After a decade of being assured that the short support window for Android devices had a root cause in the lack of support device makers received from Qualcomm, they are the last company that I want to see buy Intel.
Meanwhile x86 has maintained backwards compat for (checks calendar) yeah decades. Literally decades. With standards up the wazoo to avoid the disaster that is the Arm ecosystem without UEFI or something like it.
Every Arm SoC being a snowflake needing special attention by the OS is a huge hassle. There's a reason there's no simplified Arm installer for operating systems.
Historically this backwards compatibility was a competitive moat for intel – having a large supply of weird instructions with some undocumented behaviour thrown in makes it more expensive to make competing chips.
The ISA isn't what makes x86 easier for operating systems to support than ARM SoCs. It's things like generic ACPICA instead of hardcoded devicetrees.
To be clear, I wasn’t claiming anything about operating systems. Merely that adding many weird instructions was a strategy intel used to try to make the jobs of amd and centaur and suchlike harder.
Some ARM systems (mainly servers) do support ACPI; allowing for one image to run on multiple processors and devices.
However… ACPI is apparently a pretty awful thing to implement. When it doesn’t work, or mistakes are made (looking at my own 13th gen HP laptop right now - borked ACPI tables means unpatchable broken sleep on Linux), then it’s pure frustration.
Device trees on the other hand are much more binary. Either everything generally works or it doesn’t at all. It’s a valid approach.
Flawed implementations of open specs can be worked around with things like quirk tables. A spec held hostage by a non-cooperating vendor cannot. In the world of ARM SoCs, bad vendors won't even provide a device tree, just a binary image compiled from a patched kernel.
You can override broken ACPI tables. The keyword to search for is "dsdt". For example, if you have an HP Omen this repo has a fix:
https://github.com/j0hnwang/OMEN-Transcend-16-ACPI-fix
But yes, device trees are far nicer to work with.
nobody is a saint on this matter. both parties act the way they do to support their business position.
short-term support helps qualcomm sell more because their customers want to sell more phones to the same users. on the other hand, for the industrial partners there is an "LTS" model of support given (side effect being fairphone 5 https://news.ycombinator.com/item?id=37320800).
long-term support through backwards compatibility helps intel because they can sell new chips with the assurance that the ancient, unmaintained industrial software continues to run in a shinier box.
there's no simplified Linux installer either and every distro creates one and keeps re-inventing it
Intel squandered its dominance on the CPU market for decades. Qalcomm sucking the remaining life of it would be a fitting end for a player that lost its way.
Wonder if the increasing backwards compatibility became too much to bear, but IMO it never really tried to tread new grounds for risk of losing a comfortable position.
Qualcomm main business is mobile phone chips and 5g modems.
Majority of it's revenue goes to Taiwan for TSMC as margins.
Having Intel fab, will cut the middle man, and revenue will skyrocket, all while no money leaves the USA.
[dead]
Qualcomm still hasn't shipped any of the Snapdragon X dev kits, two months and counting. If they can't deliver on their promises (that and CoPilot PCs having very disappointing sales), how could they do anything besides further drag down Intel?
Not only that, it sounds like a major customer (Apple) is close to finally ditching Qualcomm's wireless chips? (At least that's been rumored [1])
[1] https://appleinsider.com/articles/24/07/24/apple-has-reporte...
Qualcomm's RF design is best in class. This is their bread and butter and they have been consistently good at it forever.
Apple purchased intel's RF baseband division,which was awful, and has been working on it in secret for years. It remains to be seen how this will go for Apple. It is attractive to Apple for cost and efficiency reasons (theoretically they can bury this all on a single SoC if they wish to) not because Qualcomm is bad.
It bears in mind that just because you are good at one thing does not imply you will be good at another. For instance, Intel's networking is mediocre to bad depending on the product or various entities trying to produce MIPS and ARM products failing time and time again.
> Intel's networking is mediocre to bad depending on the product
That's an interesting statement, as their Wi-Fi cards are some of the best on the market and common laptop-purchasing wisdom says to buy anything with an Intel Wi-Fi adapter and avoid everything else.
> Apple … has been working on it in secret for years.
