Tuesday, February 21st 2023

Intel Defers 3 nm Wafer Orders with TSMC, Pushes "Arrow Lake" Rollout to 2025?

Intel has reportedly deferred its orders for 3 nm wafers with TSMC, sources in PC makers tell Taiwan-based industry observer DigiTimes. Built on the TSMC N3 node, the wafers were supposed to power the Graphics tiles (containing the iGPU), of the upcoming "Arrow Lake" processors, which were originally on course for a 2024 release. The DigiTimes report detailing this development says that Intel's 3 nm wafer orders have been deferred to Q4-2024, which would realistically mean a 2025 launch for whatever product was designed to use 3 nm tiles. Advance orders for next-gen wafers by high-volume clients such as Intel, are usually placed several quarters in advance, so the foundry could suitably scale up its capacity.
Source: DigiTimes
Add your own comment

32 Comments on Intel Defers 3 nm Wafer Orders with TSMC, Pushes "Arrow Lake" Rollout to 2025?

#1
Daven
Shouldn’t this also effect Meteor Lake?
Posted on Reply
#2
Calmmo
DavenShouldn’t this also effect Meteor Lake?
It has been rumored for a while now that this might happen and that the next cpu's might be on the same node / just a raptor lake refresh (and potentially on the same platform as well..) (i will also admit that with all the "lakes" im at a loss as to which is lake is meant to be which..) So i guess they're just delaying everything
Posted on Reply
#3
usiname
Intel and delay, who would have thought...
Posted on Reply
#4
TheinsanegamerN
Oh great. This means yet another generation of the same xe igpu, and another year until AMD makes a truly powerful APU in response. The 680m just isn’t enough for me.
Posted on Reply
#5
bonehead123
We're drowning in the lakes intel, enough is enough, just name the new gen something else & be done with all the water bodies already !
Posted on Reply
#6
Denver
Happy buying AMD stock. :P
Posted on Reply
#7
Chaitanya
Even when Intel goes begging to 3rd party Fabs they have delays.
Also whatever happened to billions of taxpayers Dollars they leeched recently as subsidy.
Posted on Reply
#8
thesmokingman
usinameIntel and delay, who would have thought...
This is the way...
Posted on Reply
#9
armit
AMD's stability is very seriously unstable, and

Intel's electricity and heat are too serious, the same as high-performance graphics cards.

"A person who bought two Ryzen processors and two Intel 13th generation processors posted this post."
Posted on Reply
#10
trsttte
DenverHappy buying AMD stock. :p
Intel stock pays better dividends though ;)
Posted on Reply
#11
Minus Infinity
Ha ha, people called me a troll when I said a few months ago 2024 Arrow Lake was BS. No surprises at all. Gosh Zen 5 might on a refresh node by the time AL comes out. Too bad it's delayed though, I have high hopes for Arrow Lake's efficiency gains, let alone the new Meteor Lake and Arrow Lake P cores getting huge IPC uplifts.
Posted on Reply
#12
Cooe
TheinsanegamerNOh great. This means yet another generation of the same xe igpu, and another year until AMD makes a truly powerful APU in response. The 680m just isn’t enough for me.
... Uhh... Did you just forget the current 780m exists in AMD's brand new Phoenix APU or something? It's already like +30% faster than the 680m and prolly more like >=40% when overclocked (as RDNA 3 overclocks like a bat out of hell).
armitAMD's stability is very seriously unstable, and

Intel's electricity and heat are too serious, the same as high-performance graphics cards.

"A person who bought two Ryzen processors and two Intel 13th generation processors posted this post."
"Stability is unstable"? Uhh... wat??? o_O I'm calling bullshit on that one. Something was wrong with your specific PC config or your specific motherboard. Aside from X570's USB wonkiness that's since been fixed Ryzen hasn't been properly "unstable" since LITERALLY the early Zen 1 days with its super BIOS unoptimized and then barely functioning memory controller.
Posted on Reply
#13
BoboOOZ
Minus InfinityHa ha, people called me a troll when I said a few months ago 2024 Arrow Lake was BS. No surprises at all. Gosh Zen 5 might on a refresh node by the time AL comes out. Too bad it's delayed though, I have high hopes for Arrow Lake's efficiency gains, let alone the new Meteor Lake and Arrow Lake P cores getting huge IPC uplifts.
Unfortunately, it seems that Intel cannot seem to deliver on time no matter what, and that is having consequences on the whole market, unfortunately, regardless if you're buying blue or red.
Posted on Reply
#16
Wirko
bonehead123We're drowning in the lakes intel, enough is enough, just name the new gen something else & be done with all the water bodies already !
America will run out of lakes before Intel ceases making Lakes but wait, that won't be the end of the world just yet...

Apparently there are river rapids there, too.
Posted on Reply
#18
Hardware Geek
I just expect everything to be delayed from Intel at this point. The need to reevaluate how they are creating their roadmaps. They have to stop making promises they can't keep because delays have become their SOP.
Posted on Reply
#19
hs4
Other articles indicate that there was a problem with production capacity on TSMC's side, and Intel seems to be on the receiving end of the delay request.

