Friday, October 24th 2025

Capacity Constraints Hit Intel as Demand Outstrips Intel 10/7 Node Supply

Intel's Q3 earnings delivered a solid number, but a few interesting things are happening at Intel's Fabs that aren't related to any leading-edge node production. According to Intel CFO David Zinsner, "In Q3, Intel Foundry delivered Intel 10 and 7 volume above expectations," indicating that demand for Intel's products built on these nodes is significantly exceeding supply. Zinsner noted, "Capacity constraints, especially on Intel 10 and Intel 7, limited our ability to fully meet demand in Q3 for both data center and client products," suggesting strong demand for Core and Xeon processors built on these nodes.

Intel first introduced Intel 7 node, previously known as 10 nm Enhanced SuperFin, with 12th Generation "Alder Lake." However, that is now discontinued and Intel continues to ship 13th Generation "Raptor Lake" and 14th Generation Core "Raptor Lake Refresh" client CPUs on the Intel 7 node. For Xeon server chips, the company uses the Intel 7 node for the 4th Generation "Sapphire Rapids," 5th Generation "Emerald Rapids" Xeon Scalable and manufactures the I/O die for the latest "Granite Rapids" Xeon 6, where the compute dies are made on the Intel 3 node. High demand from the gaming sector for 13th and 14th Gen gaming CPUs, along with strong demand for Xeon processors, has created a manufacturing bottleneck. Intel will prioritize delivering data center products first and gaming products second.
Intel produces its Intel 7 node at several locations, including Intel D1X in Hillsboro, Oregon, and Fabs 12, 32, and 42 in Arizona. Despite having multiple fabrication sites, demand is consuming all available capacity. As supply tightens and demand rises, prices have increased. In September, Intel implemented a 10% price hike on its 13th and 14th Generation processors. Standard RPL-S SKUs, previously priced around $150-160, have risen to about $170-180 due to tight inventories and increased upstream costs for maintaining older and multiple product lines. Most capacity will be allocated to the data center product portfolio, led by Xeon processors, as it represents a higher-margin business and a crucial revenue source for Intel.
Source: Intel Q3 Earnings Transcript
Add your own comment

23 Comments on Capacity Constraints Hit Intel as Demand Outstrips Intel 10/7 Node Supply

#1
Mr. Perfect
Any idea what they're using the 10nm for? I tried finding a product that still uses it, but the last one seems to be Tiger Lake, which was discontinued in January 2024. Or is 10nm and 7nm just lumped together in their reporting?
Posted on Reply
#2
wNotyarD
Mr. PerfectAny idea what they're using the 10nm for? I tried finding a product that still uses it, but the last one seems to be Tiger Lake, which was discontinued in January 2024. Or is 10nm and 7nm just lumped together in their reporting?
Aren't the 7 and 10 processes also used for Altera FPGAs?
Posted on Reply
#3
Wirko
I understand that 10nm was renamed to Intel 7, but what node was renamed to Intel 10?
Posted on Reply
#4
AleksandarK
News Editor
WirkoI understand that 10nm was renamed to Intel 7, but what node was renamed to Intel 10?
Regular 10 nm SuperFin. The enhanced version was renamed to Intel 7
Posted on Reply
#5
Squared
10nm (Ice Lake)
10SF (Tiger Lake)
10ESF (renamed Intel 7, Alder Lake and Raptor Lake)
Posted on Reply
#6
dgianstefani
TPU Proofreader
Squared10nm (Ice Lake)
10SF (Tiger Lake)
10ESF (renamed Intel 7, Alder Lake and Raptor Lake)
Raptor apparently had further improvements.
Posted on Reply
#7
TheLostSwede
News Editor
Mr. PerfectAny idea what they're using the 10nm for? I tried finding a product that still uses it, but the last one seems to be Tiger Lake, which was discontinued in January 2024. Or is 10nm and 7nm just lumped together in their reporting?
Intel makes a lot of peripheral chips, like networking, which 10 nm is still good enough for, although some of that has moved to 7 nm. That said, anything that isn't recent and 10 Gbps, seems to be 28 nm still.
Posted on Reply
#8
Squared
dgianstefaniRaptor apparently had further improvements.
I always wondered but never saw this. Thanks!
Posted on Reply
#9
wNotyarD
TheLostSwedeIntel makes a lot of peripheral chips, like networking, which 10 nm is still good enough for, although some of that has moved to 7 nm. That said, anything that isn't 10 Gbps, seems to be 28 nm still.
Do you have info on what process is used to manufacture their chipsets?
Posted on Reply
#10
TheLostSwede
News Editor
wNotyarDDo you have info on what process is used to manufacture their chipsets?
Unfortunately no.

