Sporadic Amazon UK chart update
- Normal PS5 disc sku in stock for last 36 hours at least (now #31 in the monthly charts, moving up slowly)
- GoW stuck in the best sellers daily chart at #100
Sporadic Amazon UK chart update
- Normal PS5 disc sku in stock for last 36 hours at least (now #31 in the monthly charts, moving up slowly)
- GoW stuck in the best sellers daily chart at #100
The availability of the standard PS5 edition means that premium bundles will likely make up a small portion of overall sales. This is already reflected in Amazon hourly chart.
Additionally, bundle deals often included games at no extra cost at this point in previous generations.
AMD sold Sony some $3.776 billion worth of chips for PlayStation 5 game consoles in 2022, which accounted for 16% of the company's revenue for the year, according to the company's filing with the SEC.
Series X (standalone) is actually in stock at Microsoft's store right now. But yeah, they really need to fix the supply issues of the console. It's strange that this is still such an issue. I wonder what's going on.
So Sony and Microsoft?
Does only mention Gaming and Client. You would assume Microsoft also has a lot of Datacenter revenue. Or maybe they don't order those directly from AMD.
Anyways the numbers make sense. Series X SOC is bigger, so more revenue for AMD.
This year sales are down and shift towards the smaller Series S SOC leads to dropping below 10%.
Sonys 6nm is cheaper, yet they increased share AND nominal revenue over the year ago.
Ampere is expecting normal supply for Series X only on H2. Don't know what type of informations they have, but if they're right the lack of stock will remain for a while.
So Sony and Microsoft?
Does only mention Gaming and Client. You would assume Microsoft also has a lot of Datacenter revenue. Or maybe they don't order those directly from AMD.
Anyways the numbers make sense. Series X SOC is bigger, so more revenue for AMD.
This year sales are down and shift towards the smaller Series S SOC leads to dropping below 10%.
Sonys 6nm is cheaper, yet they increased share AND nominal revenue over the year ago.
Ohh, you're right.
Totally forgot they changed reporting segments.
Gaming and Client are two different ones.
2021 had the Laptop and PC Corona boom. Lenovo makes sense.
But Sony being responsible for over 50% of total yearly gaming revenue is madness.
Shows how bad AMDs dedicated GPU business is doing.
Series X (standalone) is actually in stock at Microsoft's store right now. But yeah, they really need to fix the supply issues of the console. It's strange that this is still such an issue. I wonder what's going on.
Seems they thought $300 trumped all. Going by the internet discourse before launch and in the media they probably felt their strategy was perfect. Get the hype of the world’s most powerful console and the bring in the masses with the cheapest next gen console. Then pandemic hit and there was an even greater pivot to series S.
At the moment I don’t think MS wants to manufacture a lot of series x and is waiting to push it in the second half of year when they have the software to justify increasing X production.
The only reasonable justification for Series X production being so awful is they are working on a revision. PS5 sales grew a ton once they started working on 6nm.
Xbox being so late to moving on from 7nm maybe points to them wanting to go with 5nm, but that probably means we don't see that happening until August/September, which lines up with PS5 last year, PS4/XB1 slim in 2016, and the PS3 slims.
If they are working on 5nm, I see two scenarios happening. 1, they release the new 5nm X and S at the same time as Starfield, September 6, and treat it like a relaunch. 2, they release the X revision first, in June (like the 360 slims), and then the S revision is launched with Starfield.
The reason why Microsoft would want 5nm, and why it's likely that they are working on it and not 6nm, is because of the size of the X chip.
PS5 7nm SOC was 308nm2, and the 6nm SOC is 260nm2.
XB1 S SOC was 240nm2, and the PS4 slim was 209nm2.
Series X 7nm SOC is 360nm2. A 6nm revision would bring it down to ~310nm2, and a 5nm revision would bring it down to ~260nm2.
That's probably happening. The S would also go from 197nm2 to ~140nm2.
As a rough example, let's say Microsoft have scheduled 1M Series X and 3M Series S for production. On 7nm, they would need to order the equivalent of 360,000,000nm2 for the X, and 591,000,000nm2 for the S, for a total of 951,000,000nm2. Let's say for a 5nm revision, they stick with producing only 3M Series S, and the rest go to the X. Microsoft use the same total of 951M.
While keeping Series S unit production the same on 5nm as 7nm, ordering the same total of silicon as 7nm on 5nm, and using the rest for the X, can produce 1M additional units.
I hope my little game theory is accurate. Otherwise Microsoft is intentionally throwing away all the potential they had from 2021 to mid 2022. I'd also go one step further and say they should drop the disc drive, go digital only with the X revision, and price it at $399. Do something interesting. The One S did so well in 2016 because it was an interesting and exciting rebrand for Xbox One.
Makes only sense if they already preplanned to do that anyways.
