Array
(
    [content] => 
    [params] => Array
        (
            [0] => /forum/threads/how-intel-lost-the-sony-playstation-business.21007/
        )

    [addOns] => Array
        (
            [DL6/MLTP] => 13
            [Hampel/TimeZoneDebug] => 1000070
            [SV/ChangePostDate] => 2010200
            [SemiWiki/Newsletter] => 1000010
            [SemiWiki/WPMenu] => 1000010
            [SemiWiki/XPressExtend] => 1000010
            [ThemeHouse/XLink] => 1000970
            [ThemeHouse/XPress] => 1010570
            [XF] => 2021770
            [XFI] => 1050270
        )

    [wordpress] => /var/www/html
)

How Intel lost the Sony PlayStation business

I mean... the article notes backwards compatibility issues, pricing, and I noticed that it said the decision was made in 2022. 2022 was also the year that Intel ARC GPUs came out and flopped. Sony probably made their decision based upon ARC's horrible performance and game compatibility at launch.

Even though this is a big loss for Intel - they never competed in console previously because they lacked GPUs. Their GPU design flopped and was clearly not mature enough yet. Sony likely made their decisions based upon that.

If the PlayStation 6 came out with the sort of issues that Intel ARC had - the brand would be absolutely ruined. Intel has not delivered compelling iGPU performance at all (the independent reviews are not out for Lunar Lake yet, so that might change shortly) so Sony's decision makes perfect sense.

The article's wording seems to suggest Intel would fabricate the wafers for the PS6 also. The only way Intel could have maybe gotten the deal was through an extremely competitive price, though I'm not so sure that even if Intel essentially sold at cost they would have been more attractive. Sony definitely could have pointed at the mess ARC was and demanded even lower prices.

Think of the risks Sony would be undertaking - betting on Intel actually succeeding on the process side of things (a risky bet for sure - but at least the process teams have proven results and delivered on time before the 10nm fiasco) , AND ALSO then having to bet that the GPU design team would get their shit together in time for a console chip to launch? The flop that was ARC would definitely swing the scales in TSMC and AMD's favor. To me, it seems like this is one more reason for Pat to be mad at Raja.
 
I mean... the article notes backwards compatibility issues, pricing, and I noticed that it said the decision was made in 2022. 2022 was also the year that Intel ARC GPUs came out and flopped. Sony probably made their decision based upon ARC's horrible performance and game compatibility at launch.

Even though this is a big loss for Intel - they never competed in console previously because they lacked GPUs. Their GPU design flopped and was clearly not mature enough yet. Sony likely made their decisions based upon that.

If the PlayStation 6 came out with the sort of issues that Intel ARC had - the brand would be absolutely ruined. Intel has not delivered compelling iGPU performance at all (the independent reviews are not out for Lunar Lake yet, so that might change shortly) so Sony's decision makes perfect sense.

The article's wording seems to suggest Intel would fabricate the wafers for the PS6 also. The only way Intel could have maybe gotten the deal was through an extremely competitive price, though I'm not so sure that even if Intel essentially sold at cost they would have been more attractive. Sony definitely could have pointed at the mess ARC was and demanded even lower prices.

Think of the risks Sony would be undertaking - betting on Intel actually succeeding on the process side of things (a risky bet for sure - but at least the process teams have proven results and delivered on time before the 10nm fiasco) , AND ALSO then having to bet that the GPU design team would get their shit together in time for a console chip to launch? The flop that was ARC would definitely swing the scales in TSMC and AMD's favor. To me, it seems like this is one more reason for Pat to be mad at Raja.
Still, can't have done any harm to Sony's negotiating position with AMD to have Intel in the mix. And is this sort of business actually as profitable for the silicon vendors as x86 PC/laptops and servers ?

Seems curious that this all happened 2 years ago and the story is only coming out now ?!
 
Still, can't have done any harm to Sony's negotiating position with AMD to have Intel in the mix. And is this sort of business actually as profitable for the silicon vendors as x86 PC/laptops and servers ?
Not nearly as profitable, but probably good enough to keep employees busy and pay them.

Seems curious that this all happened 2 years ago and the story is only coming out now ?!
Running out of dirt throwing at Intel.
 
It's a pretty recurring thing where Intel has been known to refuse any business below a certain margin threshold. This is because they have a monopolistic "price setter" mindset where if you want to do business with Intel you have to do things on Intel's terms.

Intel is not a monopoly anymore so companies can easily say "thanks but not thanks, we'll take our business somewhere else". Intel needs to get used to actually competing for business again, which means making comprimises on margin to make sure you get the customer.
 
