News Intel spins off Network and Edge group as standalone business after posting $2.9b loss — seeking investors for NEX division amid massive internal r...

[Admin]

Intel will spin off its Network and Edge Group (NEX) into a standalone company, seeking external investors while keeping a minority stake. The move follows Intel’s $2.9B quarterly loss, a 15% workforce cut, and warnings that its 14A node could be canceled without a major external partner.

Intel spins off Network and Edge group as standalone business after posting $2.9b loss — seeking investors for NEX division amid massive internal r... : Read more

 

[JRStern]

Been seeking a buyer for how long now, a year or so?

 

[jg.millirem]

With every article about Intel I read now, Taps is playing in my mind.

 

[hotaru251]

honesty Intel's spiral is not friendly for investors ESPECIALLY after the statement about leaving cutting edge nodes as a possibility.

Why would you invest in a company who is in such dire straights they are gutting their company & even AFTER that is in such a bad place it risks them leaving an entire market?

That is insanely high risk all together. You could easily invest in Nvidia or amd and get profit w/o any of that risk.

 

[acadia11]

Ok,the head of their networking and edge computing division was given the reigns of their AI plans … what does this mean than for the the executive shake up? Can’t remember the guys name … but does he go with the new NEX company?

 

[razor512]

If that section of their business includes their PCIe NICs and WiFi adapters, then part of Intel's business issue is really bad marketing.

For example, they the nothing to make it easier for people to get proper BE200 WiFi adapters, and they do not put enough pressure on stores to properly label the requirements and limitations of their XX1adapters and the CNVio requirement.

They need to be more clear about their networking hardware and market them at least as much as they market their video cards.

They need to push for lower cost 10GbE NICs and use their connections to get them on store shelves of every major electronics store, same for their WiFi adapters.

They need to push for multiple home user tiers of networking hardware. For example, Ease up on the price gouging and push some of their QSFP28 NICS and some Tofino 2 based switches as a higher end solution for home wired networking.

They came out in 2018 and that is more than enough time for them to drop the insane markups and target home users. It will make them stand out among the home networking market by offering something unique that can improve the user experience.

 

[TerryLaze]

honesty Intel's spiral is not friendly for investors ESPECIALLY after the statement about leaving cutting edge nodes as a possibility.

Why would you invest in a company who is in such dire straights they are gutting their company & even AFTER that is in such a bad place it risks them leaving an entire market?

Doing a $100bil + dollar investment out of your own pocket isn't easy but the fact that they made $100bil to put into this investment is the reason to invest.
If intel knows one thing it's how to make money.
The next few months will be the best time to invest in intel to try and make money since the release of 18A CPUs will be the 'make or break' point, either the stock will go up after that or at least you won't have spend that much on the stocks.

That is insanely high risk all together. You could easily invest in Nvidia or amd and get profit w/o any of that risk.

Nvidia and amd stock are so pumped up right now that there is no way to make any money from it...
Buy low and sell high is how you make profit, not buy high and sell low.

 

[palladin9479]

Hmm looks like this is being used as a way to generate capital, remember they still maintain ownership of the NEX company just not full ownership as they can now sell part of that ownership to a bank or investment fund. That capital can be used to stablish and fund core business units like processor design or possibly fabrication.

They would do this if they expect the spun off division not to be a candidate for future growth.

 

[Moores_Ghost]

Edge? Sure. Network? Tan has lost his mind. I understand he's a penny pincher and they need that but he's pinching the wrong pennies, firing the wrong people, and hasn't changed Intel's internal culture, which is toxic at best.
This is another move in a long list of moves designed to keep them afloat, not move them forward. He believes Intel's recovery to be a marathon when it doesn't have to be plus, this is a person that said he doesn't believe in "If you build it they will come" , a Field of Dreams quote. If you build it, as in do it right for the right price, they WILL come. AMD and Nvidia are proof enough.
They need a new corporate culture, they need extreme restructuring and now I feel they need a new CEO. Tan will bankrupt that company, but it will be later than sooner. I seriously doubt shareholders that believe in Intel want that. I'm shocked so few see it. Sadly by the time they do it may be too late.

