Google Readies Custom Chip for Pixel Phones and Chromebooks

[JosephAW]

Feel sorry for those who adopt and the the project is canceled in a few years. Reminds me of the Windows phones Firefox phones and others. Maybe this will push Apple to innovate more. Might be a good thing. Let's just hope they don't have a killer app.

 

[rjohnstone]

Man Qualcomm will take a beating if Google leave.

Yeah. It might cost them a whole weeks worth of chip production. 😂

 

[ghanwani]

Technology-wise Google’s best days are behind it. Their top talent, including their founders, are gone. Right now they are busy defending their monopoly.

 

[tbayrgs]

Man Qualcomm will take a beating if Google leave.

Google’s quantity of chips is barely a studder in Qualcomm’s production. Unless they plan to license their chips to other OEMs, and they’re actually any good (Google hasn’t shown any consistency in their ability to product decent hardware of any sort), this won’t cause Qualcomm to lose a minute of sleep.

Outside of the negative optics of possibly losing a big named customer, this would hardly even register on Qualcomm's radar. Pixel's marketing and brand recognition is far, far, far greater than it's sales impact. Pixel sales have never even gotten large enough to get out of the "Other" category in market share analysis. In that sense, Pixel and Surface brands share the same boat.

Exactly.

I suspect, as a few others have mentioned, that a version for use in the consumer products is just a byproduct of their development for use in other areas.

 

[az431]

The reason why restaurant food tastes the same everywhere is because they source their ingredients from the same place and they hire the same people that keep moving from one restaurant to another.

It is the same with tech.

No, not the same at all because one is food and the other is computer hardware. That's like saying horses and cars are the same because they both can be used for traveling.

 

[Aston441]

Even with the vast vitriol you speak of... all but the most looney, even on an Android fan site have been begrudgingly giving Apple their processor props for the last few years!
That being said- I think there is a common anti-consumer enemy here in Qualcomm. They’ve had a stranglehold on chip design, tech, & manufacturing for long enough to stifle innovation from any smaller company trying to crack that market...as consumers, we should all rejoice that some of the VERY few companies with the capabilities to buck the shackles of these guys are doing exactly that.
Go Apple!
Go Google!

Agree.
Qualcomm went from being an incredible innovator in the late 1990s to being just another evil oligarch today.

kind of like Apple

 

[lkrupp]

Man Qualcomm will take a beating if Google leave.

Not at all. As has been pointed out Pixel phones are a niche market and don't sell in any significant numbers. That they are constantly touted by the tech media is irrelevant.

 

[az431]

Man Qualcomm will take a beating if Google leave.

Wrong. Qualcomm sells approximately half a billion Snapdragon chips a year to over 100 manufacturers. Google buys maybe 7 million a year, less than 1.5%. Relative to Qualcomm's overall revenue from all products, Google's chip purchases barely register.

 

[lkrupp]

Yay for competition!

Pixel phones are in no way shape or form competition to Apple or Samsung or Huawei. They are insignificant. So let Google make its own SOC but how will they do it? If they trample on Apple’s A series patents or Qualcomm’s Snapdragon patents the crap will hit the fan, especially from Qualcomm.

In my opinion this is a rumor that will never see the light of day.

 

[rp2011]

While this is certainly interesting, I see it going about as far as Amazon's Fire phone.

Exynos processors are currently lagging behind Qualcomm. I don't see them catching up any time soon.

In fact, I suspect that Samsung's processors use a lot of Qualcomm's technology. Google will most likely be paying royalties to Qualcomm anyway.

The only benefit I see is allowing Google to at least appear to not be last in line. Pixel's are currently released in October where Samsung's Galaxy S series is released in February and uses Qualcomm's latest processor. By the time Google releases their Pixel Qualcomm is already announcing their next chip.

Qualcomm won't provide the chips any sooner than that because the other, larger manufacturers are prioritized.

This may be Google's attempt to jumped to the head of the line or at least appear they are not at the end of the line.
[automerge]1586872061[/automerge]


Why would you want to use a low power, mobile processing chip in a server?

Everyone wants low power chips. That’s a lot of money running and cooling them. ARM chips are the future of cloud infrastructure

 

[oneMadRssn]

Right now estimates are that Google only sells about 6 million Pixel phones per year.

Samsung sells more phones in a couple weeks than Google sells in an entire year.

It might be good for Chromebooks, though. But are Chromebooks searching for new processor technologies?

I'll be shocked if this endeavour is still around in three years... knowing how Google cancels stuff.

How much will this cost Google to develop their own processors for their own phones?