The reason Sony went from 7nm to 6nm is because of the easy transition and the same design rules. It's really cheap for a nice benefit if the volume is there.
I don't think Xbox Series X has the volume to make the 7nm -> 5nm shrink worthwhile.
Series S is already cheap. Especially everything around the SOC. That's where Series X has the most potential for cost reductions.
Nothing makes sense with Microsoft and Xbox supply.
What I believe could make sense. Move the Series S from 7nm to 6nm. Move Series X from 7nm to 5nm with an extremely reduced chassis and other components.
Right now RTX4000 and RX7000 series GPUs, as well as Zen 5 are using TSMC 5nm nodes. Those are expensive as hell.
I hope my little game theory is accurate. Otherwise Microsoft is intentionally throwing away all the potential they had from 2021 to mid 2022. I'd also go one step further and say they should drop the disc drive, go digital only with the X revision, and price it at $399. Do something interesting. The One S did so well in 2016 because it was an interesting and exciting rebrand for Xbox One.
it would be even better imo if they just copy sony and do a detachable drive. this completely eliminates the fomo of buying a digital only console as you can "always buy the disc drive later if you want". The vast majority of course will likely never end up buying the disc, but the option is very powerful.
Either way like you I believe they need to do something to shake things up and I'd love to see a 399 X DE
The only reasonable justification for Series X production being so awful is they are working on a revision. PS5 sales grew a ton once they started working on 6nm.
Xbox being so late to moving on from 7nm maybe points to them wanting to go with 5nm, but that probably means we don't see that happening until August/September, which lines up with PS5 last year, PS4/XB1 slim in 2016, and the PS3 slims.
If they are working on 5nm, I see two scenarios happening. 1, they release the new 5nm X and S at the same time as Starfield, September 6, and treat it like a relaunch. 2, they release the X revision first, in June (like the 360 slims), and then the S revision is launched with Starfield.
The reason why Microsoft would want 5nm, and why it's likely that they are working on it and not 6nm, is because of the size of the X chip.
PS5 7nm SOC was 308nm2, and the 6nm SOC is 260nm2.
XB1 S SOC was 240nm2, and the PS4 slim was 209nm2.
Series X 7nm SOC is 360nm2. A 6nm revision would bring it down to ~310nm2, and a 5nm revision would bring it down to ~260nm2.
That's probably happening. The S would also go from 197nm2 to ~140nm2.
As a rough example, let's say Microsoft have scheduled 1M Series X and 3M Series S for production. On 7nm, they would need to order the equivalent of 360,000,000nm2 for the X, and 591,000,000nm2 for the S, for a total of 951,000,000nm2. Let's say for a 5nm revision, they stick with producing only 3M Series S, and the rest go to the X. Microsoft use the same total of 951M.
While keeping Series S unit production the same on 5nm as 7nm, ordering the same total of silicon as 7nm on 5nm, and using the rest for the X, can produce 1M additional units.
I hope my little game theory is accurate. Otherwise Microsoft is intentionally throwing away all the potential they had from 2021 to mid 2022. I'd also go one step further and say they should drop the disc drive, go digital only with the X revision, and price it at $399. Do something interesting. The One S did so well in 2016 because it was an interesting and exciting rebrand for Xbox One.
It's just as possible that MSFT massively overestimated how price conscious the console gaming market was. They may have thought Series S with its ~$200 discount would be a big seller vs XSX/PS5, and so allocated most of their chip orders to Series S instead of X.
PS5 sales grew a ton due to demand being there. Capacity =/ demand
On node moves, it doesn't make much sense. TSMC capacity on both 6/7nm and 4/5nm is declining and notably under full capacity. It's not like there is a shortage of manufacturing capacity here.
Furthermore, Sony going to 6nm is quite different to MSFT going to 5nm.
6nm represented ~25-30% increase in chips per wafer while likely being a ~10% increase in cost
5nm is 60% more expensive than 7nm
It makes no sense to go to 5nm. 6nm is substantially cheaper and gives notable capacity increases, but again its not necessary as TSMC capacity is not substantially under utlised and Xbox does not have the volume to maximise said capacities either.
No to mention 6nm is actually considered the same line as 7nm meaning its an easier switch. 5nm is not, further incurring more losses and maybe fault rates.
It's just as possible that MSFT massively overestimated how price conscious the console gaming market was. They may have thought Series S with its ~$200 discount would be a big seller vs XSX/PS5, and so allocated most of their chip orders to Series S instead of X.
PS5 sales grew a ton due to demand being there. Capacity =/ demand
On node moves, it doesn't make much sense. TSMC capacity on both 6/7nm and 4/5nm is declining and notably under full capacity. It's not like there is a shortage of manufacturing capacity here.
Furthermore, Sony going to 6nm is quite different to MSFT going to 5nm.
