Qualcomm explored acquiring pieces of Intel chip-design business

[IanSzot]

Summary

Qualcomm has explored purchasing parts of Intel chip design business. They're interested in the client PC designs but everything's on the table (Altera, for example). They're not interested in the server segment.

Intel said they're "commited to their PC business" and Qualcomm refused to comment on the matter.

 

Quotes

Quote

Qualcomm has explored the possibility of acquiring portions of Intel's ns new tab
design business to boost the company's product portfolio, according to two sources familiar with the matter.

The mobile chipmaker has examined acquiring different pieces of Intel, which is struggling to generate cash and looking to shed business units and sell off other assets, the people said.

Intel’s client PC design business is of significant interest to Qualcomm executives, one of the sources said, but they are looking at all of the company’s design units.

 

Intel's board is set to meet next week to weigh a proposal from Intel CEO Pat Gelsinger and other executives on how to trim its operations in an attempt to save cash. Potential options include a sale of its programmable chip unit, Altera, Reuters reported.

 

My thoughts

Everyone knows Intel has been struggling lately and they need cash asap. Qualcomm has over double Intel's market cap so they have plenty of cash to buy parts of Intel if they want to.

I doubt Intel would sell their PC and Server designs as these are their most lucrative business but I could see they trying to get rid of Altera. They purchased Altera in 2015 for 16 billion dollars but still struggle to conquer market share against AMD's Xilinx (from what I researched Xilinx has about 60% market share on FPGAs). Marvell also seems like they're interested in purchasing Altera.

 

Sources

Main story: https://www.reuters.com/technology/qualcomm-has-explored-acquiring-pieces-intel-chip-design-business-sources-say-2024-09-06/

CEO wants to sell Altera: https://www.reuters.com/technology/intel-ceo-pitch-board-plans-shed-assets-cut-costs-source-says-2024-09-01/

[Agall]

7 minutes ago, IanSzot said:

Summary

Qualcomm has explored purchasing parts of Intel chip design business. They're interested in the client PC designs but everything's on the table (Altera, for example). They're not interested in the server segment.

Intel said they're "commited to their PC business" and Qualcomm refused to comment on the matter.

 

Quotes

 

My thoughts

Everyone knows Intel has been struggling lately and they need cash asap. Qualcomm has over double Intel's market cap so they have plenty of cash to buy parts of Intel if they want to.

I doubt Intel would sell their PC and Server designs as these are their most lucrative business but I could see they trying to get rid of Altera. They purchased Altera in 2015 for 16 billion dollars but still struggle to conquer market share against AMD's Xilinx (from what I researched Xilinx has about 60% market share on FPGAs). Marvell also seems like they're interested in purchasing Altera.

 

Sources

Main story: https://www.reuters.com/technology/qualcomm-has-explored-acquiring-pieces-intel-chip-design-business-sources-say-2024-09-06/

CEO wants to sell Altera: https://www.reuters.com/technology/intel-ceo-pitch-board-plans-shed-assets-cut-costs-source-says-2024-09-01/

Their Core Ultra 200V chips seem to look really good, so that gives me hope they won't die. The iGPU performance so far beats out even the 890M in the new HX 370 chip.

 

I haven't been able to confirm it yet, but i think Intel makes their Core Ultra 200V series on their own fab and not TSMC?

[Forbidden Wafer]

8 minutes ago, Agall said:

Their Core Ultra 200V chips seem to look really good, so that gives me hope they won't die. The iGPU performance so far beats out even the 890M in the new HX 370 chip.

 

I haven't been able to confirm it yet, but i think Intel makes their Core Ultra 200V series on their own fab and not TSMC?

The reports says everything new is in TSMC.
https://videocardz.com/newz/intel-core-ultra-200v-lunar-lake-enters-production-at-tsmc

[Agall]

3 minutes ago, Forbidden Wafer said:

The reports says everything new is in TSMC.
https://videocardz.com/newz/intel-core-ultra-200v-lunar-lake-enters-production-at-tsmc

Makes more sense  thanks for finding that.

[ImorallySourcedElectrons]

1 hour ago, IanSzot said:

Everyone knows Intel has been struggling lately and they need cash asap.

They do? https://www.intc.com/financial-info/cash-flow This isn't bad for a company working on massive fab projects, not great but it's not like they're going to have to start selling the kitchen sink anytime soon.