I keep hoping Apple will release a MacBook with a 5G chipset. The rumors are saying their in house one will ready in 2026 at the earliest. It sure seems like a long road given they bought the intel RF division in 2019.
My understanding is that Intel's chips weren't great and making power efficient 5g chips is wildly difficult. Thus ends my understanding of these issues, though.
And Qualcomm patents. I just don't see how you can patent anything related to complying with a radio spec, there has to be limited ways to comply
Companies should do what they say they're going to do, but these dev kits are an example of something that's relevant to HN but not to Qualcomm's business.
A large part of the success of this new platform is how fast devs can adapt / fix their apps to work natively. Apple, for instance, provided dev mules for OSX ARM --- and their rollout of Apple Silicon was smoother than anyone could have hoped.
Windows ARM -- still borked in SOOO many ways -- and its 10+ years old now.
> Windows ARM -- still borked in SOOO many ways -- and its 10+ years old now.
Internally, nearly 20 years. It was kept alive for a long time by a single individual as a side project. When I first got out of college I actually helped update tests that were being used for it (I maintained the ARM compiler test harness, and it was being used for some Windows on ARM stuff as well).
Microsoft has never went fully in on arm, whereas Apple was willing to burn bridges and start brand new.
To be fair to Microsoft Apple did have the advantage of controlling the stack from top to bottom.
How can you expect your products to have good software support by developers then?
Snapdragon laptops have been available for a few weeks already. Although laptops cost more than this dev kit they're also more usable as a daily driver. If Qualcomm wants real adoption they'd send them out for free, not require devs to pay.
Dev shops would rather not have piles of laptops with batteries to maintain sitting in their racks/shelves for build and remote testing.
For individual devs laptops are fine, usually, but there's also no solid "reference" platform, since all the laptops are targeting different consumer lines.
That's a bit beside the point though, the Dev kits should've come out months before the consumer products were launched... and failing that at the same time.
Qualcomm and Arrow said the units would ship "tomorrow" in July... and it took over a month (after accepting many orders) before they even updated the stock to a more realistic timeframe, late September.
When this platform was heralded as the “AI” desktop, I pre-ordered both the dev kit and a laptop. Like many of you, I’ve experienced a months-long delay in the delivery of the dev kits. Although I STILL don't have my devkit, I received my laptop pretty much on time. -- and I quickly discovered that despite Windows on ARM (WOA) being over a decade old, the support for open-source tooling is as complete as Swiss cheese. Key Python modules are missing, and even the Git command-line (git bash) client isn’t functional yet!
I mean, forget about basic open-source development, let alone performing AI inference work on your new Snapdragon laptop.
After some digging, I’ve learned that just four overworked developers in Prague make up the core team unclogging this tooling dependency log-jam. Gah!
For what it’s worth, WSL2 (Linux on Windows) actually runs quite impressively on the Snapdragon X.
Why run Windows on ARM, when Linux on ARM is so.much more mature? Or are you buildings a Windows-specific product?
My plan was to run the laptop Windows and the 'devkit' Linux. This first batch of laptops (AFAICT) can NOT dual boot.
My thinking is this, if Windows ARM is a success -- there will be more units out there that can ALSO run Linux too. If Windows ARM is a failure, then Linux will suffer too.
> After some digging, I’ve learned that just four overworked developers in Prague make up the core team unclogging this tooling dependency log-jam. Gah!
What an embarrassment. So basically, it’s not a serious product.
... Intel wasn't able to ship on EUV, so they would be in great company.
[dead]
The headline seems intentionally bombastic and false. The text specifies that they are interested in lines of business, for example consumer computing, not the entire entity.
Qualcomm consists of at least 4 lawyers for every engineer. If this happens expect a lot more lawsuits making anything involving hardware much more expensive for everybody.
Not plausible. I worked at Qualcomm for several years in engineering and office of the chief scientist, and that would be insane inversion of division headcounts.
Qualcomm makes lots of their money by holding a monopoly on wireless chip patents. They use lawyers to bully other companies out of the space.
You can compare this with the patent wars of the companies in Silicon Valley which came to halt when the orgs realized they were effectively giving money to lawyers instead of innovating.
Qualcomm doesn’t really have real competition in Southern California. It’s cheaper for them to bully smaller companies with lawyers than employ more engineers (not sure if it’s possible to employ more engineers in the wireless space regardless).