In a non distant past, Intel displaced AMD in a last minute move and won the XBOX contract:

"AMD has since gone on to be a stalwart of Xbox CPUs. The switch from Intel to AMD happened with the Xbox One in 2013 and continued for the Xbox Series X and Series S. The chipmaker also provided the CPUs for Sony's PlayStation 4, released in 2013,"

Those two wins in 2013 were lifesavers for an AMD that was in a near death state, far worse than even Intel right now. They have been the incumbent ever since. Intel probably would have had to offer ridiculous pricing and support to retake the business.
 
It's a pretty recurring thing where Intel has been known to refuse any business below a certain margin threshold. This is because they have a monopolistic "price setter" mindset where if you want to do business with Intel you have to do things on Intel's terms.

Intel is not a monopoly anymore so companies can easily say "thanks but not thanks, we'll take our business somewhere else". Intel needs to get used to actually competing for business again, which means making comprimises on margin to make sure you get the customer.

Yet, Intel announced that it won Amazon's chips design and foundry contract today. Compare to the lost Sony contract, Amazon contract is much smaller in terms of total wafers needed every year. It makes me believe that Intel has more untold reasons to decline Sony's request, such as fab capacity, limitation of CapEx, and technology complexity.

 
Last edited:
"AMD has since gone on to be a stalwart of Xbox CPUs. The switch from Intel to AMD happened with the Xbox One in 2013 and continued for the Xbox Series X and Series S. The chipmaker also provided the CPUs for Sony's PlayStation 4, released in 2013,"

Those two wins in 2013 were lifesavers for an AMD that was in a near death state, far worse than even Intel right now. They have been the incumbent ever since. Intel probably would have had to offer ridiculous pricing and support to retake the business.

You’ve raised an important point. Does Intel actually profit from its foundry services, especially under heavy price pressure from TSMC customers (direct or indirect) who are using Intel Foundry as leverage? If that’s the case, it makes sense for Intel to walk away from the massive Sony PlayStation 6 order and instead accept a much smaller order for producing Amazon's chips.

Intel and Intel Foundry can afford to lose some money on smaller orders to build market share, but it wouldn’t make sense to incur significant losses on large orders.
 
Last edited:
You’ve raised an important point. Does Intel actually profit from its foundry services, especially under heavy price pressure from TSMC customers (direct or indirect) who are using Intel Foundry as leverage? If that’s the case, it makes sense for Intel to walk away from the massive Sony PlayStation 6 order and instead accept a much smaller order for producing Amazon's chips.

Intel and Intel Foundry can afford to lose some money on smaller orders to build market share, but it wouldn’t make sense to incur significant losses on large orders.
The question is whether Intel was doing Intel CPU + GPU or if they were partnering with Nvidia. If they were working with Nvidia, then they wouldn't have have nearly as much leverage on pricing - and I don't think Intel would be willing to subsidize Nvidia's profits and take a loss just to load their fabs.
 
Still, can't have done any harm to Sony's negotiating position with AMD to have Intel in the mix. And is this sort of business actually as profitable for the silicon vendors as x86 PC/laptops and servers ?

Seems curious that this all happened 2 years ago and the story is only coming out now ?!

"Seems curious that this all happened 2 years ago and the story is only coming out now ?!"

My guess is that Intel intentionally revealed this information because it had some good news to offset this bad news, such as the Amazon chips deal, making IFS a subsidiary, and the $3 billion 'Secure Enclave' award. If Intel didn’t preemptively reveal it, the information would eventually come out, potentially causing unpredictable and negative impacts.
 
You’ve raised an important point. Does Intel actually profit from its foundry services, especially under heavy price pressure from TSMC customers (direct or indirect) who are using Intel Foundry as leverage? If that’s the case, it makes sense for Intel to walk away from the massive Sony PlayStation 6 order and instead accept a much smaller order for producing Amazon's chips.
Gaming consoles traditionally haven’t been a very profitable business compared to client PCs - custom engineered, mid-volume (Sony sold 60M over the past 4 years) SoCs that are essentially high-end graphics PCs. Not sure which Intel node Playstation 6 was best suited for, but probably not 18A. Guessing that the Amazon deal was a better fit for where Intel wants to be - AI plus Xeon 6 on 18A.
 
Yet, Intel announced that it won Amazon's chips design and foundry contract today. Compare to the lost Sony contract, Amazon contract is much smaller in terms of total wafers needed every year. It makes me believe that Intel has more untold reasons to decline Sony's request, such as fab capacity, limitation of CapEx, and technology complexity.


Amazon is a big chip buyer and presumably has a lot of clout to force Intel's hand. Also Intel might be looking for a big name foundry customer to generate good PR.
 
Amazon is a big chip buyer and presumably has a lot of clout to force Intel's hand. Also Intel might be looking for a big name foundry customer to generate good PR.

Amazon is always a tough buyer to negotiate the price, not just the chips.
 
Back
Top