Intel needs certain vulture capitalists need to stop running a crappy PR campaign designed to pressure Intel into splitting off its foundry business, and a lot of Intel's actions now are related to pleasing certain investor groups to prevent lawsuits from said investors and hostile actions from those groups. Simultaneously, journalists need to stop enabling said investors. Heck if we apply the same logic regarding cash flow, Microsoft should be putting itself up for sale right about now.

 

1 hour ago, IanSzot said:

Qualcomm has over double Intel's market cap so they have plenty of cash to buy parts of Intel if they want to.

That ain't how that works, by that logic NVidia should be ready to just straight-up buy Intel right about now.

 

1 hour ago, IanSzot said:

They purchased Altera in 2015 for 16 billion dollars but still struggle to conquer market share against AMD's Xilinx (from what I researched Xilinx has about 60% market share on FPGAs). Marvell also seems like they're interested in purchasing Altera.

Altera was always the second player, primarily due to garbage toolchain software and closed off mentality. Altera improved a lot under Intel, but companies like Qualcomm and Marvel would not really be able to run that business, they're already a nightmare to deal with in the electronics industry due to how closed their ecosystems are. Give Altera to either of them and its product line is dead within years. Then again, regulators should have never allowed Intel and AMD to snatch up Xilinx and Altera in the first place.

[sounds]

11 hours ago, ImorallySourcedElectrons said:

... companies like Qualcomm and Marvel would not really be able to run that business, they're already a nightmare to deal with in the electronics industry due to how closed their ecosystems are. Give Altera to either of them and its product line is dead within years. Then again, regulators should have never allowed Intel and AMD to snatch up Xilinx and Altera in the first place.

Xilinx and Altera's product lines are rarely seen in high volume products because as soon as the device is getting popular, there are lots of ways to cut the FPGA from the design.

 

Intel gobbling up Altera and AMD swallowing Xilinx definitely needed more regulation, but that time period saw lots of iffy consolidation. Even ARM was up for sale to Nvidia...

[ImorallySourcedElectrons]

4 hours ago, sounds said:

Xilinx and Altera's product lines are rarely seen in high volume products because as soon as the device is getting popular, there are lots of ways to cut the FPGA from the design.

I disagree, you might want to crack open some mid range multimedia equipment sometime. For example, Philips used a small FPGA for their ambilight system for quite a while there. You'll also find them in some camera's, and we use them at work in fairly high volume for products you wouldn't even go looking in. FPGAs are quite affordable in volume, but it's not the easiest component to put into a design without eating up your profit margin, I do agree with that.

 

4 hours ago, sounds said:

Intel gobbling up Altera and AMD swallowing Xilinx definitely needed more regulation, but that time period saw lots of iffy consolidation. Even ARM was up for sale to Nvidia...

And that's without getting into Analog, NXP, Microchip, and TI their buying sprees, that they were allowed to acquire National, Linear, Maxim, Hittite, Freescale, Atmel, microsemi, etc. 

[sounds]

Yeah, agreed, there are volume products using FPGAs. Doesn't the original Nvidia G-Sync require a proprietary circuit from nvidia with an FPGA?

 

Maybe what I could have said is, it eats up the profit margin so the company will be looking for a way to cut it from the design, since that's usually an "easy" optimization. (If reimplementing your core logic is easy.)

[ImorallySourcedElectrons]

17 minutes ago, sounds said:

Yeah, agreed, there are volume products using FPGAs. Doesn't the original Nvidia G-Sync require a proprietary circuit from nvidia with an FPGA?

 

Maybe what I could have said is, it eats up the profit margin so the company will be looking for a way to cut it from the design, since that's usually an "easy" optimization. (If reimplementing your core logic is easy.)

Yeah, and another reason to go with the FPGA is that it buys you an interesting upgrade track, you sometimes see this with wireless communication devices. That expensive FPGA can suddenly become very profitable if you're able to corner the market on day one while your competitors are rushing hardware redesigns.

[Haswellx86]

This shows that Intel themselves are admitting that their process node technology just isn't good enough. I wouldn't use words like "inferior" when compared to TSMC. They still might have the grunt over TSMC to push high performance and clocks, and there is also more flexibility if all the R&D is done in Intel's land.

 

But the fact that the whole Lunar Lake processors are fully designed in TSMC means at least in terms of efficiency, TSMC's manufacturing is superior.

