Nvidia to make ARM processors for Windows machines

[picpicmac]

Newswire story tonight:

Nvidia to make Arm-based PC chips in major new challenge to Intel​

Nvidia has quietly begun designing central processing units (CPUs) that would run Microsoft’s (MSFT.O) Windows operating system and use technology from Arm Holdings(O9Ty.F), , two people familiar with the matter told Reuters.

The AI chip giant's new pursuit is part of Microsoft's effort to help chip companies build Arm-based processors for Windows PCs. Microsoft's plans take aim at Apple, which has nearly doubled its market share in the three years since releasing its own Arm-based chips in-house for its Mac computers, according to preliminary third-quarter data from research firm IDC.


Advanced Micro Devices (AMD.O) also plans to make chips for PCs with Arm technology, according to two people familiar with the matter.

...

Interesting.

If Microsoft and Nvidia (and maybe AMD) can move most of the PC market off of Intel architectures and onto ARM, I wonder if the performance increases will provide hard competition for Apple.

Also, will future "Hackintosh" projects use Nvidia ARM processors????

 

[picpicmac]

Newswire story tonight:

Nvidia to make Arm-based PC chips in major new challenge to Intel​

Interesting.

If Microsoft and Nvidia (and maybe AMD) can move most of the PC market off of Intel architectures and onto ARM, I wonder if the performance increases will provide hard competition for Apple.

Also, will future "Hackintosh" projects use Nvidia ARM processors????

The Reuters wire story writer clearly is not very deep in their knowledge of Nvidia, by calling it an "AI" chip maker. Sure, Nvidia does that, but its bread and butter are graphics chips and graphics cards.

 

[tis100]

This article seems to be written to boost Nvidia's stock. Nvidia has had a line of ARM APUs for a while now called Tegra. They even run Windows. See the Microsoft Surface 2. And more recently, Microsoft switched to an ARM processor co-developed with Qualcomm. See the Microsoft Surface Pro 9.

I suspect this is simply a product update and not a new product line. Perhaps it will used in the upcoming Nintendo Switch 2 to come out next year.

 

[diamond.g]

The Reuters wire story writer clearly is not very deep in their knowledge of Nvidia, by calling it an "AI" chip maker. Sure, Nvidia does that, but its bread and butter are graphics chips and graphics cards.

Nah, their bread and butter is datacenter. Gaming (the graphics cards) could go away and they would still be making pretty good money.

Not sure how easily Grace Hopper can be scaled down for "general desktop" use though.

 

[Spaceboi Scaphandre]

Newswire story tonight:

Nvidia to make Arm-based PC chips in major new challenge to Intel​

Interesting.

If Microsoft and Nvidia (and maybe AMD) can move most of the PC market off of Intel architectures and onto ARM, I wonder if the performance increases will provide hard competition for Apple.

Also, will future "Hackintosh" projects use Nvidia ARM processors????

No it won't lmao

But Nvidia making new ARM chips is important for another device: The Nintendo Switch. The Switch currently uses a Tegra X1, a modified version of the Tegra that was used in the shortlived Nvidia Shield. But that SoC is ancient and outdated. The new chip for the Switch 2 has been rumored to be a brand new SoC Nvidia made just for the Switch 2 that has similar specs to a gaming PC.

 

[MRMSFC]

Interesting to say the least. I swore they said something to the same effect when the M1 released though.

No doubt NVidia has serious gpu chops, but work on the cpu side remains to be seen. (Doesn’t the tegra use standard ARM cores?)

I do think with Windows 11 ARM compatibility and Microsofts dev kit, the x86 manufacturers have been put on notice.

Still, any new players in the market are going to have to iterate fast to catch up. They don’t control the whole stack like Apple and can’t necessarily guarantee everything works.

 

[Joe Dohn]

This article seems to be written to boost Nvidia's stock. Nvidia has had a line of ARM APUs for a while now called Tegra. They even run Windows. See the Microsoft Surface 2. And more recently, Microsoft switched to an ARM processor co-developed with Qualcomm. See the Microsoft Surface Pro 9.

I suspect this is simply a product update and not a new product line. Perhaps it will used in the upcoming Nintendo Switch 2 to come out next year.

I wouldn't be so sure that this is simply a product update. ARM chips are starting to take off as viable desktop alternatives.

And Microsoft really IS better potitioned than Apple here. If they do get the edge on ARM processors, it means their power efficiency will be just as good as the Apple MX product line, BUT they can leverage from all Windows legacy software. Their x86-x64 ARM translation layer is not perfect, but it is good enough that the main applications work.

