News Qualcomm announces Nuvia-powered PC chip - competitive with Apple M series

Page 2 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

NTMBK

Lifer
Nov 14, 2011
10,208
4,940
136

Image_from_iOS__1_.jpg


I'm excited to see this pan out. If we can get CPUs as good as the M1 on PCs, I'm all for it. More competition is good!
 

NTMBK

Lifer
Nov 14, 2011
10,208
4,940
136
Anyway, the fact there is a 5G modem already tells me this is not a laptop chip, it is very likely this chip has small cores in it, a laptop chip cant have something like the A510 in it.

It's a laptop chip. It's claiming to be "setting the benchmark for Windows PCs"- as in, they think it will be faster than Intel and AMD.

Qualcomm spent $1.4bn acquiring Nuvia. They weren't going to throw that investment in the bin and use A510 crap.
 

Shivansps

Diamond Member
Sep 11, 2013
3,835
1,514
136

Because a modem is not something you need to target laptops, what to me, it means they are also targeting smaller devices, ultra-portables, 2-in-1, tablets... but for those you actually need small cores, i really hope thats not the case here.
Just look at the 8CX, a SoC designed to run ARM Windows, they placed 4 (four!!!) A55 cores in it along with a modem. This is because it was designed for the Surface Pro X.

Its similar to the diference from the M1 to the M1 Pro in Apple, and Apple has the huge advantage of having a small core that is waaay faster on top.
 

moinmoin

Diamond Member
Jun 1, 2017
4,933
7,619
136
Go to the next line: Leadership in sustained performance and battery life.

Sounds like competition for everyone.
The big difference always will be that Apple does its own stuff in its own "magical" world. Apple is a lifestyle company, and even without having any leadership in performance and/or efficiency Apple can easily fall back to "only" being a popular manufacturer of lifestyle products. Qualcomm, Intel and AMD can't do that.
 

Ajay

Lifer
Jan 8, 2001
15,332
7,792
136
You NORMALLY don't throw this kind of money at a problem without having some sort of business case for it.
So, if you haven't noticed, throwing this kind of money at a problem fails as often as it succeeds. The M&A world is littered with corpses. I doubt Qualcomm will have sufficient additional earning from that 1.4B price tag so that it pays itself back in any reasonable time frame. I think it's more about showing stock holders that the C-suite is doing all it can to acquire top talent, patents and attention, all to keep the stock price up. It's a bit of a shell game, IMHO. Maybe I'm wrong - we'll see.
 

Ajay

Lifer
Jan 8, 2001
15,332
7,792
136
Does anyone really believe that Qualcomm of all companies won't do everything in their power to normalize 5G in laptops? Offering 5G at cost or even at a moderate loss may be worth it in their calculations if they can create some market momentum behind 5G laptops and get people who buy x86 laptops to start demanding built in 5G. Because it is pretty obvious who a leading supplier of those 5G modems would be.

Fair point.

If they expect to lose that much with the 2023 iPhone, then it sounds like they expect the first Apple designed modems to appear with next year's iPhone. If so, maybe Apple starts including 5G in Macs with the M2 or M3 generation...

It makes sense as the 5G rollout continues - particularly in cities. The data rates, when the signal is good, it quite a bit better than LTE and much more useable out of the office.
 

JasonLD

Senior member
Aug 22, 2017
485
445
136
The big difference always will be that Apple does its own stuff in its own "magical" world. Apple is a lifestyle company, and even without having any leadership in performance and/or efficiency Apple can easily fall back to "only" being a popular manufacturer of lifestyle products. Qualcomm, Intel and AMD can't do that.
Well, that is not entirely true. Apple still has to lead the performance and/or efficiency vs competition in order to justify their premium pricing. Otherwise, PowerPC to Intel to AS transition wouldn’t have happened.
 

Joe NYC

Golden Member
Jun 26, 2021
1,893
2,192
106
So, if you haven't noticed, throwing this kind of money at a problem fails as often as it succeeds. The M&A world is littered with corpses. I doubt Qualcomm will have sufficient additional earning from that 1.4B price tag so that it pays itself back in any reasonable time frame. I think it's more about showing stock holders that the C-suite is doing all it can to acquire top talent, patents and attention, all to keep the stock price up. It's a bit of a shell game, IMHO. Maybe I'm wrong - we'll see.

It would be funny if the same employees who walked out on Apple and other companies to form this Nuvia, walked out on Qualcomm to start yet another Arm start up...
 