 

Intel is still losing money (their stock is actually rising a bit now though). Their foundries are a huge toll on them. They better make Arc more compelling or else their GPU division might also fail. But it looks like they are probably doing good there.

[sounds]

1. Hardware is hard. Intel has always led in the hardware space but it's never a guarantee

2. Intel going with TSMC for Lunar Lake reflects more on how their foundries were doing 2-3 years ago

3. That still doesn't mean Intel's latest foundry process is better in 2024. I am too lazy to google it right now but there are ways to figure out if Intel can ship what they promised.

[ImorallySourcedElectrons]

On 9/12/2024 at 8:05 PM, Gat Pelsinger said:

This shows that Intel themselves are admitting that their process node technology just isn't good enough.

No, that's a very bad take on all of this.

On 9/12/2024 at 8:05 PM, Gat Pelsinger said:

I wouldn't use words like "inferior" when compared to TSMC. They still might have the grunt over TSMC to push high performance and clocks,

Intel actually tends to lead on density, for raw performance you might actually go to someone else than TSMC and Intel...

On 9/12/2024 at 8:05 PM, Gat Pelsinger said:

But the fact that the whole Lunar Lake processors are fully designed in TSMC means at least in terms of efficiency, TSMC's manufacturing is superior.

No, it's really common in the electronics industry to sell your own fab capacity to external customers and run your own chips in someone else's fab. The economics can get surprisingly intricate if you care about maximizing profits.

On 9/12/2024 at 8:05 PM, Gat Pelsinger said:

Intel is still losing money (their stock is actually rising a bit now though). Their foundries are a huge toll on them.

No, the foundries are a massive asset to them, stop repeating the vulture capitalist fund their punchline. And losing money and stock price have little to do with each other - they may be correlated but causation is a different thing. If you want to see money loss, see M$'s balance sheet for the last years.

[Mark Kaine]

On 9/11/2024 at 8:33 PM, ImorallySourcedElectrons said:

That ain't how that works, by that logic NVidia should be ready to just straight-up buy Intel right about now.

and they would if they were allowed to lol.

 

only reason that's "not how it works" are regulations to prevent monopolies...

 

[ImorallySourcedElectrons]

On 9/20/2024 at 9:07 AM, Mark Kaine said:

and they would if they were allowed to lol.

 

only reason that's "not how it works" are regulations to prevent monopolies...

 

Again, that ain't how buying a company works, companies like Intel are primarily owned by institutional investors. You can't just oblige them to sell their shares just like that because you got a big wad of cash laying around, especially for companies which have consistently paid out pretty good dividends.

[leadeater]

On 9/20/2024 at 7:07 PM, Mark Kaine said:

and they would if they were allowed to lol.

 

only reason that's "not how it works" are regulations to prevent monopolies...

 

You can't buy what isn't up for sale.

[Mark Kaine]

7 minutes ago, leadeater said:

You can't buy what isn't up for sale.

 

?

 

 

 

[Nimoy007]

23 minutes ago, Mark Kaine said:

 

To quote others in this forum, "That ain't how it works."

[starsmine]

1 hour ago, Mark Kaine said:

 

?

 

 

 

go ahead, buy a controlling share with the outstanding shares.

(if a share is not outstanding, it can not be bought)


edit: I want to be clear, there are ways to do it. But they rarely go well for the company. they are Hostile Takeovers.

[Mark Kaine]

1 hour ago, starsmine said:

go ahead, buy a controlling share with the outstanding shares.

(if a share is not outstanding, it can not be bought)


edit: I want to be clear, there are ways to do it. But they rarely go well for the company. they are Hostile Takeovers.

So you're saying those shares that get sold daily... aren't actually up for sale? huh...

 

1 hour ago, starsmine said:

But they rarely go well for the company. they are Hostile Takeovers

That's what im talking about all along.

 

And the us government would *NEVER* allow it, sorry that was my point from the get go. 

 

"THEY WOULD IF ALLOWED TO"

(but they aren't)

[starsmine]

16 minutes ago, Mark Kaine said:

So you're saying those shares that get sold daily... aren't actually up for sale? huh...

What? No, that is not at all what I am saying. The amount of shares that are floating on the exchange are rarely ever even approaching 10% of the total amount of shares in existence. 

16 minutes ago, Mark Kaine said:

That's what im talking about all along.