And to top it off, Microsoft is much more flexible than Apple in several conditions. Apple wants to dictate too much on how users should use their computers, even in cases where it would be perfectly valid.

 

[Falhófnir]

With their GPUs sucking back so much power now, being able to have a CPU that draws maybe 15W rather than 60W under load frees up a huge amount of headroom for gaming laptops. That's a big prize up for grabs, potentially.

 

[targetnovember]

Since Nvidia is a public company, we can look at real numbers. For 2023, the page I found had their graphics revenue at 11.9 billion. Compute and networking (I'm assuming datacenter products) is at 15.1 billion. It's pretty evenly split. So is Nvidia an AI company? I guess so?

 

[diamond.g]

Since Nvidia is a public company, we can look at real numbers. For 2023, the page I found had their graphics revenue at 11.9 billion. Compute and networking (I'm assuming datacenter products) is at 15.1 billion. It's pretty evenly split. So is Nvidia an AI company? I guess so?

Is that for FY23? Cause the numbers look wildly different for FY24 (so far) unless I am reading page 23&24 of their 10-Q wrong.

 

[targetnovember]

Is that for FY23? Cause the numbers look wildly different for FY24 (so far) unless I am reading page 23&24 of their 10-Q wrong.

Yes, the page I found had that for 2023

 

[Analog Kid]

I'm not sure why Qualcomm's announcement was front page but this was not. Two PC veterans seriously eyeing Arm as a contender.

Intel needs to find a strategy fast.

 

[Falhófnir]

I'm not sure why Qualcomm's announcement was front page but this was not. Two PC veterans seriously eyeing Arm as a contender.

Intel needs to find a strategy fast.

Honestly wouldn't surprise me if they already have their own Arm chips under development behind securely closed doors. They will probably only be deployed in a last ditch can't beat em = join em scenario, though we are maybe edging ever nearer to that the more news we hear like this.

 

[jz0309]

I'm not sure why Qualcomm's announcement was front page but this was not. Two PC veterans seriously eyeing Arm as a contender.

Intel needs to find a strategy fast.

maybe it has to do with Qualcomm positioning themselves as Apple competitor - which they aren't.

As for Intel, the stories I saw recently had to do with them getting into the AI biz ... they've been losing marketshare in both laptop/desktop and data centers and I think the thought is with node advancements they'll regain marketshare.

But, since they also followed the strategy of getting in to all the taxpayer money for new fabs, they HAVE to win some big foundry customers very soon, their own chips will never fill up those fabs ...

 

[Ursadorable]

Also, will future "Hackintosh" projects use Nvidia ARM processors????

Doesn't Apple Silicon have the T2 on the M chip now? Pretty sure that will come into play, if it isn't already, to prevent future hackintoshs.

 

[Analog Kid]

Honestly wouldn't surprise me if they already have their own Arm chips under development behind securely closed doors. They will probably only be deployed in a last ditch can't beat em = join em scenario, though we are maybe edging ever nearer to that the more news we hear like this.

Maybe, but Intel does not have a great history in developing new architectures... Not due to lack of talent, but because of never really being able to commit to leaving the past. AMD, on the other hand, strikes me as much less fanatical and more flexible in their thinking.

Intel might have something brewing, but this wouldn't be the first time they've tried and would be the first time they succeeded.

maybe it has to do with Qualcomm positioning themselves as Apple competitor - which they aren't.

As for Intel, the stories I saw recently had to do with them getting into the AI biz ... they've been losing marketshare in both laptop/desktop and data centers and I think the thought is with node advancements they'll regain marketshare.

But, since they also followed the strategy of getting in to all the taxpayer money for new fabs, they HAVE to win some big foundry customers very soon, their own chips will never fill up those fabs ...

If Intel sees their future in AI accelerators, they're almost doomed from the start. Nvidia is an AI juggernaut because an AI accelerator looks a lot on silicon like a post-modern graphics accelerator and they're a decade or more ahead on the software support for the stack. The AI industry just thinks Nvidia when they're looking for compute and that will be a hard habit to break. I'd love to see someone challenge Nvidia for AI dominance, for the same reasons I'd like to see someone challenge TSMC, but Intel is starting from well behind and trying to fight their way forward.

And it would be nice to have another good foundry in the world. It all seems a bit precarious with TSMC having such a huge technological lead-- if they stumble (or get tripped) then it feels like the pace of technological progress will slow with them. Samsung and Intel are the two most likely backup plans, but Intel has been a bit of a mess in their fabrication business as well and they have a reputation for being rather mercurial with their business partners and interests that they're going to have to overcome for companies to bet their future on an Intel partnership. Last I saw, Qualcomm and Tesla both passed them over as a supplier.

