Business Qualcomm approached Intel about a takeover in recent days, source says

  • Want to keep track of this thread?
    Accounts can bookmark posts, watch threads for updates, and jump back to where you stopped reading.
    Create account
Qualcomm approached Intel about a takeover in recent days, source says
Article | Archive
By Max A. Cherney
September 20, 2024 9:11 PM PDT Updated 16 hours ago

SAN FRANCISCO, Sept 20 (Reuters) - Qualcomm (QCOM.O) has in recent days approached Intel (INTC.O)
to explore a potential acquisition of the troubled chipmaker, a source familiar with the situation said on Friday, in what could be a transformational deal in the sector but faces many hurdles.

Qualcomm CEO Cristiano Amon is personally involved in the negotiations to acquire five-decade-old Intel, according to the source who was briefed on the matter. Another person familiar with the situation said Amon has been actively examining various options for a deal for the company.

Earlier this month, Reuters reported that Qualcomm explored the possibility of acquiring portions of Intel's design business and that its PC design unit was of particular interest. Qualcomm executives were examining Intel's entire portfolio of businesses. The conversations with Intel are at an early stage. The San Diego-based company has not made a formal offer for Intel, according to third person familiar with the matter.

The sources requested anonymity as the discussions are confidential.
Intel declined to comment. Qualcomm did not immediately respond to a Reuters request for comment.
Intel's shares closed up 3.3%, while Qualcomm fell 2.9%.
The approach by Qualcomm comes at a moment of weakness for Intel, which was once the most valuable chipmaker in the world, but whose shares have lost nearly 60% of their value since the start of the year.

A deal, should it go ahead, would likely invite scrutiny from antitrust regulators in the United States, China and Europe. Qualcomm may be required to divest parts of Intel in order to gain regulatory approvals.
A bid would mark the biggest takeover attempt in the technology industry since Broadcom sought to buy Qualcomm for $142 billion in 2018, before President Donald Trump nixed the tie-up, citing national security risks.

Reuters could not determine how Qualcomm, which has a market value of $188 billion, would finance a bid for Intel, which is valued at $122 billion, including its debt.
Qualcomm has roughly $13 billion in cash, according to recent company filings.
It is also unclear how Qualcomm would handle the takeover of Intel's contract manufacturing business. To build chips with an atomic level of precision, Intel has invested hundreds of billions of dollars over decades on its fabrication process and amassed tens of thousands of engineers to do it.

Qualcomm has never operated a chip factory, or fab, and currently contracts the likes of Taiwan Semiconductor Manufacturing Co (2330.TW)
and uses designs and other technology supplied by Arm Holdings (O9Ty.F)

INTEL'S WOES​

Once the dominant force in chipmaking, Intel ceded its manufacturing edge to Taiwanese rival TSMC and failed to produce a widely desired chip for the generative AI boom capitalized on by Nvidia (NVDA.O) and AMD (AMD.O)

Intel has been attempting to turn its business around by focusing on AI processors and creating a chip contract manufacturing business, known as a foundry.

As part of a memo from CEO Pat Gelsinger, Intel released a series of announcements that stemmed from a board meeting last week. Gelsinger and other executives presented a plan to shave off businesses and restructure the company, Reuters has previously reported.
The company plans to pause construction on factories in Poland and Germany, and reduce its real estate holdings. Intel also said it had reached a deal to make a custom networking chip for Amazon.com's AWS.

The Wall Street Journal reported on Qualcomm's talks with Intel earlier on Friday.
 
China #1

It wouldn’t change new management since Intel was already run by China. Let them merge.
 
Once the dominant force in chipmaking, Intel ceded its manufacturing edge to Taiwanese rival TSMC and failed to produce a widely desired chip for the generative AI boom capitalized on by Nvidia (NVDA.O) and AMD (AMD.O)

Intel has been attempting to turn its business around by focusing on AI processors and creating a chip contract manufacturing business, known as a foundry.
*Loses to Nvidia and AMD in AI chip development*
*decides the best course of action is to "try harder" at creating AI processors*

Yeah, Intel is fucked. Rather than pivot into making products the average person actually wants at reasonable prices, something both NVIDIA and AMD have let go by the wayside to chase the AI dragon (that will totally be worth the Billions of investment dollars, trust me bro), they think its smarter to double down on a losing strategy that hasn't even paid off for its competitors (in terms of actual usable products I mean).