No you are not. those are not simple things, a hostile take over requires a lot of underhanded tactics to pick up enough shares to have a controling stake, and a lot of business politics and secret purchases. You cant just do it, and doubly so if its not for sale.

16 minutes ago, Mark Kaine said:

 

And the us government would *NEVER* allow it, sorry that was my point from the get go. 

 

"THEY WOULD IF ALLOWED TO"

(but they aren't)

They wouldnt. 

What are they even thinking about buying that they dont already have? Mobile eye?

[leadeater]

3 hours ago, Mark Kaine said:

?

Doesn't matter how much money anyone has, you cannot buy what is not for sale. No money in the world bypasses "not for sale"

 

Also FYI Nvidia is critically over valued share price.  Nvidia cannot be valued more than Intel when even now their existence is underpinned by Intel existing at all. Nvidia doesn't have the stronger hand in anything when Intel is the one allowing their products to be used at all. If we did a bit of speculative role play and made Intel not exist at all over night and all their products disappeared in to a black hole Nvidia would immediately collapse and cease to exist.

 

Speculative company value != actual company value.

 

Spoiler

No ARM and Nvidia's own ARM efforts would not keep them alive right now

 

 

2 hours ago, Mark Kaine said:

So you're saying those shares that get sold daily... aren't actually up for sale? huh...

Daily trade is irrelevant, it's also not how you purchase controlling stakes of a company or do business mergers. The other thing is the volume of shares traded per day is 1.99% of the total available and those will pretty much be the same shares each day. So even if you brought all 1.99% being traded in a day the next day it would be near zero since you just purchased them all. Do you think you could force Intel to sell the fabrication business unit with 1.99% company shares?

 

Quote

The ownership structure of Intel (INTC) stock is a mix of institutional, retail and individual investors. Approximately 59.70% of the company’s stock is owned by Institutional Investors, 0.68% is owned by Insiders and 39.61% is owned by Public Companies and Individual Investors.

 

Buying shares and company mergers and sales literally are not the same thing at all.

 

 

You might entice some of the Mutual Fund holders to sell but you won't have any remote chance of getting all the Institutional Holders to sell.

 

 

 

[porina]

28 minutes ago, leadeater said:

Also FYI Nvidia is critically over valued share price.

Agreed. While market cap is one measure of a company size or worth, it isn't the only or even ideal way to do it.

 

28 minutes ago, leadeater said:

Nvidia cannot be valued more than Intel when even now their existence is underpinned by Intel existing at all. Nvidia doesn't have the stronger hand in anything when Intel is the one allowing their products to be used at all. If we did a bit of speculative role play and made Intel not exist at all over night and all their products disappeared in to a black hole Nvidia would immediately collapse and cease to exist.

Care to expand on this? I guess the key part of this is "not exist at all over night and all their products disappeared in to a black hole" which, be honest, is totally unrealistic. That wouldn't just impact Nvidia, but much of the world would be in chaos.

 

Somewhat more realistic hypothetical scenarios are:

1, Say Intel does run out of cash in near future. US gov says, we'll bail out the fabs cos they're still nice, but CPUs go bye bye. IP gets sold off bit by bit and doesn't go anywhere fast. Existing Intel systems will keep running without further support. Users big and small would switch over to AMD if it has to be x86, and ARM or others could see uplift otherwise.

 

2, Intel never existed at all. Then we would see the computing industry take a very different path over the years. The AMD we know will not exist either in that case, since they were enabled by their LEGAL copying of Intel products. AMD could still exist as they did other things, but it wouldn't be x86 based. Who knows what architecture we'd be using instead.

 

28 minutes ago, leadeater said:

No ARM and Nvidia's own ARM efforts would not keep them alive right now

If Xeon stopped being offered for whatever reason, couldn't they move to Epyc? Again, it wouldn't be an overnight thing so AMD could have time to adjust capacity accordingly.

[Mark Kaine]

40 minutes ago, leadeater said:

Doesn't matter how much money anyone has, you cannot buy what is not for sale. No money in the world bypasses "not for sale"

 

Also FYI Nvidia is critically over valued share price.  Nvidia cannot be valued more than Intel when even now their existence is underpinned by Intel existing at all. Nvidia doesn't have the stronger hand in anything when Intel is the one allowing their products to be used at all. If we did a bit of speculative role play and made Intel not exist at all over night and all their products disappeared in to a black hole Nvidia would immediately collapse and cease to exist.