If the chip is versatile and cost-effective, it can be used in their other hardware too. Home/Nest, Chromecast, and Wifi.

 

[Rigby]

The Pixel phones are great, probably the best Android phones out there. They have the Titan M security chip, come with a clean version of Android and always get quick updates, and you can unlock and re-lock the bootloader easily to install custom ROMs without losing security (which ironically makes them one of the best Android phones do "de-google"). The high end version has competitive hardware and performance, while the lower-end "a" version offers a lot for little money. I'm frankly surprised that Google doesn't sell more of them.

 

[DNichter]

Good first step for Google, but I'd imagine they are pretty behind when it comes to chip development. Apple really made all the right moves when it comes to iOS. I am not sure they will ever be beaten at this point.

 

[mtneer]

Doesn't Qualcomm require any phones sold in US to have Snapdragons? Isn't that why Samsung is forced to ship Snapdragons in North America while they use Exynos everywhere?

 

[RickDEGH]

Even with the vast vitriol you speak of... all but the most looney, even on an Android fan site have been begrudgingly giving Apple their processor props for the last few years!
That being said- I think there is a common anti-consumer enemy here in Qualcomm. They’ve had a stranglehold on chip design, tech, & manufacturing for long enough to stifle innovation from any smaller company trying to crack that market...as consumers, we should all rejoice that some of the VERY few companies with the capabilities to buck the shackles of these guys are doing exactly that.
Go Apple!
Go Google!

Absolutely! Only a few companies have the mettle to actually compete in certain spaces these days. I tried looking for cloud photo storage services (which is affordable and simple) and couldn’t get anything past the hegemony of Google/Apple/Amazon/Microsoft.

 

[CarpalMac]

I have owned every Pixel produced and can say they should take their heads out of their backsides and fix their failures before focussing on new areas to cock up and underwhelm in.

The camera and software has been key to me, as well as the pure Android experience. The camera is only marginally better than the Pro 11 now though (apart from Nightsight at x2 zoom as Apple don't do that), the hardware otherwise though is laughable. The screen flicker and hysterically bad battery life shows that they did absolutely zero usability testing. Only a cretin would have said "yeah, 48% battery remaining by 12pm and light use is fine".

Armchair analyst comments aside, they could have got a numpty like me in and I could have spec'd up a better and more desirable device.

Add to this, Google's ADT approach to shiny new things and then quickly getting bored. I cannot see how this will be anything other than another fail opportunity. Here's hoping I am wrong.

 

[brendu]

Man Qualcomm will take a beating if Google leave.

No they won’t. Google barely sells any phones.

 

[mi7chy]

Google is years behind Apple.

 

[name99]

Right now estimates are that Google only sells about 6 million Pixel phones per year.

Samsung sells more phones in a couple weeks than Google sells in an entire year.

It might be good for Chromebooks, though. But are Chromebooks searching for new processor technologies?

I'll be shocked if this endeavour is still around in three years... knowing how Google cancels stuff.

How much will this cost Google to develop their own processors for their own phones?

I'm not sure this is the correct framework.
THE fault line in tech is not between Apple and Google, or between TSMC and Samsung, it is between the engineering companies and the finance companies.
The engineering companies (Apple, Google, Facebook, SS, TSMC, AMD, AMZ, ...) care about changing the world, and believe a version of "it you build it they will come".
The finance companies (Intel, Qualcomm, pretty much all the telcos cable co's) care ONLY about the bottom line, and don't care about crippling or killing products the second they stop making money. They have zero understanding of a changing world (all their models are based on an assumption that the world stays pegged to exactly how it worked the day the model was created) and zero confidence that new products will create new markets.
(MS? As always a split personality, half the company stuck in the old model, half in the new...)

You may not LIKE the priorities of some of the engineering companies (Facebook...), others of them may sometimes be so incompetent that you have to wonder how long they can survive (SS), but THEY are the companies that change the world; the others only watch (and grab what they can...)

So look at Google in this context. On the mobile side they're stuck with QC and their utter lack of ambition (64-bit ARM, cancelling Falkor) and insistence now on pushing tech not based on quality but on their monopoly (everything about mm wave and the 5G chipsets). On the s86 side Intel is hardly an inspiring partner.
If you believe you can change the world, why not create your own chips? There is a lot of space of variation here. It doesn't have to mean your own core (A4, A5 were not Apple custom cores). Get a good core from ARM, a good GPU from somewhere else, max out the cache (unlike that cheapskate QC), add the largest NPU/TPU known to mankind (you know that's something Google cares about), add a HW AV1 encoder/decoder ...