"Apple has sharply reduced the size of the order it has with TSMC". PhoneArena. Feb 18, 2023.
The report says that in order to give Apple most of the 3nm production capacity later this year, Intel agreed to change its roadmap and delay receipt of its 3nm orders.
"Apple feels the heat – significantly reduces its advanced chip order with TSMC". Gizchina. Feb 18, 2023
However, the mass-produced N3 in the fourth quarter of this year do not have enough production capacity. Intel had to readjust and revised its roadmap so that the current main customer of the 3nm process is Apple.
These articles suggest that Intel will make the N5 iGPU for Arrow lake. Rather, AMD should be using N3 in Zen5, but is heading in a direction that will make that more difficult.
Posted on Reply
#20
TheinsanegamerN
Cooe... Uhh... Did you just forget the current 780m exists in AMD's brand new Phoenix APU or something? It's already like +30% faster than the 680m and prolly more like >=40% when overclocked (as RDNA 3 overclocks like a bat out of hell).
....Uhh.....no. As of yet there is not a single product on the market with said SOC. All the claims of it being "30%" faster are based on time spy, which is a synthetic benchmark well known for not tracking alongside games.

Given AMD's decoder ring BS and how hard they are pushing zen 2/3 design wins; I'm going to guess that said phoenix point products will not be in large supply for some time. Then we have to wait for industry designs, and then hope they are made available. They also didnt clarify what memory speed the 680m and 780m were running at, if the 680m was using 4800 or 5200 mhz DDR5 and the 780m was using 7500 mhz LPDDR5x, that would make a massive difference. Based on the 7900 series I dont see these new iGPUs being 30-40% faster in IRL tasks. Even so, the only reason we got the 680m was intel's xe GPUs, and we never saw it on desktop. To get another big iGPU push well need intel to get off of their hind quarters and release that xe2-LGP based 192eu iGPU that meteor lake was supposed to get.
Posted on Reply
#21
BoboOOZ
TheinsanegamerN....Uhh.....no. As of yet there is not a single product on the market with said SOC. All the claims of it being "30%" faster are based on time spy, which is a synthetic benchmark well known for not tracking alongside games.

Given AMD's decoder ring BS and how hard they are pushing zen 2/3 design wins; I'm going to guess that said phoenix point products will not be in large supply for some time. Then we have to wait for industry designs, and then hope they are made available. They also didnt clarify what memory speed the 680m and 780m were running at, if the 680m was using 4800 or 5200 mhz DDR5 and the 780m was using 7500 mhz LPDDR5x, that would make a massive difference. Based on the 7900 series I dont see these new iGPUs being 30-40% faster in IRL tasks. Even so, the only reason we got the 680m was intel's xe GPUs, and we never saw it on desktop. To get another big iGPU push well need intel to get off of their hind quarters and release that xe2-LGP based 192eu iGPU that meteor lake was supposed to get.
The problem with AMD laptop APU's is not performance, it's very low availability.
The demand largerly surpasses supply for newer AMD laptop chips. Most probably Phoenix will have similar availability to the 6800H/ U series, which is very small. Looks like AMD is playing it safe and allocating most capacity to taking market share in server from Inteal which means that great products for great prices will have to wait in the laptop market.
Indeed, Intel is much more likely to have good laptop APUs with good availability at this point. Given the number of announcements, we can guess they are fighting hard for the laptop market and will be bringing the more interesting stuff.
Posted on Reply
#22
hs4
BoboOOZThe problem with AMD laptop APU's is not performance, it's very low availability.
The demand largerly surpasses supply for newer AMD laptop chips. Most probably Phoenix will have similar availability to the 6800H/ U series, which is very small. Looks like AMD is playing it safe and allocating most capacity to taking market share in server from Inteal which means that great products for great prices will have to wait in the laptop market.
Indeed, Intel is much more likely to have good laptop APUs with good availability at this point. Given the number of announcements, we can guess they are fighting hard for the laptop market and will be bringing the more interesting stuff.
Zen3: Server - N7, Laptop - N6
Zen4: Server - N5, Laptop - N4
Zen5: Server - N4, Laptop - N3 (according to roadmap slides)

AMD has been allocating better nodes for laptop separately from those for servers. The opinion that laptop supply is low due to prioritizing server use will not be true.
Posted on Reply
#23
BoboOOZ
hs4Zen3: Server - N7, Laptop - N6
Zen4: Server - N5, Laptop - N4
Zen5: Server - N4, Laptop - N3 (according to roadmap slides)

AMD has been allocating better nodes for laptop separately from those for servers. The opinion that laptop supply is low due to prioritizing server use will not be true.
You do understand the difference between node size and capacity?
Posted on Reply
#24
Zareek
Cooe"Stability is unstable"? Uhh... wat??? o_O I'm calling bullshit on that one. Something was wrong with your specific PC config or your specific motherboard. Aside from X570's USB wonkiness that's since been fixed Ryzen hasn't been properly "unstable" since LITERALLY the early Zen 1 days with its super BIOS unoptimized and then barely functioning memory controller.
Right, seriously! I've been on Ryzen since the first series. My old 1800x is still in service, but it's definitely getting long in the tooth and it is still seriously memory picky. My 3800x/x570 had a few minor issues as I was a launch day buyer. They got sorted pretty quickly, and it has been smooth sailing since. If anything, my current machine is more stable than any platform I have ever owned, Intel or AMD based!
Posted on Reply
#25
TheinsanegamerN
hs4Zen3: Server - N7, Laptop - N6
Zen4: Server - N5, Laptop - N4
Zen5: Server - N4, Laptop - N3 (according to roadmap slides)

AMD has been allocating better nodes for laptop separately from those for servers. The opinion that laptop supply is low due to prioritizing server use will not be true.
So AMD is using superior nodes with low availability for their newest mobile parts. Thus, availability of these parts would be......low!

If AMD had tons of room on their fabs for Zen 4 CPUs, do you rally think we'd see 90% of the 7000 series lineup being recycled older gen parts?
Posted on Reply
Add your own comment
May 21st, 2024 16:51 EDT change timezone

New Forum Posts

Popular Reviews

Controversial News Posts