They will also never be acquired by Qualcom. Not because of anything like integrity or something impressive like that, but because the US government will 100% block the acquisition like they have done time and time again with Qualcom.
 
China #1

It wouldn’t change new management since Intel was already run by China. Let them merge.
Qualcomm is already an American company

They will also never be acquired by Qualcom. Not because of anything like integrity or something impressive like that, but because the US government will 100% block the acquisition like they have done time and time again with Qualcom.
Qualcomm had numerous previous acquisitions go though. What the government blocked previously was Broadcom attempting to acquire Qualcomm because it would put an American company under a company that was (previously) headquartered in Singapore.
 
Check Qualcomm on Wikipedia because I've never heard of the company before

Click on company founders page

Early life and career

Mfw
 
Sounds good. Let Q take over. Q has never designed a processor and never run a fab so they are sure to fuck things up even worse and crater shit completely.
 
I'd much rather nvidia bought out intel. Leather jacket man seems pretty good at running a company.
 
There are roughly two possibilities here.

1: Qualcomm is being led by absolute morons who could be outwitted by a pajeet phone scammer.

Or…

2: They’re trolling and stirring the pot and have no intentions to buy Intel.
(There’s plenty bad blood between Intel and Qualcomm. Intel tried to steal their modem business and Qualcomm is currently trying to steal Intel’s CPU market share.)

There’s zero chance this takeover would happen, the whole idea is retarded.

There’s virtually zero synergy between the two, except for the modem part. And conversely there’s a humongous amount of overlap: THREE kinds of CPUs, two different GPU designs, two NPUs, etc. etc.

They’d likely have to fire half the workforce if this ever happened because of the amount of overlap.

Yeah, Intel is fucked. Rather than pivot into making products the average person actually wants at reasonable prices,
Totally fucked. Gelsinger sounds like a fucking moron when he starts sperging about his “AI centrino moment”.

Yeah sure Pat… Consumers just can’t wait to get one of those AIPC laptops, so they can finally blur the background in Slack and Zoom through an NPU instead of through the GPU.

And instead of using the die space on their latest laptop CPUs for something useful, they slapped a gigantic NPU on there, because “MUH AI PEECEE REVOLUTION!!” Sure Pat! It’s right around the corner! Just like last year and the year before!
 
Last edited:
*Loses to Nvidia and AMD in AI chip development*
*decides the best course of action is to "try harder" at creating AI processors*

Yeah, Intel is fucked. Rather than pivot into making products the average person actually wants at reasonable prices, something both NVIDIA and AMD have let go by the wayside to chase the AI dragon (that will totally be worth the Billions of investment dollars, trust me bro), they think its smarter to double down on a losing strategy that hasn't even paid off for its competitors (in terms of actual usable products I mean).

They will also never be acquired by Qualcom. Not because of anything like integrity or something impressive like that, but because the US government will 100% block the acquisition like they have done time and time again with Qualcom.
The average person is retarded and tries to do everything on a phone, gets a confused look when something doesn’t work, and dusts off a mid-2010s laptop or uses their work laptop to do whatever it is they need to do.

Intel needs growth, not making the same chips with minor improvements like they’ve been doing for the past twenty years.
 
Intel needs growth, not making the same chips with minor improvements like they’ve been doing for the past twenty years.
Agreed completely, but extend this to all the chip manufacturers, in pretty much all spaces, and add the caveat that "Nobody knows how to do this". Part of the reason they're so dicks hard for AI is that its finally a potential market offering that actually needs new structural silicon.

This matters a great deal, because we're pretty much up near the edge of what the current stuff can manage, and while there are a few promising avenues for new paradigms over the fence, they're all still facing the great thermals problem and nobody has a good solution for it. Silicon is getting too hot, and being smaller doesn't actually matter much in high power applications - Datacenters are limited by thermals and interface bandwidth, not chip physical size. Smaller chips offered better efficiency, but that's starting to run into hard physical limitations with preventing the electrons from just crossing the gate anyway, and prevalling solutions like gate all around are proving difficult to manufacture, to the point that TSMC has the proven ability to do so, and is choosing not to.

