Baytrail wouldn't work on another. Baytrail-T products just ran things very slowly when used in hardware applications that were inappropriate for the SoCs. Not that Baytrail-D was that much better, but oh well.

Now imagine if Baytrail-T had been missing some key ISA features of standard Intel x86 CPUs of the day (and I'm not talking cutting-edge-for-its-time SIMD like AVX). Like . . . I don't know, SSE3 or SSE4. There are probably a fair number of x86 programs out there that won't even boot without SSE4 today, and there were probably some in Baytrail's time that required at least SSE2 or SSE3. A dodgy PC vendor could sell you a Baytrail-T for a song, preloaded with an OS that would boot and some software carefully selected to work without needing whatever ISA extension was missing from this hypothetical Baytrail (let's call it Bayfail). Try loading some third-party x86 software on there and whabam, now you can't run it because . . . SSE2/3 unsupported? What's that? Consumer is confused. Not like the OEM would care. Of course this hypothetical Bayfail would only be intended for, I don't know, an embedded application or a closed-OS consumer device that will never run third-party software anyway (without some hackery). If Intel couldn't stop real-life Baytrail-T from finding itself in applications for which it was never intended, do you think the RISC-V world will fare any better?

Show me a RISC-V vendor that tries to sell a cheap phone SoC that doesn't support RISC-V P, and I'll show you a chincey OEM that will try to slip that SoC into phones that can access an app store with apps that (sometimes) use packed SIMD (presumably as an equivalent to 128b NEON). Someone buys this mysteriously cheap phone, only to discover that now they can't load a whole raft of applications that are supposed to be RISC-V compatible.

ARM dodges this neatly by setting standards for its own designs that are targeted to a specific market. For consumer devices, you can either have Cortex-M for embedded microcontrollers, or you get an A-series chip. And . . . that's about it! As long as the OEM chooses an A-series SoC, all the same instructions will be supported no matter what across that generation. No exceptions. There are some weird circumstances where you have ARM with custom ISA extensions (notably Apple), but that's for a closed software environment. You run exactly one OS from one vendor with apps that follow that vendor's design guidelines (or else). Bottom line is, though, that any application that will run on A76 or A77 will also run on A53. It may be slow on A53, but it'll run. It'll even support NEON.

RISC-V doesn't even try to set standards like that. It really is aimed at embedded where maybe the OEM doesn't want a full-fledged Cortex-M, opting instead for something simpler and cheaper that only needs support for a fraction of the functions offered by ARM's cheapest offering. Also, at least the last time I checked, not all of the RISC-V ISA extensions are "frozen", which means they are subject to change. That could be a mess. And then you have the possibility of propietary ISA extensions . . . ugh.