6nm represented ~25-30% increase in chips per wafer while likely being a ~10% increase in cost
5nm is 60% more expensive than 7nm
It makes no sense to go to 5nm. 6nm is substantially cheaper and gives notable capacity increases, but again its not necessary as TSMC capacity is not substantially under utlised and Xbox does not have the volume to maximise said capacities either.
No to mention 6nm is actually considered the same line as 7nm meaning its an easier switch. 5nm is not, further incurring more losses and maybe fault rates.
A TSMC 7nm 300mm wafer is $10,000 and and 5nm is $17,000 from what I could find.
300mm wafer is 70695mm2
360mm2 7nm Series X SOC would mean 196 total chips per wafer, or $51.02 per chip of the wafer cost.
A 260mm2 5nm Series X SOC would mean 271 total chips per wafer, or $62.73
An increase of $11.71 for the SOC, but that's would be countered by the removal of the BD drive. A standard BD drive was $18 back in 2016, and the Xbox One S UHD BD drive was estimated to be $33.50
Despite a small increase in bill-of-materials cost, the new Xbox One S delivers for consumers in terms of improved performance and features, according
www.businesswire.com
I couldn't find any breakdown for how much the Series X or PS5 BOM was, but assuming the lowest possible savings and the UHD drive fell down to $18, and the high of the One S $33.50, removing that would mean a total saving of anywhere between $6.29 and $21.79 when subtracting the price increase of the new SOC.
There's then the additional cost savings that would come from no disc drive and smaller SOC, such as a significant decrease in potential space for plastic and potential standard board design.
The Series S motherboard looks like this
That's a standard motherboard for any console. Unfortunately for Microsoft, the Series X is anything but standard. They have a split motherboard design and need to have a dedicated aluminum block to separate the two
I think a 5nm revision would allow for a significant rethink of how to simplify the Series X internals and the cost savings from this, plus the removal of the BD drive, should ultimately come out as a net cost reduction for the Series X.
For the Series S, using the same wafer math above, they can make 360 chips per 300mm 7mm wafer (SOC is 197mm2), or $27.78 per chip. 5nm would be 141mm2, so 501 chips per wafer, or $33.93 per chip.
A $6.15 increase, but with the potential savings from a total smaller motherboard and even smaller plastic chassis, it's possible a 5nm Series S is ultimately a wash in BOM, if not slightly cheaper.
Sony did order more wafers along with the 6nm revision, but if we use napkin math, we could see what Microsoft could spend on 7nm for a year and what happens if they spend the same on 5nm.
Recap: 7nm wafer is $10,000 and 5nm is $17,000
Let's say for 2022 Microsoft produced 4M Series X and 6M Series S. Microsoft could make at most 196 X chips per $10,000 wafer, and 360 S chips per $10,000 wafer
6M / 360 = 16,667 minimum wafers, or $166,670,000
4M / 196 = 20,409 minimum wafers, or $204,090,000
That's $370,760,000 total Microsoft spent on wafers, and 37,076 wafers. Series X would have 55% of wafers allocated to it. Let's say they spend exactly this much for 5nm. They would only buy 21,809 wafers at $17,000.
Let's say the same 55:45 allocation remains. That would be 11,995 X wafers and 9,814 S wafers.
Math above showed 271 X chips per 5nm wafer, and 501 S chips.
271 * 11,995 = 3,250,645 X chips
501 * 9,814 = 4,916,814 S chips
So if Microsoft spent exactly as much on 7nm and 5nm, Xbox would produce less total units by 19% on the X and 18% on the S.
There could be an argument that Microsoft can reinvest the costs saved by lower BOM of a digital 5nm Series X and smaller Series S back into buying more 5nm wafers, but the total cost decrease would be nearly impossible to truly work out.
If we look at the BD Drive costing anywhere between $18 and $33.50, that's a potential $72M and $134M in savings (from 4M X). Throwing that into wafer purchases would look like this.
$370,760,000 + $72,000,000 or $134,000,000 = $442,760,000 or $504,760,000
That's anywhere between 26,044 and 29,691 5nm wafers.
Let's say the same 55:45 allocation remains. That would be 14,324 to 16,330 X wafers and 11,719 to 13,360 S wafers.
Math above showed 271 X chips per 5nm wafer, and 501 S chips.
271 * 14,324 to 16,330 = 3,881,804 to 4,425,430 X chips
501 * 11,719 to 13,360 = 5,871,219 to 6,693,360 S chips
So reinvesting the cost saved from just the BD drive could bring Xbox to producing slightly more units by about 11-12%. With overall cost reduction from less plastic and simpler Series X internals, reinvesting that back into buying more wafers could see an additional 20% or higher units produced. Maybe.
If Microsoft wanted the same number of wafers on 5nm as 7nm (37,076 but let's round it down to 37,000) that would cost $629,000,000.