I have no time for petty squabbling so if you do not like my opinion, please move on.

 

[baboma]

>They would do this if they expect the spun off division not to be a candidate for future growth.

I think Tan clearly laid it out in his email to employees--really to all stakeholders, since it was published on the Intel website. His "three strategic pillars of growth" are 1) get foundry capex under control, 2) revitalize x86 (the current breadwinner), and 3) do AI (the future breadwiner).

https://newsroom.intel.com/corporate/lip-bu-tan-steps-in-the-right-direction

Those are core. Anything that isn't core will be shed. It's not about whether the networking div will be profitable. That, and Intel can use any help to bulk up its cash reserves.

Relevant to the interests of this forum, Tan specifically mentioned Nova Lake--"We also must drive continued progress on Nova Lake to close gaps in the high-end desktop space."--which is good news on that front, the relentless focus on core and lack of any mention of consumer GPUs means that consumer dGPU is in all likelihood, done.

There's some tentative movement on pursuing edge AI with the repurposing of Battlemage into the B50/60 Pro products. Tan has mentioned there'll be more announcements on this front, so I expect to see follow-ups with Celestial/Druid/etc. My assessment would be same, that this space is Intel's best bet to compete in for AI.

 

[Mr Majestyk]

I hope betting the entire house on AI blows up in his face. Another wannabe AI mercenary that missed the boat by a decade and despite saying he doesn't think Intel can catch up, is betting everything on it.

 

[virgult]

the relentless focus on core and lack of any mention of consumer GPUs means that consumer dGPU is in all likelihood, done.

That would be monumentally stupid. The one ace they have up their sleeve is offering a competitor to CUDA on reasonably priced consumer GPUs. That's their entry point into the AI market. They've already developed it, all they need to do is support it and market if for what it is - then the entire hobbyist and educational/university market would be theirs.

 

[dalek1234]

...he's pinching the wrong pennies, firing the wrong people, and hasn't changed Intel's internal culture, which is toxic at best.
...

He did say few days ago that he wants to change the company culture, but he didn't say how he is going to achieve that. Even if he does have a plan, I doubt it is doable. The toxic culture is such a big part of Intel's existence, I don't think it can be changed. He would need to fire everyone and hire new people, which is not a solution.

 

[bit_user]

For Tan, the NEX spin-off is emblematic of a broader philosophy shift: Intel is exiting everything that’s not essential to its x86 and AI ambitions.

It's funny how Nvidia and AMD recently moved in the opposite direction. Nvidia bought Mellanox about 6 years ago, and AMD bought Pensando like 3 years later. That's because they both understand how essential networking is to scale AI training.

Even though Intel will still retain part ownership, what they lose by bringing in other investors is that the entity must now prioritize its own interests. Intel can no longer twist its product roadmap or pricing structure to support its efforts to sell its servers or AI accelerators in ways that might not maximize shareholder value for the new company.

For instance, if the new company wants to work with another AI chip company, like Tenstorrent, I think Intel can't block that, even with a majority stake, unless they want to open themselves up to litigation by other investors claiming that Intel isn't making decisions in the new company's best interests.

 

[bit_user]

I think Tan clearly laid it out in his email to employees--really to all stakeholders, since it was published on the Intel website. His "three strategic pillars of growth" are 1) get foundry capex under control, 2) revitalize x86 (the current breadwinner), and 3) do AI (the future breadwiner).

Item 1 will kill the foundry business.

Item 2 will never return Intel to its prior greatness, because huge swaths of the market are moving on from x86 and cannot be won back.