I'm not an analyst, or even terribly well informed on these companies businesses, I'm mostly just someone with an opinion, but it seems to me like the window for Intel to reinvent themselves and remain a dominant force in the industry is closing fast. I think the x86-S spec is interesting, but it might be 5 years too late...

 

[vanc]

The Switch currently uses a Tegra X1, a modified version of the Tegra that was used in the shortlived Nvidia Shield

I wouldn't call it shortlived. It's still alive and kicking. I have two older models (2015) at home, and they're still working very well, even with the now ancient Tegra X1.

NVIDIA has been in the ARM CPU market for over a decade. They also provide CPUs for the auto industry, at least for Audi and Tesla as far as I know.

 

[picpicmac]

It's pretty evenly split. So is Nvidia an AI company? I guess so?

Yes, I guess it is. I hadn't check recently what was going on at Nvidia, so sometime in the past couple of years it looks like their corporate balance has shifted to the data center.

 

[jz0309]

Maybe, but Intel does not have a great history in developing new architectures... Not due to lack of talent, but because of never really being able to commit to leaving the past. AMD, on the other hand, strikes me as much less fanatical and more flexible in their thinking.

Intel might have something brewing, but this wouldn't be the first time they've tried and would be the first time they succeeded.



If Intel sees their future in AI accelerators, they're almost doomed from the start. Nvidia is an AI juggernaut because an AI accelerator looks a lot on silicon like a post-modern graphics accelerator and they're a decade or more ahead on the software support for the stack. The AI industry just thinks Nvidia when they're looking for compute and that will be a hard habit to break. I'd love to see someone challenge Nvidia for AI dominance, for the same reasons I'd like to see someone challenge TSMC, but Intel is starting from well behind and trying to fight their way forward.

And it would be nice to have another good foundry in the world. It all seems a bit precarious with TSMC having such a huge technological lead-- if they stumble (or get tripped) then it feels like the pace of technological progress will slow with them. Samsung and Intel are the two most likely backup plans, but Intel has been a bit of a mess in their fabrication business as well and they have a reputation for being rather mercurial with their business partners and interests that they're going to have to overcome for companies to bet their future on an Intel partnership. Last I saw, Qualcomm and Tesla both passed them over as a supplier.

I'm not an analyst, or even terribly well informed on these companies businesses, I'm mostly just someone with an opinion, but it seems to me like the window for Intel to reinvent themselves and remain a dominant force in the industry is closing fast. I think the x86-S spec is interesting, but it might be 5 years too late...

I've worked in the semiconductor industry for over 30 years, 17 of those at Intel, but that is now, 17 years ago
could go on but this thread is about Nvidia and ARM ...
But I totally agree - competition is a good thing and is what drives innovation in a lot of cases

 

[sw1tcher]

The Reuters wire story writer clearly is not very deep in their knowledge of Nvidia, by calling it an "AI" chip maker. Sure, Nvidia does that, but its bread and butter are graphics chips and graphics cards.

They refer to Nvidia as an "AI" chip company because that's where:
(1) Nvidia's growth is coming from, and
(2) the segment where Nvidia gets most of its sales from now.

Just look at Nvidia's Q2 2024 earnings report where Jensen Huang said: "A new computing era has begun. Companies worldwide are transitioning from general-purpose to accelerated computing and generative AI."

Nvidia's "Data Center" segment which includes their AI chips brought in $10.32 billion in revenue, up 141% from the previous quarter and up 171% from a year ago.

Their "Gaming" segment only did $2.49 billion in sales (up 11% from the previous quarter and up 22% from a year ago).

Since Nvidia is a public company, we can look at real numbers. For 2023, the page I found had their graphics revenue at 11.9 billion. Compute and networking (I'm assuming datacenter products) is at 15.1 billion. It's pretty evenly split. So is Nvidia an AI company? I guess so?

Data Center revenue for FY2023 was $15.01 billion, an increase of 41% vs FY2022 Data Center revenue of $10.61 billion.

Gaming revenue for FY2023 was $9.07 billion, a decrease of 27% vs FY2022 Gaming revenue of $12.46 billion.

$15.01 billion is 65.5% more than $9.07 billion. That's not an even split.


The growth in Data Center revenue for FY2024 vs FY2023 and the decline in Gaming revenue for FY2024 vs FY2023 will be even greater. We already saw it in their Q1 2024 and Q2 2024 earnings reports.

Q1 2024:

• Data Center revenue was (at the time) a record $4.28 billion, up 14% from a year ago.
• Gaming revenue was $2.24 billion, down 38% from a year ago.