And their own chip (again, not necessarily their own core) allows Google to grow beyond the blinkered limitations of their partners. They can try to create some halo Chromebooks that aren't just cheap junk? Maybe try for the tablet market again? Maybe downscale to an Android Wear device that doesn't feel like it belongs in 2013? There's even scope to use such chips (if well designed) in the data center for some of the more menial tasks.

The point is you don't understand Google (or Apple, or Facebook; though you do understand INTC or QC) by asking "what's the addressable market today, and how much revenue can we squeeze from it", but by asking "what totally new things would this allow us to do"?
[automerge]1586887552[/automerge]

Google is years behind Apple.

Maybe so, but the competitor is not Apple, it is Chinese Android, it is MS, it is Facebook, it is Amazon...

 

[Michael Scrip]

I'm not sure this is the correct framework.
THE fault line in tech is not between Apple and Google, or between TSMC and Samsung, it is between the engineering companies and the finance companies.
The engineering companies (Apple, Google, Facebook, SS, TSMC, AMD, AMZ, ...) care about changing the world, and believe a version of "it you build it they will come".
The finance companies (Intel, Qualcomm, pretty much all the telcos cable co's) care ONLY about the bottom line, and don't care about crippling or killing products the second they stop making money. They have zero understanding of a changing world (all their models are based on an assumption that the world stays pegged to exactly how it worked the day the model was created) and zero confidence that new products will create new markets.
(MS? As always a split personality, half the company stuck in the old model, half in the new...)

You may not LIKE the priorities of some of the engineering companies (Facebook...), others of them may sometimes be so incompetent that you have to wonder how long they can survive (SS), but THEY are the companies that change the world; the others only watch (and grab what they can...)

So look at Google in this context. On the mobile side they're stuck with QC and their utter lack of ambition (64-bit ARM, cancelling Falkor) and insistence now on pushing tech not based on quality but on their monopoly (everything about mm wave and the 5G chipsets). On the s86 side Intel is hardly an inspiring partner.
If you believe you can change the world, why not create your own chips? There is a lot of space of variation here. It doesn't have to mean your own core (A4, A5 were not Apple custom cores). Get a good core from ARM, a good GPU from somewhere else, max out the cache (unlike that cheapskate QC), add the largest NPU/TPU known to mankind (you know that's something Google cares about), add a HW AV1 encoder/decoder ...

And their own chip (again, not necessarily their own core) allows Google to grow beyond the blinkered limitations of their partners. They can try to create some halo Chromebooks that aren't just cheap junk? Maybe try for the tablet market again? Maybe downscale to an Android Wear device that doesn't feel like it belongs in 2013? There's even scope to use such chips (if well designed) in the data center for some of the more menial tasks.

The point is you don't understand Google (or Apple, or Facebook; though you do understand INTC or QC) by asking "what's the addressable market today, and how much revenue can we squeeze from it", but by asking "what totally new things would this allow us to do"?

I'm just giving my day-one skepticism.

But I agree that it would be great if we had more competition in the silicon market. Let's hope Google can do it!

Cheers

 

[Cosmosent]

My guess, Google is providing a Netlist of its Custom Hardware Blocks to Samsung, & chose this route primarily because it's the best low-power solution that keeps their IP protected.

In other words, NOT because it's the lowest-cost option.

 

[vmistery]

I’ve noticed people discounting this as pointless due to low sales but I wonder if they are looking at using these in more than just phones. Amazon are making their own DC chip designs it would make sense that google might do the same for that purpose and then extend it to phones and laptops. Or they might sell them to phone manufacturers to guarantee longer support cycles for the chips.

 

[teknikal90]

I think they'd be a bit late in the game. designing processors isn't something you just decide to do one year. Apple had to first buy PA semi and got into the Arm race pretty early with the A4 (a repackaged existing architecture) to be where they are now, with almost 10 years of accumulated design experience.

Even Samsung with their foundries and experience gained from that can't match what Apple and Qualcomm are doing.

 

[konqerror]

My guess, Google is providing a Netlist of its Custom Hardware Blocks to Samsung, & chose this route primarily because it's the best low-power solution that keeps their IP protected.

In other words, NOT because it's the lowest-cost option.

I agree. Basically what Apple did on the A4 and A5. Allows them to get what they want specifically with low levels of custom engineering.

Apple still uses a lot of 3rd party IP blocks today.

 

[Macaholic868]

The finest chip to power mobile spyware and adware that has ever been created!!!