  • Like
Reactions: Tlh97 and Ajay

naukkis

Senior member
Jun 5, 2002
702
570
136
Because a modem is not something you need to target laptops

All laptops don't have LTE/5G modems in it because Intel/AMD doesn't have integrated one yet. But always connected laptop is the way things are going, even those x86-laptops will need to have LTE/5G modems soon.
 

Doug S

Platinum Member
Feb 8, 2020
2,201
3,405
136
Well, that is not entirely true. Apple still has to lead the performance and/or efficiency vs competition in order to justify their premium pricing. Otherwise, PowerPC to Intel to AS transition wouldn’t have happened.

Why? Despite having x86 for the past 15 years they have been behind what you could get in PC laptops most of that time because they only updated their products every year or two. Meanwhile Intel was releasing new SKUs every few months, and there are a list of PC laptops from multiple companies available with those new SKUs almost immediately. The only time Apple actually had a "lead" was a few times where they got Intel to provide them a custom SKU - like they did when they introduced the Macbook Air and basically created a new laptop category. Most of the time when Apple introduced new Macs they were only catching up to where the PC market was, or maybe front running the Dells of the world for a month or two since Intel gave them special treatment.

Apple had no choice with the PPC transition. The non-embedded segment of the PowerPC market had shrunk to IBM and Apple, and IBM only cared about workstations and servers starting at $5K and up (and up and up and up!) from there. The performance of PPC wasn't a problem, they were just targeting a totally different market segment. They would have had to fund the R&D for all the chips they needed for Macs since no one else would be using them. That was not really doable at the time when they were only a few years removed from near bankruptcy and almost all their R&D was going towards this crazy idea Jobs had of Apple selling smartphones.
 

Doug S

Platinum Member
Feb 8, 2020
2,201
3,405
136
It would be funny if the same employees who walked out on Apple and other companies to form this Nuvia, walked out on Qualcomm to start yet another Arm start up...


I've said before I expect this to happen. If Qualcomm doesn't enter the server market, GW III and his two sidekicks who left Apple because they wanted to do a datacenter design aren't going to stick around at Qualcomm for one minute beyond their lockup period. If they'd wanted to design smartphone and laptop SoCs they could have stayed at Apple.

Whether they found another company or retire in luxury depends on whether they're the type who always have to be doing something, or decide "work" is not something they ever care to do again once they have more money than they know what to do with.
 

scannall

Golden Member
Jan 1, 2012
1,944
1,638
136
I've said before I expect this to happen. If Qualcomm doesn't enter the server market, GW III and his two sidekicks who left Apple because they wanted to do a datacenter design aren't going to stick around at Qualcomm for one minute beyond their lockup period. If they'd wanted to design smartphone and laptop SoCs they could have stayed at Apple.

Whether they found another company or retire in luxury depends on whether they're the type who always have to be doing something, or decide "work" is not something they ever care to do again once they have more money than they know what to do with.
My gut feeling tells me the whole Nuvia thing was just a way to get a big payday. And that the server CPU's spiel was just fluff. I don't expect them to stay at Qualcomm any longer than they have to.
 
  • Like
Reactions: CHADBOGA

Ajay

Lifer
Jan 8, 2001
15,332
7,792
136
Apple had no choice with the PPC transition. The non-embedded segment of the PowerPC market had shrunk to IBM and Apple, and IBM only cared about workstations and servers starting at $5K and up (and up and up and up!) from there. The performance of PPC wasn't a problem, they were just targeting a totally different market segment. They would have had to fund the R&D for all the chips they needed for Macs since no one else would be using them. That was not really doable at the time when they were only a few years removed from near bankruptcy and almost all their R&D was going towards this crazy idea Jobs had of Apple selling smartphones.

I remember that well. Performance was an issue, though Jobs hyped up that problem 10 fold to justify to users the need to switch (I know, shocking ;)) . His team did kind of jab IBM in the eye by insisting that IBM get the G5 up to 1GHz without any alterations to the Altivec engine (which was presenting problems with scaling). Kind LOL today.

Apple was already recovering well before the switch the x86. Jobs pushed for creative new Macs designs and had top notch marketing at the time (well, and since). I almost invested in 2002 based dramatic changes within Apple and seeing some very successful investors dump a ton of money on Apple (I still regret passing on that). Moving over to the x86 platform, for a host of reasons as you well point out, was the only way to keep the Mac business profitable; and it succeeded.
 

Eug

Lifer
Mar 11, 2000
23,583
996
126
In other news, Apple will begin selling repair parts to individuals. They're starting with recent iPhones, but will eventually include Apple Silicon Macs. No mention of Intel Macs.