The problem is that even if they cross this threshold, datacenters need cost efficiency, not sheer power - They're already renting an entire building, they don't really care if the chip is a quarter of a palm instead of a quarter of a fingernail, what matters is if it has good thermals and cheap costs, and if they can get ten thousand of them. Only the bleeding edge of enthusiast compute is interested in even more power, the vast majority of enthusiast users just float a few years behind the curve, and they only make up a tiny fraction of the wider consumer base, who don't actually give a shit about processing power - most craptop, android and iphone users have no clue what a gigaflop is, they just want more energy efficiency so the battery lasts longer. Making stuff run well under those constraints is the app store guys problem, its exceedingly rare that a consumer thinks "I need to buy a new phone so [app] runs faster", they just think app's shit.

I could keep going, but the crux of my point is that the compute market itself doesn't need faster, smaller, crazier chips, it needs cheaper chips, in absolutely bonker qualities. If you could release a chip with the power of an i7 from 4 years ago, but generate a profit off of an MSRP of ~$100, and have a deep supply, you'd print fucking money. But doing so requires basically rethinking chips from the ground up, a paradigm shift that nobody even has an inkling of how to approach - Closest maybe is RISCV, but by its nature its not a drop in solution.

AI is the lazy and quick way out. They can apply the same solutions they've already invented, but in a new configuration for a specific new problem. No need to face foundational industry changes, just make some new patterns and send them to TSMC to print out, off you go. Unfortunately, its a doomed venture - AI accelerators offer two general options, neither of which are particularly valuable to the consumer or the datacenter. You got neural net acceleration for error prone, prebaked and generally user immutable models. This is your copilot PC's building spybooks of everything you've done, or your chatbot telling you to be racist only to the right kinds of targets. The vast majority of consumer users are going to treat this as a novelty, most people don't organize their life enough to even try to make use of a personal assistant, and chatbots already have a bad reputation as data sources, so people will just go to google. Datacenters have even less interest, if they're offering AI services they're going to invest in AI hardware proper, like the shit Nvidia's putting out, CPU die ones are pointless to them. The other option is generative content models like Stable Diffusion, Llama, etc, which consumers also have no real need for. Popping SD to shitpost generate is fun for a fraction of people as a one time novelty, but it doesn't really do anything useful for you as a consumer. At the datacenter level, again, you'd just prefer the power of a dedicated hardware solution and not a silicon wart.

And before "Business computing for individuals" comes up, any organizational procurement team that spends this much on the rank and file laptop fleet instead of cheap dells needs to be lynched, the accepted business AI solution is already cloud for a good reason. You can give the people who actually use it licenses, ignore the rest, and when some retard spikes their laptop and smashes it, you don't have to buy an entire new AI hardware set. They'll just license power from the datacenter providers, and those providers are buying bespoke hardware.

Intel and the rest of the industry can try to come up with a new fancy way to draw the eyes of the average consumer, but the average consumer doesn't need more, and compute is too obscure to them to even try to educate them on why they might want it for an upsell.
 
you've never heard of a snapdragon? they make smartphone processors. how do you not know who qualcomm is?
I owned a snapdragon processor phone one time, was the only time it crossed my mind lol.

Edit: did 2 mins of research to find it has infected literally everything, fucking crazy. I blinked and all the sudden they appear in my mirror closer than they appear lol.
 
I owned a snapdragon processor phone one time, was the only time it crossed my mind lol.

Edit: did 2 mins of research to find it has infected literally everything, fucking crazy. I blinked and all the sudden they appear in my mirror closer than they appear lol.
That happened to me with Samsung. Had no idea they made washing machines, fridges, and basically anything you could want that's electronic. That was the day I learned about South Korea's Samsung cyberpunk dystopia. Guess I never noticed because I used to find appliances boring.

LG is also part of the Korean cyberpunk dystopia.
 
Back
Top Bottom