 

Speculative company value != actual company value.

 

  Reveal hidden contents

No ARM and Nvidia's own ARM efforts would not keep them alive right now

 

 

Daily trade is irrelevant, it's also not how you purchase controlling stakes of a company or do business mergers. The other thing is the volume of shares traded per day is 1.99% of the total available and those will pretty much be the same shares each day. So even if you brought all 1.99% being traded in a day the next day it would be near zero since you just purchased them all. Do you think you could force Intel to sell the fabrication business unit with 1.99% company shares?

 

 

Buying shares and company mergers and sales literally are not the same thing at all.

 

 

You might entice some of the Mutual Fund holders to sell but you won't have any remote chance of getting all the Institutional Holders to sell.

 

 

 

Ok so its difficult,  but the news says "CEO wants to sell... Altera..."?

 

So in theory is he allowed to or not? I.e. What if intel wants to sell (piecemeal) ? (that's what the news says, right?)

 

Im not saying its gonna happen anytime soon, on the contrary im still thinking even if intel wanted there are probably tons of roadblocks, not at last government intervention, but in theory its possible...

Alternatively intel just gives up and stops producing anything  - i just can't see how the company survives long-term,  but ya'll are saying "nah will never happen..."

 

Ima say in 20 or so years there won't be a chip producing company called intel if they don't change course very very swiftly... 

 

who's gonna even buy their products anymore if they don't work? 

 

 

im kinda thinking of other companies that were "too big to fail" like IBM for example 

 

https://leaders.com/articles/business/ibm-history/#:~:text=A failure to capitalize on,all contributed toward IBM's decline.

 

 

yeah, they still not dead, but are they really relevant anymore? 

 

"building a quantum computer"

 

I mean sure, it's nice to want things i guess, but will it work?

 

Microsoft is also indeed in a similar situation, sure they got tons of money and income,  but do they have a vision for the future beyond "the cloud"? 

 

 

[porina]

2 hours ago, Mark Kaine said:

Ok so its difficult,  but the news says "CEO wants to sell... Altera..."?

 

So in theory is he allowed to or not? I.e. What if intel wants to sell (piecemeal) ? (that's what the news says, right?)

If Intel wants to offer to sell parts of its business, they can explore that at any time. The point Leadeater was making earlier is the other way around. You can't just turn up and expect to buy if something isn't offered on sale. It doesn't hurt to ask, but don't get your hopes up.

 

2 hours ago, Mark Kaine said:

im kinda thinking of other companies that were "too big to fail" like IBM for example 

IBM of today is a successful company. They're not consumer facing any more so it is off radar to most people.

 

While I said market cap is not a good measure of size, they're still up there. (USD equivalent)

Apple 3.5T

Microsoft 3.2T

Nvidia 2.8T

TSMC 0.8T

ASML 320B

AMD 252B

IBM 200B

Qualcomm 188B

Intel 93B

 

Even before the recent troubles I've felt Intel was under valued for many years. Nvidia is over valued and it'll be interesting to see how long that lasts.

[leadeater]

4 hours ago, porina said:

Care to expand on this? I guess the key part of this is "not exist at all over night and all their products disappeared in to a black hole" which, be honest, is totally unrealistic. That wouldn't just impact Nvidia, but much of the world would be in chaos.

It was supposed to be unrealistic simply to highlight how Nvidia can't exist without Intel. You really shouldn't have Nvidia being valued so much more than Intel when the market is largely underpinned by Intel and Intel is also a leading technology fabricator. How it makes sense to value Nvidia so much more when they can't make anything without others or be used without others is outright illogical valuation 

 

It would be fine if Intel and Nvidia were valued relatively the same, company value is wishy washy after all but that's not the case and it's a ~30x difference.

 

4 hours ago, porina said:

If Xeon stopped being offered for whatever reason, couldn't they move to Epyc? Again, it wouldn't be an overnight thing so AMD could have time to adjust capacity accordingly.

AMD couldn't supply the market right now no matter how much they tried, But I guess on the other hand since Nvidia couldn't use so much TSMC allocation anymore they could. Who knows but it would be chaos and I don't think the market would survive as it is.

 

To have Intel and it's products go poof  overnight would be truly apocalyptic.

Edited September 22 by leadeater