I remember that well. Performance was an issue, though Jobs hyped up that problem 10 fold to justify to users the need to switch (I know, shocking ;)) . His team did kind of jab IBM in the eye by insisting that IBM get the G5 up to 1GHz without any alterations to the Altivec engine (which was presenting problems with scaling). Kind LOL today.

Apple was already recovering well before the switch the x86. Jobs pushed for creative new Macs designs and had top notch marketing at the time (well, and since). I almost invested in 2002 based dramatic changes within Apple and seeing some very successful investors dump a ton of money on Apple (I still regret passing on that). Moving over to the x86 platform, for a host of reasons as you well point out, was the only way to keep the Mac business profitable; and it succeeded.
Actually in the keynote he said the G5 would hit 3 GHz in year, when he announced the 2 GHz G5. So, not up to 1 GHz but that it would increase by 1 GHz. And he didn't tell IBM ahead of time he would say that. Also, that was the time that one of the VPs coined the phrase "premature specification", when the G5 specs showed up briefly on Apple's live website before the keynote.

Also, a huge hint about the switch that we ALL missed was Jobs demonstrating Sherlock or something, showing a shipment to Apple. It was all in plain sight, but nobody caught on... that the origin of the shipment was from Intel's headquarters. He had used a real shipping waybill number, and the number was shown. This was before the Intel switch announcement.

BTW, I did invest in Apple in the mid 2000s... but then cashed out when they went up something like 50%. I thought that was a pretty good profit. If you had told me at that time that Apple would become a trillion dollar company, I would have called a psychiatrist to get you committed.

Fun fact: Apple released its first Intel Mac in January 2006. On December 28, 2006, the day before their quarterly statement, AAPL closed at $80.87.

Intel_C8087.jpg
 

Ajay

Lifer
Jan 8, 2001
15,332
7,792
136
Actually in the keynote he said the G5 would hit 3 GHz in year, when he announced the 2 GHz G5. So, not up to 1 GHz but that it would increase by 1 GHz. And he didn't tell IBM ahead of time he would say that. Also, that was the time that one of the VPs coined the phrase "premature specification", when the G5 specs showed up briefly on Apple's live website before the keynote.
Now I remember (so, 'I remember it well' was off, lol!). Jobs was a cagey ah heck!
 
  • Like
Reactions: Eug

NTMBK

Lifer
Nov 14, 2011
10,208
4,940
136
All laptops don't have LTE/5G modems in it because Intel/AMD doesn't have integrated one yet. But always connected laptop is the way things are going, even those x86-laptops will need to have LTE/5G modems soon.

Who the heck is going to pay for a second 5G data plan just to use their laptop off WiFi? When you can just tether to your phone, for the 2 times a year you might use it? It doesn't make sense, especially now that more people are working from home.
 
  • Like
Reactions: Thunder 57

Roland00Address

Platinum Member
Dec 17, 2008
2,196
260
126
The big difference always will be that Apple does its own stuff in its own "magical" world. Apple is a lifestyle company, and even without having any leadership in performance and/or efficiency Apple can easily fall back to "only" being a popular manufacturer of lifestyle products. Qualcomm, Intel and AMD can't do that.
What does this mean?

Most customers do not do a head to head comparison between products to see which is better, most customers do not do research.

Likewise the biggest competitor for Apple is their own products, are they good enough? And was the experience satisfactory that the customer will buy a new apple product and either do so for a new feature or they kept the device for 3 to 7 years (laptops, phones, etc have different turnover rates.)
 

moinmoin

Diamond Member
Jun 1, 2017
4,933
7,619
136
@Roland00Address Exactly.

My gut feeling tells me the whole Nuvia thing was just a way to get a big payday. And that the server CPU's spiel was just fluff.
The startups not trying to go for a big payday are a huge exception indeed. And I don't think the focus on server CPU was just fluff, for a startup it was the biggest possible TAM of potential buyers since there are more huge companies running their own more an more customized huge datacenters than there are companies that create more or less custom consumer SoCs on a large scale. So in a way it's "unfortunate" Qualcomm picked them up considering how many possible buyers there were in the datacenter space, but I doubt they mind considering the amount of money that moved.
 
  • Like
Reactions: Joe NYC

Eug

Lifer
Mar 11, 2000
23,583
996
126
Who the heck is going to pay for a second 5G data plan just to use their laptop off WiFi? When you can just tether to your phone, for the 2 times a year you might use it? It doesn't make sense, especially now that more people are working from home.
I have a 4G data plan for my wife's iPad 7. I could never teach her to tether the damn thing. Having the data plan is so much easier. Mind you, it's only a 4 GB plan and I only pay $5 a month for it.