Again, 55:45 allocation
271 * 20,350 = 5,514,850 X chips
501 * 16,650 = 8,341,650 S chips
Now at this point, the allocation probably goes way more into the X's favor. I don't think 8M S is necessary,
13,000 S wafers = 6.5M
24,000 X wafers = 6.5M
Maybe. I'd say Microsoft needs to spend more in production as it is, the X is still hard to find for some reason and in its largest markets is not consistently available.
Also, just rechecking, but it's possible Microsoft just keeps buying only 40K wafers since 2020
An update on TSMC's upcoming 5-nanometer process technology.
fuse.wikichip.org
Using a 9% defect rate, a wafer of 196 X 7nm chips would actually mean only 178 are usable. 360 S chips would turn into 327.
So to get to 4M X and 6M S, Microsoft would need
4,000,000 / 178 = 22,472
6,000,000 / 327 = 18,349
That totals 40,821 wafers, which might actually be close to how many wafers Microsoft buys a year? Let's just drop it to 22K and 18K wafers each then, which would equal 3,916,000 X and 5,886,000 S.
40,000 7nm wafers at $10,000 is $400,000,000
If Microsoft spent exactly that much on 5nm wafers ($17K) they could only buy 23,529.
Let's say the same 55:45 allocation remains. That would be 12,940 X wafers and 10,588 S wafers.
Math above showed 271 X chips per 5nm wafer, and 501 S chips. I can only find 5nm defect rate before 2021 and it's reported 5nm was doing better than 7nm at a similar timeframe after production started, so at this point we could go with the same .09 defect rate. That would mean 246 X chips and 455 S chips usable per wafer.
246 * 12,940 = 3,183,240 X chips
455 * 10,588 = 4,817,540 S chips
If we look at the BD Drive costing anywhere between $18 and $33.50, that's a potential $70.5M and $131.2M in savings (from 3,916,000 X). Throwing that into wafer purchases would look like this.
$400,000,000 + $70,488,000 or $131,186,000 = $470,488,000 or $531,186,000
That's anywhere between 27,675 and 31,246 5nm wafers.
Let's say the same 55:45 allocation remains. That would be 15,221 to 17,284 X wafers and 12,454 to 14,142 S wafers.
Math above showed 246 usable X chips per 5nm wafer, and 455 S chips.
246 * 15,221 to 17,284 = 3,744,366 to 4,251,864 X chips
455 * 12,454 to 14,142 = 5,666,570 to 6,434,610 S chips
So reinvesting the cost saved from just the BD drive could bring Xbox to producing slightly more units by about 9%. With overall cost reduction from less plastic and simpler Series X internals, reinvesting that back into buying more wafers could see an additional 20% or higher units produced. Maybe.
If Microsoft wanted the same number of wafers on 5nm as 7nm (40,000) that would cost $680,000,000.
Again, 55:45 allocation
246 * 22,000 = 5,412,000 X chips
455 * 18,000 = 8,190,000 S chips
I think a better use of wafers would be 25K and 15K each. I don't think Microsoft is going to need 8M S.
246 * 25,000 = 6,150,000 X chips
455 * 15,000 = 6,825,000 S chips
This post is long winded since I just used it to try some math out that might not even be accurate but eh. A move to 5nm allows Microsoft to basically build a new Series X which is going to have a lot of cost reductions thanks to the smaller SOC and if they go digital only. I think it's worth it, otherwise I don't think they can match Series X demand this holiday. Microsoft are always looking into how to reduce costs of the console over time and I don't know what reason they would have to not go down to 6nm already if that's what they are planning.
Also, just using the same math for PS5, PS5 7nm SOC was 308mm2. That's 229 potential chips per wafer, but 208 with the defect rate, or 16,640,000 producible PS5. PS5 specifically could have a higher fail rate than Xbox due to the overclocking required for the PS5 GPU.
On a 6nm SOC, PS5 has a 264mm2 chip or 267 chips per wafer, and 242 chips after defects. That would equal out to 19,360,000 possible PS5 with 80,000 wafers. If Sony secured 100,000 wafers, they could produce 24,200,000 PS5.
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.
www.gameswirtschaft.de
www.tomshardware.com
fuse.wikichip.org
but eh. A move to 5nm allows Microsoft to basically build a new Series X which is going to have a lot of cost reductions thanks to the smaller SOC and if they go digital only. I think it's worth it, otherwise I don't think they can match Series X demand this holiday. Microsoft are always looking into how to reduce costs of the console over time and I don't know what reason they would have to not go down to 6nm already if that's what they are planning.
direct.playstation.com
![[Bloomberg] Nintendo Switch 2 Set for Gaming’s Biggest Ever Launch Even at $400-Plus](/forums/data/news_thumbnails/0/90.jpg?1741939443)