Item 3 is a massive uphill battle, given how far behind Intel is on AI. AMD is much further along and still way behind Nvidia. Furthermore, I bet Intel will continue trying AMD's losing strategy of making a cut-rate Nvidia knockoff product. You can never beat Nvidia at their own game. The only chance to beat them is doing an endrun, like Cerebras did.

the relentless focus on core and lack of any mention of consumer GPUs means that consumer dGPU is in all likelihood, done.

Recent Linux development for upcoming Battlemage dGPUs suggests otherwise, but you might be right.

My assessment would be same, that this space is Intel's best bet to compete in for AI.

We'll see what AMD comes up with, in UDNA. I wonder if they're trying to merge XDNA into it, as well. It sure would be neat if they could consolidate all of the compute and memory resources between their GPU and NPU blocks, but I don't know if that's the plan.

 

[bit_user]

If that section of their business includes their PCIe NICs and WiFi adapters, then part of Intel's business issue is really bad marketing.
...
They need to push for lower cost 10GbE NICs and use their connections to get them on store shelves of every major electronics store, same for their WiFi adapters.

They need to push for multiple home user tiers of networking hardware. For example, Ease up on the price gouging and push some of their QSFP28 NICS and some Tofino 2 based switches as a higher end solution for home wired networking.

They clearly don't see profit in the consumer networking market, IMO. I think they fear Realtek and Marvel will just undercut whatever they try to do, there.

To hit lower price points, one thing they ought to do is a refresh on a smaller node. Their current 10GBase-T chips run too hot and really need active cooling for a proper consumer product.

 

[TerryLaze]

It's funny how Nvidia and AMD recently moved in the opposite direction. Nvidia bought Mellanox about 6 years ago, and AMD bought Pensando like 3 years later. That's because they both understand how essential networking is to scale AI training.

That's because nvidia and amd didn't have any of that tech.
Intel did/still does. It's not part of NEX.

https://newsroom.intel.com/artificial-intelligence/intel-unveils-first-integrated-optical-io-chiplet

Item 1 will kill the foundry business.

You can't kill something that didn't even start yet....
It will slow it down from the initially way too highly estimated growth.

Item 2 will never return Intel to its prior greatness, because huge swaths of the market are moving on from x86 and cannot be won back.

Just look at how much x86 emulation happens in the arm market, arm users are thirsting to run x86 on their devices, the only thing stopping them is batteries, but by the time arm gets up to the performance of x86 they will be using the same amount of power at which point users will choose x86 over arm, because that's what they really want.
They only stick with arm for now because batteries can't keep up.

Servers that are on arm will stay on arm, but on servers GPUs and AI will be the big money makers in the future and not CPUs.

Item 3 is a massive uphill battle, given how far behind Intel is on AI.

It is, but they also don't need to be market leader in the field from day one just to make money, as long as they can sell enough AI they can make enough money.

Recent Linux development for upcoming Battlemage dGPUs suggests otherwise, but you might be right.

Having mass market GPUs is the best way for them to get feedback and bug reports that will make them better at making GPUs in general, so including AI/server GPUs.

 

[logainofhades]

I fear this bean counter was a bad choice as CEO.

 

[punkncat]

If Lip is trying to undermine the company, cost investors, and ruin what is left of an already smoldering wreckage then good job to him.

Completely wrong mind and word set to be trying to help anything. Probably on the wall over at AMD as ongoing employee of the month.

 

[bit_user]

You can't kill something that didn't even start yet....

That's obviously not true. Projects, businesses, etc. still in the planning phase get canceled or undermined every single day of the week.

It will slow it down from the initially way too highly estimated growth.

He said Intel will cancel 14A, if they don't find a big external customer. That's it. Done. Intel would then have to sell its fabs at a firesale price.

Just look at how much x86 emulation happens in the arm market, arm users are thirsting to run x86 on their devices, the only thing stopping them is batteries, but by the time arm gets up to the performance of x86 they will be using the same amount of power at which point users will choose x86 over arm, because that's what they really want.
They only stick with arm for now because batteries can't keep up.