Plus a few things:

1) The iPad on 4G doesn't use much extra power. It seems to last just as long as usual as it would be on WiFi, or at least close to it.

2) When I tether my iPad Pro to my iPhone, it kills the iPhone battery in short order. It's not a viable solution for extended periods when traveling.

3) Despite the supposed automatic tethering in iOS, there still is user interaction required to make it work. I have to start the tethering, and then after the iPad Pro is unused for a while, I have to start it again. Minor irritations, but irritating nonetheless.

4) My phone plan allows the addition of a cellular tablet or cellular watch for CAD$10 a month, sharing the phone's data plan. It gets its own phone number but cannot make phone calls or send/receive SMS. However, anything using data is fully supported. Basically it's full internet access for just US$8 a month, up to my 20 GB per month on my phone plan (of which I normally used less than half per month on my iPhone).

Based on this, I was determined to get my next iPad Pro with 5G... but then the pandemic happened and I stopped travelling. At this point I don't even need a new iPad Pro, much less a new iPad Pro with 5G. Now I use < 2 GB data on my iPhone per month, since I work from home 95% of the time, and the iPad Pro sits idle while I sit in front of my 30" Cinema Display at home. ;)

BTW, it might interest you to know that Apple actually designed a few 3G MacBook Pro prototypes. A couple of them even made it into the wild, but none were ever sold.


Screen Shot 2021-11-17 at 7.53.39 PM.png
 

lobz

Platinum Member
Feb 10, 2017
2,057
2,856
136
They didn't always have to in the past. Their basis is the religion they've managed to built to last even through obvious underperforming times.
That being said they're really up there right now, in the technological sense too.
 
  • Like
Reactions: Tlh97 and Joe NYC

moinmoin

Diamond Member
Jun 1, 2017
4,933
7,619
136
I still do not know what you mean by a lifestyle brand, and what heuristic you are trying to convey that is intuitive to you but I can not see.

————

I do not think it matters, and I am moving on since you say we agree.
Yes, I think we agree. Maybe we put a different name on it, but nothing you wrote I disagree with. :D My point was the difference of Apple's competitive environment to that of Qualcomm, Intel and AMD. You elaborated that well for Apple.
 

DrMrLordX

Lifer
Apr 27, 2000
21,582
10,785
136
Because a modem is not something you need to target laptops, what to me, it means they are also targeting smaller devices, ultra-portables, 2-in-1, tablets... but for those you actually need small cores, i really hope thats not the case here.
Just look at the 8CX, a SoC designed to run ARM Windows, they placed 4 (four!!!) A55 cores in it along with a modem. This is because it was designed for the Surface Pro X.

Its similar to the diference from the M1 to the M1 Pro in Apple, and Apple has the huge advantage of having a small core that is waaay faster on top.

Most of the WARM lappies out there that I researched were available almost exclusively from Verizon or AT&T and could only be bought with a mobile data plan. Qualcomm wants to be able to continue to service that market.
 

Eug

Lifer
Mar 11, 2000
23,583
996
126
Personally I think with the laptop market, the next evolutionary step is 5G support with lower power silicon. Apple already has this in its iPad Pros and is already developing their own 5G chip, with IP purchased from Intel. 5G is Qualcomm's forte, and given the direction of the technology, it makes perfect sense for Qualcomm to double down on it, as it is a major differentiator from the competition, and with favourable cell plan structures from some carriers, it makes sense for pros, prosumers, and even some consumers to get 5G non-tethered.

Rumour has it that Apple will not be integrating their 5G chip into their A series SoC as of yet, probably because only sells a subset of devices using A series chip have 5G activated and they're new at this type of engineering. However, eventually when Apple gets better at 5G low power design and the silicon cost drops, I wouldn't be completely surprised to see 5G integrated. In the meantime, Qualcomm gets to be first to market on this, and dominates in this segment on the Windows side.
 

Doug S

Platinum Member
Feb 8, 2020
2,201
3,405
136
Personally I think with the laptop market, the next evolutionary step is 5G support with lower power silicon. Apple already has this in its iPad Pros and is already developing their own 5G chip, with IP purchased from Intel. 5G is Qualcomm's forte, and given the direction of the technology, it makes perfect sense for Qualcomm to double down on it, as it is a major differentiator from the competition, and with favourable cell plan structures from some carriers, it makes sense for pros, prosumers, and even some consumers to get 5G non-tethered.