Q2 2024:

• Data Center revenue hit a record $10.32 billion, up 171% from a year ago.
• Gaming revenue was $2.49 billion, up 22% from a year ago.

If the sales trend continues, Nvidia will bring in less revenue for Gaming for all of FY2024 than FY2023.

 

[picpicmac]

If the sales trend continues, Nvidia will bring in less revenue for Gaming for all of FY2024 than FY2023.

Yeah, I hadn't kept up with how Nvidia's corporate revenues have been changing.

The Qualcomm presser got all the attention in the latest ARMs race, but I will not count Nvidia out as a possible dark horse to be an important Microsoft enabler in the next few years.

Microsoft is aiming to be the system software monopoly for both Intel and ARM camps.

 

[deconstruct60]

Microsoft is aiming to be the system software monopoly for both Intel and ARM camps.

How? SoCs with Arm design 'CPU' cores in them for personal computing devices pragmatically covers tablets and phones at this point also. [ there are many 10's of millions of folks for whom their phone is their sole primary 'computer' device]

Microsoft has no hope of monoploticaly covering the phone market. Plainy evident by all of the Android integration suppose the are dramatically increasing in Windows. It is 'phone wars' are over; Windows didn't win.

Windows also certainly does not have 'monopoly' status on modern Macs either. Modern Macs can't (and don't particularly want to) natively boot Windows on raw iron. Kind of hard to be a 'monopoly' when you can only run in an actively support role as a 'guest' OS.

Microsoft having a 'monopoly' on the Windows subset is somewhat of a circular argument. You've defined the subset of there product so inside that 'circle' there is only Microsoft.

Arm systems here is only helping to keep the Windows subset from shrinking. It isn't really defining at 'monopoly'. It is more a 'stop loss'.

 

[Falhófnir]

How? SoCs with Arm design 'CPU' cores in them for personal computing devices pragmatically covers tablets and phones at this point also. [ there are many 10's of millions of folks for whom their phone is their sole primary 'computer' device]

Microsoft has no hope of monoploticaly covering the phone market. Plainy evident by all of the Android integration suppose the are dramatically increasing in Windows. It is 'phone wars' are over; Windows didn't win.

Windows also certainly does not have 'monopoly' status on modern Macs either. Modern Macs can't (and don't particularly want to) natively boot Windows on raw iron. Kind of hard to be a 'monopoly' when you can only run in an actively support role as a 'guest' OS.

Microsoft having a 'monopoly' on the Windows subset is somewhat of a circular argument. You've defined the subset of there product so inside that 'circle' there is only Microsoft.

Arm systems here is only helping to keep the Windows subset from shrinking. It isn't really defining at 'monopoly'. It is more a 'stop loss'.

Seems more like Microsoft realised how badly they missed the boat on mobile/tablet/Arm computing and are now scrambling to make sure that at least desktop Windows can tap into it going forward. It would also certainly make having another crack at a smartphone OS easier if Windows apps already had native Arm versions by and large.

 

[deconstruct60]

Seems more like Microsoft realised how badly they missed the boat on mobile/tablet/Arm computing and are now scrambling to make sure that at least desktop Windows can tap into it going forward. It would also certainly make having another crack at a smartphone OS easier if Windows apps already had native Arm versions by and large.

"having another crack". ... probably not happening. It isn't just integration they are running Android apps on Windows. The 'war' is over. Even if the governments come in and 'blow up' Android/iOS tight grip on the smartphone market the app inertia is extremely high. Additionally, Android is open source which have even more traction if the play store and core google services have barriers thrown at them. The barrier to entry to other phone players is quite low.

Likely just going to a concophony of layers on top of Android (and to a lessor extent iOS) rather than opening the door to yet another OS with completely different app API.

And Windows 'race to the bottom' on OS pricing ( free upgrades 'forever' ) is a weakend state to expansion. Dramatically even more divergent hardware platforms to chase for 'almost free' isn't really going to help Windows. Just going to lower the margins on it even more.


The relatively larger 2-in-1 tablet area is where they still have traction. Android still stumbles there on a regular basis and Apple is deeply committed to keeping some substantive separation between the iPadOS and macOS. blow up that traction off chasing some pie-in-the-sky tangent and Windows would be in world of 'hurt'.


Windows didn't displace an OS to get where it is, it largely walk into a vaccum/greenfield. It isn't pushing Linux out of servers and 'back office high end compute' . (bulk of Azure services isn't Windows based). Not pushing out a healthy macOS. iOS/Android ... nope.