I found this though. Can't say I remember this having been announced.

Posted on Reply
#11
Squared
TheLostSwedeIntel makes a lot of peripheral chips, like networking, which 10 nm is still good enough for, although some of that has moved to 7 nm.
Is that where all of Intel's fab capacity is going? Just looking at client and server CPUs fab allocation doesn't make sense to me. Intel used to produce Apple laptop chips and most chips for the laptop, server, and desktop markets and they would all begin using a new node within a few months.

Today all server tiles are made by Intel but on the client side Intel makes only one of four Meteor Lake active tiles, zero active Lunar Lake and Arrow Lake tiles, and one in three or two in three (depending on SKU) Panther Lake tiles.

Also Intel 7 demand remains strong somehow and four fabs can't keep up, and to my knowledge after years of collecting EUV machines for 4/3 and 18A, only Fab 34 in Ireland is producing 4/3 and Fab 52 in Arizona is alone producing 18A. Doesn't Intel have around 21 fabs? Where is the capacity for modern nodes?
Posted on Reply
#12
TheLostSwede
News Editor
SquaredIs that where all of Intel's fab capacity is going? Just looking at client and server CPUs fab allocation doesn't make sense to me. Intel used to produce Apple laptop chips and most chips for the laptop, server, and desktop markets and they would all begin using a new node within a few months.
I doubt it, but for for some of those nodes, it might be, as the demand for networking chips are at an all time high, since all the LLM training requires the servers to talk to each other, no?
SquaredAlso Intel 7 demand remains strong somehow and four fabs can't keep up, and to my knowledge after years of collecting EUV machines for 4/3 and 18A, only Fab 34 in Ireland is producing 4/3 and Fab 52 in Arizona is alone producing 18A. Doesn't Intel have around 21 fabs? Where is the capacity for modern nodes?
Only 10 that are used for mass production of chips, with two more coming in Arizona and then that one that may or may not be built in Ohio. Another seven seem to be for R&D of some kind and that last one, I'm not sure. Plus one in Penang that does packaging.
en.wikipedia.org/wiki/List_of_Intel_manufacturing_sites
Posted on Reply
#13
londiste
Didn't Intel go for an aggressive move with fabs to latest node? Basically, 10/7 is likely capacity constrained because there isn't that much of a capacity.
Posted on Reply
#14
user556
londisteDidn't Intel go for an aggressive move with fabs to latest node? Basically, 10/7 is likely capacity constrained because there isn't that much of a capacity.
It still suggests that demand for those older products has increased recently. Which I find extraordinary given the over-voltage problem was never resolved. Those gen 13/14 products are time-bombs.
Posted on Reply
#15
thestryker6
wNotyarDDo you have info on what process is used to manufacture their chipsets?
Z690/790 and the rest of those generations used 14nm. Z890 has similar features, but a different package and I haven't seen any confirmations on manufacturing node though 14nm seems likely. Previously they used 22nm and potentially still use it (or the revamped version Intel 16) for lower cost parts that don't need the density.
londisteDidn't Intel go for an aggressive move with fabs to latest node? Basically, 10/7 is likely capacity constrained because there isn't that much of a capacity.
None of the EUV manufacturing capacity has been taken from the DUV nodes. They did not increase DUV capacity as originally planned under Gelsinger though and instead swapped some machines early. The current situation is due to OEMs wanting cheaper parts and buying stuff using Intel 7 as opposed to Intel 3/TSMC N3B for that reason.
Posted on Reply
#16
londiste
thestryker6None of the EUV manufacturing capacity has been taken from the DUV nodes. They did not increase DUV capacity as originally planned under Gelsinger though and instead swapped some machines early.
Didn't 10/7 also become EUV in their later/current iterations?
user556Which I find extraordinary given the over-voltage problem was never resolved. Those gen 13/14 products are time-bombs.
While Raptor Lake had a problem, why would that be a problem with manufacturing? Other than specific problematic batches the problems with 13/14 gen the problems are/were about settings rather than dies themselves.
Posted on Reply
#17
thestryker6
londisteDidn't 10/7 also become EUV in their later/current iterations?
Nope they just made changes to the node design and I believe the biggest shift here was their FinFET design.
Posted on Reply
#18
user556
londisteWhile Raptor Lake had a problem, why would that be a problem with manufacturing? Other than specific problematic batches the problems with 13/14 gen the problems are/were about settings rather than dies themselves.
It's a product design problem, not the manufacturing nor specific batches. If Intel doesn't know the reason for the flaw then they can't fix it. Supposedly a micro-code update would likely fix it (provide a workaround) but until the reason is solved they're all still ticking time bombs that irreversibly destroys the CPUs - In a piecemeal fashion from what I've read. The don't completely die outright.
Posted on Reply
#19
londiste
user556It's a product design problem, not the manufacturing nor specific batches. If Intel doesn't know the reason for the flaw then they can't fix it. Supposedly a micro-code update would likely fix it (provide a workaround) but until the reason is solved they're all still ticking time bombs that irreversibly destroys the CPUs - In a piecemeal fashion from what I've read. The don't completely die outright.
Intel had at least 3 different problems based on what has been discovered:
1. Bad batches with an oxidation problem, which is a manufacturing issue but limited, and now they know about it.
2. Overvoltage due to the microcode problem(s) with Vmin shift.
3. The motherboard/CPU settings resulting in excessive current and voltage.

The point I was trying to make was that other than oxidation issue, it wasn't a problem with manufacturing process.
thestryker6Nope they just made changes to the node design and I believe the biggest shift here was their FinFET design.
FinFET itself came a lot earlier, I believe it even started at 22nm but obviously evolved further throughout nodes.
Intel did a FinFET update both in 10nm Superfin as well as its successor which became Intel 7. However, what I had in mind was that at least in Intel 7 they were doing (limited) EUV layers. I do not know if it ever got confirmed but the idea going around was that they did move to EUV specifically for some critical layers to finally avoid quad patterning.

You are right in that me saying it became EUV - implying fully EUV - isn't strictly true but they did go for some of it.
Posted on Reply
#20
thestryker6
londisteFinFET itself came a lot earlier, I believe it even started at 22nm but obviously evolved further throughout nodes.
Hence why I referred to changing the design as opposed to introducing something new...
londisteHowever, what I had in mind was that at least in Intel 7 they were doing (limited) EUV layers. I do not know if it ever got confirmed but the idea going around was that they did move to EUV specifically for some critical layers to finally avoid quad patterning.

You are right in that me saying it became EUV - implying fully EUV - isn't strictly true but they did go for some of it.
Intel did no such thing with any of their 10nm nodes however TSMC did with N7. The iteration of Intel 7 used with RPL has "3rd generation" SuperFin transistors, but no shift to using EUV for any of the layers.
Posted on Reply
#21
user556
londisteIntel had at least 3 different problems based on what has been discovered:
1. Bad batches with an oxidation problem, which is a manufacturing issue but limited, and now they know about it.
2. Overvoltage due to the microcode problem(s) with Vmin shift.
3. The motherboard/CPU settings resulting in excessive current and voltage.

The point I was trying to make was that other than oxidation issue, it wasn't a problem with manufacturing process.
And the point I'm making is #2 is unknown cause, unsolved, still a time bomb. Intel attempted something like three different patches without success.
Posted on Reply
#22
MrMilli
user556And the point I'm making is #2 is unknown cause, unsolved, still a time bomb. Intel attempted something like three different patches without success.
You are blowing this way out of proportion. I have zero returns out of hundreds 12 to 14th gen sold systems. My distributor has minimal returns and they sell tens of thousands.
Let me add that I've never sold a system with PL1/PL2 set to 4096W. That was a stupid thing that motherboard manufacturers never should have done.
There were reports that 13 and 14th had a return rate four times as high as other CPU's but considering that CPU's have a return rate of under 1%, that still is a fraction of the market. Not something as you describe as "a ticking timebomb".
Posted on Reply
#23
user556
Good to hear, I'll be staying clear though.
Posted on Reply
Add your own comment
Nov 20th, 2026 16:19 PST change timezone

New Forum Posts

Popular Reviews

TPU on YouTube

Controversial News Posts