I think people who made the switch to ARM didn't do it with an expectation to rely on emulation either for their main workloads or for very long.

It is, but they also don't need to be market leader in the field from day one just to make money, as long as they can sell enough AI they can make enough money.

The problem is that there's very little room in the market for anyone who's not the best. If you're not faster than Nvidia, then you need to at least provide more performance per $, which turns out to be almost as difficult.

Having mass market GPUs is the best way for them to get feedback and bug reports that will make them better at making GPUs in general, so including AI/server GPUs.

GPUs actually aren't ideal for AI, which makes me skeptical their long-term AI plans involve a dGPU of any sort. In spite of that, I hope they'll keep around their dGPU product line, but I'm not optimistic they will.

 

[bit_user]

I fear this bean counter was a bad choice as CEO.

If Lip is trying to undermine the company, cost investors, and ruin what is left of an already smoldering wreckage then good job to him.
...
Probably on the wall over at AMD as ongoing employee of the month.

You guys would do well to consider that the board who appointed him as CEO will have been briefed on his plans. If they didn't agree, then either he'd change plans or he'd be out of there.

That said, I don't mean to imply that any CEO they appointed would've done the exact same things. I'm just saying that it's not only him who deserves criticism, but also the board that hired him and is backing these moves.

 

[logainofhades]

I do not trust the board to make the right decisions either, given the situation Intel now finds themselves in.

 

[punkncat]

Where there is no doubt that Intel laid back on their laurels too long and created a situation they are going to have to try and survive, the way this guy is laying it out is about as optimistic as a cancer diagnosis. He could at least pretend he has aspirations of things getting better.

 

[TerryLaze]

That's obviously not true. Projects, businesses, etc. still in the planning phase get canceled or undermined every single day of the week.


He said Intel will cancel 14A, if they don't find a big external customer. That's it. Done. Intel would then have to sell its fabs at a firesale price.

Ok yeah, but they did get customers even for 18A so if they include customers in the design of 14A then there is no way they won't get enough customers for that.

Also firesale? Why?! They can keep plots of land for pretty little upkeep.
If 18a and 14A works they will have to use them for their own production as well, they wouldn't sell them.

I think people who made the switch to ARM didn't do it with an expectation to rely on emulation either for their main workloads or for very long.

Which ones? The server people? Yes, absolutely, that's why I said that those won't go back.
If you are talking about smartphones and other devices, nobody chose arm, that was just the only thing available for a very long time.

The problem is that there's very little room in the market for anyone who's not the best. If you're not faster than Nvidia, then you need to at least provide more performance per $, which turns out to be almost as difficult.

That's only true if the best can supply enough product for the whole market...
If the choices are to wait 2-3 years for the best and lose all of the money in the meantime, or get the second best and make a good piece of that money, then the choice is very clear.

 

[bit_user]

Ok yeah, but they did get customers even for 18A so if they include customers in the design of 14A then there is no way they won't get enough customers for that.

Then I guess you didn't see this article, because if that were true then there'd have been no reason for Tan to threaten canceling 14A, at the end of last week:

https://www.tomshardware.com/tech-i...-cede-leading-edge-market-to-tsmc-and-samsung
​

Also firesale? Why?! They can keep plots of land for pretty little upkeep.
If 18a and 14A works they will have to use them for their own production as well, they wouldn't sell them.

He said they'd cancel 14A, outright. Not keep it for internal use. A "firesale" is the term for when you sell something at a huge loss, which is what would happen.

That's only true if the best can supply enough product for the whole market...
If the choices are to wait 2-3 years for the best and lose all of the money in the meantime,

It's been a while since I've seen news of major backlogs on Nvidia's AI server products. I think the supply chain is currently pretty well caught-up with demand.

Furthermore, when they were backlogged, the production bottlenecks were HBM and advanced packaging. Intel is reliant on the exact same HBM suppliers.