Rumour has it that Apple will not be integrating their 5G chip into their A series SoC as of yet, probably because only sells a subset of devices using A series chip have 5G activated and they're new at this type of engineering. However, eventually when Apple gets better at 5G low power design and the silicon cost drops, I wouldn't be completely surprised to see 5G integrated. In the meantime, Qualcomm gets to be first to market on this, and dominates in this segment on the Windows side.


Huh?

Apple isn't designing a "5G modem", they are designing a cellular modem capable of 5G, and older standards like LTE (and maybe 3G, but I'll bet they ditch 3G and might even require VoLTE/Vo5G for voice as that would greatly simplify the task) Whether a customer actually uses 5G is irrelevant, every iPhone uses some sort of cellular and that's where the overwhelming majority of A series SoCs go.

I think you're probably right about keeping the modem separate for the first iteration, because if something goes wrong their backup plan would have been to continue getting modems from Qualcomm an extra year. The announcement from Qualcomm seems to indicate Apple is ready to use their modem and informed Qualcomm they wouldn't be ordering any (or at least a lot fewer) modems for newer models beyond the iPhone 13.

So we should probably expect discrete Apple modems for the iPad Pro next spring - that gives them a real world test so they can fix bugs in the baseband (which is likely where all the problems will be, the hardware is simple by comparison) they missed during internal testing. Then discrete Apple modems for all (or at least most) 2022 iPhone models, then they'll integrate them on die for the N3 based iPhone of 2023.

Having them on the A17 die when not all products using it need cellular isn't a problem. I'm not sure how the cellular patent licensing works, but if they fuse off the modem so it is permanently inoperable they presumably would not owe patent royalties for that chip, so all it would cost is the area. If they care about the area or fusing off the modem isn't good enough to avoid royalties the alternative would be to have two dies sharing a floorplan, like they do with the M1 Pro/M1 Max and the ones that don't need cellular are the "chop".

To reach 80% of iPhone sales by the intro of the 2023 model Apple would probably have to include their modem in all iPhones other than the iPhone 13 (which will be the "two year old" model at the time) and the rumored newer iPhone SE that will appear next spring. There's no way they can have it in 80% of iPhones sold without including it in at least a large number of iPhone 14s next year, because the latest model doesn't account for anywhere near 80% of their unit sales.
 

Eug

Lifer
Mar 11, 2000
23,583
996
126
Huh?

Apple isn't designing a "5G modem", they are designing a cellular modem capable of 5G, and older standards like LTE (and maybe 3G, but I'll bet they ditch 3G and might even require VoLTE/Vo5G for voice as that would greatly simplify the task) Whether a customer actually uses 5G is irrelevant, every iPhone uses some sort of cellular and that's where the overwhelming majority of A series SoCs go.
What I meant by subset of A series chips requiring 5G cellular is that most iPads currently don't include it. I realize iPad sales pale in comparison to iPhone sales though.

However, going forward after the first iteration, this could be a binning thing. iPads without cellular can use binned chips with defective cellular, as well as good chips with deactivated cellular (if yields are good).

Having them on the A17 die when not all products using it need cellular isn't a problem. I'm not sure how the cellular patent licensing works, but if they fuse off the modem so it is permanently inoperable they presumably would not owe patent royalties for that chip, so all it would cost is the area. If they care about the area or fusing off the modem isn't good enough to avoid royalties the alternative would be to have two dies sharing a floorplan, like they do with the M1 Pro/M1 Max and the ones that don't need cellular are the "chop".
That too. It would just cost die area.

Anyhow, I think this is the next step in the evolution of mobile computing devices. Back in the day, cellular was a HUGE luxury, and then everyone got a cell phone. Early on in the cell phone craze, cell plans were insanely expensive, but now in some countries (not mine) they're downright cheap, often with unlimited data along with cheap add-ons for non-phone devices for shared data. Hell, even my Apple Watch supports cellular, and I didn't even buy it for cellular per se. I bought it because it was stainless steel with sapphire front crystal, which wears a lot better than the usual aluminum and glass, and it just so happens they all include cellular too.

It is in a way like WiFi. Early on during the days of WiFi, WiFi was an expensive add-on, and even laptops didn't come standard with WiFi, whereas these days pretty much all mainstream consumer desktops come with WiFi. No, cellular won't be as ubiquitous as WiFi, but it will likely be very common on mobile devices.
 
Last edited: