News Intel 2Q25 Earnings

[Doug S]

The opposite. People avoid comparable AMD laptops because they often come with MTK junk.

And spinning it off from Intel is tolerable just don't say it has no value. Spun off they'll be a good acquisition target too.

I didn't mean "no value" to imply no one would be interested in buying it. It will be a good acquisition target as you say. I meant "no value" as it isn't relevant to whether Intel makes it through this crisis successfully or not. Everything that doesn't contribute value toward Intel turning around their x86 design issues and getting their foundry off the ground needs to go - especially the parts that are "good acquisition targets" because that means they will get some much needed cash out of it.

 

[jpiniero]

Is anyone choosing Intel laptops because they come with Intel wireless? No? Then it isn't a reason for them to keep that as part of the company!

If it's profitable and fills the fabs... why not.

 

[Doug S]

If it's profitable and fills the fabs... why not.

It isn't filling leading edge fabs. They need stuff that's made on 18A and 14A, not stuff made on old nodes and where part of it is made on an analog process. Besides nothing stops them from selling that business and having part of the deal "you make your stuff exclusively with Intel Foundry for the next five years". The buyer would kind of have to do it for the next couple at least anyway between the current products and products that are already far enough along in development that it would cost too much to 'port' them to TSMC or Samsung.

 

[oak8292]

It isn't filling leading edge fabs. They need stuff that's made on 18A and 14A, not stuff made on old nodes and where part of it is made on an analog process. Besides nothing stops them from selling that business and having part of the deal "you make your stuff exclusively with Intel Foundry for the next five years". The buyer would kind of have to do it for the next couple at least anyway between the current products and products that are already far enough along in development that it would cost too much to 'port' them to TSMC or Samsung.

I think Intel needs to fill trailing edge as well. This has been a competitive disadvantage for Intel and it is one of their biggest issues.

I think Apple went to TSMC because TSMC had the eco-system to fill trailing edge capacity economically which reduces their costs. Samsung and Apple were competing for leading edge and it is a lot more expensive to move to the next node if there aren’t people backfilling that capacity.

Over the last ten years Intel has gone from depreciating equipment in 4 years, to briefly for five years and now for eight years. Equipment depreciation is the biggest expense for leading edge. Intel needs trailing customers just to complete depreciating the equipment.

If Apple moves to Intel leading edge and Intel doesn’t have any trailing customers it just worsens their equipment cost problems. Intel needs 18A secondary customers so Intel can transition to 14A.

Apple may be looking at 14A capacity to see if there is a way to use it for five years like they do with TSMC nodes. Apple used Intel build modems with Intel’s Infineon technology that they now own. That was supposed to be trailing capacity, 14 nm but it didn’t work that way.

 

[adroc_thurston]

see if there is a way to use it for five years like they do with TSMC nodes.

Intel's trailing age user, is, well, Intel.
They sell N-1 parts for a while.

 

[oak8292]

Intel's trailing age user, is, well, Intel.
They sell N-1 parts for a while.

They have been stuck on nodes for years with +,++,+++ … which is one reason they went to five years of depreciation. However the demand should be tailing off in later years and there is still underutilized equipment. I think in the past they moved as much equipment as they could to the next node, and sold depreciated equipment into the secondary market.

However, in the last year and a half Intel is taking charges on equipment that is being taken out of service. This is what is really making foundry look horrible and unsalvageable with 50-70% losses. They don’t have the trailing demand to utilize it. I don’t know what is happening with UMC and 12 nm but Intel is writing off equipment.

 

[Doug S]

I think Intel needs to fill trailing edge as well. This has been a competitive disadvantage for Intel and it is one of their biggest issues.

Filling trailing edge isn't a problem. Or rather isn't a technical problem of "can they get good yields" like potential customers wonder about leading edge nodes. It isn't a problem of customer interest since a fab in the US that's not subject to the kind of supply chain issues that crippled automakers and many others during covid has a large built in customer base. What it is is a business problem, as they need to do the "we are a foundry and we provide all the services other foundries do for their customers" stuff. The same stuff that's required to get leading edge customers too.

If the only way they're able to fill those trailing edge foundries is with their own stuff that indicates they aren't doing the basic foundry business stuff well enough for less demanding trailing edge customers - which would imply they have zero chance of winning any business to their leading edge because those customers are VERY demanding considering how much they're paying.

Trailing edge customers won't make headlines so who knows if they're getting any but if they aren't I see little chance of them getting any leading edge business even if 14A beats the pants off TSMC.

 

[oak8292]

Filling trailing edge isn't a problem. Or rather isn't a technical problem of "can they get good yields" like potential customers wonder about leading edge nodes. It isn't a problem of customer interest since a fab in the US that's not subject to the kind of supply chain issues that crippled automakers and many others during covid has a large built in customer base. What it is is a business problem, as they need to do the "we are a foundry and we provide all the services other foundries do for their customers" stuff. The same stuff that's required to get leading edge customers too.

If the only way they're able to fill those trailing edge foundries is with their own stuff that indicates they aren't doing the basic foundry business stuff well enough for less demanding trailing edge customers - which would imply they have zero chance of winning any business to their leading edge because those customers are VERY demanding considering how much they're paying.

Trailing edge customers won't make headlines so who knows if they're getting any but if they aren't I see little chance of them getting any leading edge business even if 14A beats the pants off TSMC.

Agree with most of this but I will add that TSMC variants on initial node are a continuous ‘improvement’ process for trailing customers. Intel is working on 12 nm with UMC, this is not an Intel node, this is a variant that is hopefully fully PDK support, easy to use and yields well on Intel equipment.

Edit to remove customization example

 

[Jan Olšan]

I was saying a couple years ago that Intel needed to dump Altera, dump networking, dump everything that wasn't part of their sales of x86 PC & server SKUs and the foundry. Under Gelsinger they were pussy footing around with this stuff, like he was afraid of upsetting senior managers he'd known for years. No more pussy footing around, time to cut loose from all the barnacles that have attached themselves to Intel when they were overly fat and happy.

They've been in networking for a LONG time, I remember Intel NICs in the 90s, Centrino wireless solutions in the 00s so it has been everywhere not just the HPC level stuff. Problem is, other than at the really high end that's all commodity. Why does Intel still design and sell wifi chips? OK yes the answer is obvious - back when they were a near monopoly they wanted to extract more money out of every PC sale so they tied sales of Intel mobile CPUs to Intel wireless chips. Maybe it even made sense beyond that back then since there probably were some fairly crappy wireless chips.

But again this is all commodity stuff today. No one is creating any value add with wifi chips anymore. No one even looks at what wifi version a device supports, they know whatever it is its more than "good enough".

Why do people keep putting out these opinions without actually knowing if the business was profitable? If it was, there is no reason to just dump it because it is not cool and muh small lean company. It may be something they are forced to shed to get cash for other stuff but it harms the company. Examples: AMD forced to shed their packaging plant or license Zen to China in the last few years before turnaround. Got few hundreds million $, lost more in the long term.

Or do you know the numbers?

 

[madtronik]

I think Intel needs to fill trailing edge as well. This has been a competitive disadvantage for Intel and it is one of their biggest issues.

I think Apple went to TSMC because TSMC had the eco-system to fill trailing edge capacity economically which reduces their costs. Samsung and Apple were competing for leading edge and it is a lot more expensive to move to the next node if there aren’t people backfilling that capacity.

Over the last ten years Intel has gone from depreciating equipment in 4 years, to briefly for five years and now for eight years. Equipment depreciation is the biggest expense for leading edge. Intel needs trailing customers just to complete depreciating the equipment.

If Apple moves to Intel leading edge and Intel doesn’t have any trailing customers it just worsens their equipment cost problems. Intel needs 18A secondary customers so Intel can transition to 14A.

Apple may be looking at 14A capacity to see if there is a way to use it for five years like they do with TSMC nodes. Apple used Intel build modems with Intel’s Infineon technology that they now own. That was supposed to be trailing capacity, 14 nm but it didn’t work that way.

But do they need to fill it? If I remember correctly, even in the Bob Swan era Intel already had a lot of trailing edge work outsourced to TSMC because it was CHEAPER that doing it in-house. Until IDM 2.0 they didn't expect to reuse old nodes to offer trailing edge as a foundry. And their current old nodes are not ready to be offered as an independent foundry because they are intel-custom. So, nothing to see there? Only 18A and 14A seem relevant to the foundry business.

 

[NTMBK]

The opposite. People avoid comparable AMD laptops because they often come with MTK junk.

And spinning it off from Intel is tolerable just don't say it has no value. Spun off they'll be a good acquisition target too.

How many laptop sales do you think AMD loses due to those Mediatek WiFi chips? I sincerely doubt it's more than double digits per year. 99.9% of laptop users have absolutely no idea who makes their WiFi chip.

 

[gdansk]

How many laptop sales do you think AMD loses due to those Mediatek WiFi chips? I sincerely doubt it's more than double digits per year. 99.9% of laptop users have absolutely no idea who makes their WiFi chip.

One can't expect people to know. But it's part of the unquantifiable, nebulous AMD jank people complain about.

 

[Joe NYC]

If the only way they're able to fill those trailing edge foundries is with their own stuff that indicates they aren't doing the basic foundry business stuff well enough for less demanding trailing edge customers - which would imply they have zero chance of winning any business to their leading edge because those customers are VERY demanding considering how much they're paying.

Exactly.

In a sense, Intel recognized that Intel organization can't do basic foundry business. Which is why Intel tried to acquire Tower (after missing out on Global Foundries).

IMO, Intel dropped the ball in not acquiring Global Foundries. (Pat was asleep). The price was not even that high. All of the GloFo capacity in trailing edge and custom silicon + more advanced Intel node would have strengthened the combined range of offerings, and it would have paved the pay to perhaps IPO the foundry side under GloFo leadership.

And heavy purchase commitments from Intel, same as Global Foundries got from AMD.

It would have made it easier for this new GloFo ask for US taxpayer support, that would not be indirectly propping up Intel Products division.

Being independent of Intel would also open the door to more customers, including those who were already burned by Intel management meddling (Qualcomm)

 

[Joe NYC]

Why do people keep putting out these opinions without actually knowing if the business was profitable? If it was, there is no reason to just dump it because it is not cool and muh small lean company. It may be something they are forced to shed to get cash for other stuff but it harms the company. Examples: AMD forced to shed their packaging plant or license Zen to China in the last few years before turnaround. Got few hundreds million $, lost more in the long term.

Or do you know the numbers?

I don't have numbers.

Hallucination (of someone on Twitter) is the best I can do:

https://twitter.com/x/status/1948836183703769550

 

[DavidC1]

Why do people keep putting out these opinions without actually knowing if the business was profitable? If it was, there is no reason to just dump it because it is not cool and muh small lean company. It may be something they are forced to shed to get cash for other stuff but it harms the company.

Shedding Intel Wireless is more complicated than just selling it. Their WiFi is partly integrated into the SoC tile, which reduces overall cost(BOM). That's what the CNVio interface is for, and why they have two variants.

If they don't sell it, eventually it would be fully integrated into the SoC.

Selling such divisions may not be an immediately bad thing in the short term, but it's basically death by many small cuts. It won't help either. If they need to sell it, then it's an extremely desperate move, where you lop off your fingers to save yourself.

Intel is basically the drivers of the PC ecosystem. One of the handheld vendors when asked why they used Intel said it's because they provided very good engineering and marketing support. What AMD does is not even comparable to what Intel does.

They pretty much go as far as essentially creating an entire system for you and you just have to rebrand it and provide support which resulted in few companies being created that way. Now, tell me that doesn't have a positive impact on the market?

There are parts in a business that isn't directly quantifiable. Such as the argument iGPUs are a waste of die space, yet if they took out iGPUs, many people will simply not buy it, thus the loss is 100%.

 

[DavidC1]

Intel after 2010-11 has been the most clueless company they bought so many AI vendors / Altera / even got back into memory Business McAfee anyone it's a miracle sometimes how can someone survive after this much stupidity

The whole PC market was basically Intel, so it takes a long time to fall.

In a sane place, the Network division absolutely should not be sold. There are many reasons people get a product. You are taking those reasons away one by one. And you are opening up avenues for competitors to come in. Lip Bu Tan's future Intel better deliver absolutely stunning CPUs to make up for such losses if more cuts are to come.

One of the reasons I liked their network products is because they have the best drivers and support. Actually outside of graphics which require extreme levels of software support, Intel has great drivers and support.

I know if I configure with a Haswell or later laptop, to enable deep sleep states such as C6/C7/C8/C10, the two greatest enablers to making that happen is a right SSD and WiFi. In WiFi for example Intel WiFi is an easy product to choose to make that happen. When they used to make SSDs, their SSDs were one of the very few that actually allowed some battery life improvements over HDDs.

If the Network division is gone, it'll be one that'll be missed.

 

[Thunder 57]

The whole PC market was basically Intel, so it takes a long time to fall.

In a sane place, the Network division absolutely should not be sold. There are many reasons people get a product. You are taking those reasons away one by one. And you are opening up avenues for competitors to come in. Lip Bu Tan's future Intel better deliver absolutely stunning CPUs to make up for such losses if more cuts are to come.

One of the reasons I liked their network products is because they have the best drivers and support. Actually outside of graphics which require extreme levels of software support, Intel has great drivers and support.

I know if I configure with a Haswell or later laptop, to enable deep sleep states such as C6/C7/C8/C10, the two greatest enablers to making that happen is a right SSD and WiFi. In WiFi for example Intel WiFi is an easy product to choose to make that happen. When they used to make SSDs, their SSDs were one of the very few that actually allowed some battery life improvements over HDDs.

If the Network division is gone, it'll be one that'll be missed.

It's almost as if Lip Bu Tan has never heard of "Centrino".

 

[DrMrLordX]

If NEX is sold/spun off, it'll probably result in a company with an IP-sharing deal with Intel for the foreseeable future, enabling Intel to continue cranking out solid platforms (to the extent that they're still able).

 

[madtronik]

The whole PC market was basically Intel, so it takes a long time to fall.

In a sane place, the Network division absolutely should not be sold. There are many reasons people get a product. You are taking those reasons away one by one. And you are opening up avenues for competitors to come in. Lip Bu Tan's future Intel better deliver absolutely stunning CPUs to make up for such losses if more cuts are to come.

One of the reasons I liked their network products is because they have the best drivers and support. Actually outside of graphics which require extreme levels of software support, Intel has great drivers and support.

I know if I configure with a Haswell or later laptop, to enable deep sleep states such as C6/C7/C8/C10, the two greatest enablers to making that happen is a right SSD and WiFi. In WiFi for example Intel WiFi is an easy product to choose to make that happen. When they used to make SSDs, their SSDs were one of the very few that actually allowed some battery life improvements over HDDs.

If the Network division is gone, it'll be one that'll be missed.

I have an all AMD system with an integrated Intel ethernet adapter. For me, that the adapter is Intel is a nice plus.

 

[MasterofZen]

That's one ugly result.

 

[511]

More to come next quarter is PTL Ramp that is going to be costly along with LNL they are maintaining share by using Pricing on RPL.

 

[Joe NYC]

More to come next quarter is PTL Ramp that is going to be costly along with LNL they are maintaining share by using Pricing on RPL.

Intel is forecasting margins that are quite low. But we will have to see AMD numbers. If AMD margins are good and/or going up, then Intel margins will twice as bad...

 

[511]

Intel is forecasting margins that are quite low. But we will have to see AMD numbers. If AMD margins are good and/or going up, then Intel margins will twice as bad...

AMD doesn't have manufacturing and a complicated business to do Intel Advanced Account™.