You could also argue Qualcomms success is related to the other companies which reside around them. They have effectively built an “office moat” with their wireless patents.
so Qualcomm is basically Oracle?
Sort of, in Northern California there was a lot of office space.
In San Diego most of the land zoned for office space is owned by the Jacob brothers (although not directly connected to Qualcomm anymore). Imagine if Google and Apple had to rent their campuses from Oracle.
No way, 4x?!
Wasn't there some clause in the Intel/AMD x86-64 cross-licensing deal which voids it if either company changes ownership? I have some recollection of that being a thing.
But to be fair, Qualcomm might not care.
Deals can be amended for $$
The idea is to keep the number of x86 suppliers low, but enough patents are expired already you could probably make an x86_64 avx2 era cpu without asking
[dupe]
Some more discussion: https://news.ycombinator.com/item?id=41604817
How are MS and Qualcomm finances related ? They look like good budies with no visible reason. And if they are close then MS can have their own cpus. Win4ever !!!11 :)
But it's possible plain silicon is outdated and Intel stuff was splitted for a reason, eg. "photonics" part goes to "datacenter" division, fabs can be spun off and re-named any second.
But it's USoA ! - if you win military contract you live OK for few years and do not sell out suddenly like Sun - they did it just instantly after loosing military contract. Or maybe they (Sun managers) was preparing it 2 years ahead when dey bought Mysql :) Just theoretising :)
Strange timing - Qualcomm just announced hundreds of layoffs literally yesterday:
https://techcrunch.com/2024/09/19/chipmaker-qualcomm-lays-of...
selling intel to Qualcomm, would the equivalent of selling yourself to a vampire instead of selling blood.
the reason Apple ended up making their own wireless chips is due to Qualcomm
I wonder if Intel still have a suitable Arm architecture license that would be transferable? It seems unlikely without Arm approval, but a bundle could offset some of the cost Qualcomm might be thinking about, as that lawsuit might be getting too expensive, even for them.
For context: https://www.reuters.com/technology/arm-qualcomm-legal-battle...
A sale isn’t going to happen. Intel has had a rough quarter but their lunar lake launch looks promising, beating Qualcomm’s offerings on battery life and performance.
The issue long term is that Lunar Lake is built on TSMC, so Intel is netting a fraction per chip of what they'd make if they made it themselves.
Intel is currently investing $7B a quarter into getting their foundries competitive again, and it's not clear yet that they'll be able to really do so at scale. And even if they do, it's not clear whether those foundries can effectively serve customers that aren't Intel.
The reason people trust TSMC to make their chips is because TSMC isn't making a competing chip. If I come to an Intel Foundry with my design and work with them to spin up some new capability to get the features to work right, there isn't much of anything that stops Intel Chips from using that new capability to compete with me in a year.
Why Qualcomm’s Approach to Intel for Taking Over Semiconductor Industry Answer: Read this post: https://speadinfo.com/qualcomms-approach-to-intel-implicatio...
I really don't see it. Maybe a merger of some sort but still. Has Qualcomm anything to gain from taking intel, and likewise intel from being merged in?
They might just split and sell off the cpu and gpu divisions and keep the rest of Intel's (very wide) portfolio. Of particular interest might be the Intel foundries.
This would not be good for either party. And Qualcomm would only do this to kill x86 prematurely which I'm sure AMD will pick up on.
I want Apple to buy Intel so they can own their fabs. Instead of a billion going to TSMC, spend the billion to fix the fabs and the next several billion is profit.
US gov already put 20Bn into Intels fabs.
This is potentially bad news for open source kernels. They do lots of work there.
as outlandish as it sounds, there might be a better chance of this happening than the arm deal. the only reason being that both companies are registered in the same country, where the other parties would have a harder time blocking things.
Apple, c’mon. Get in there and just buy intel. Get the foundry working, spin out x86 into its own legacy, fabless business unit. Make your own chips, it’s only 100B!!
It’s very hard to get semiconductor factories to work. In almost all cases, it’s better to let the professionals at TSMC to do the fab than to rely on newbies to the foundry industry.
They already made the investment playing king maker for TSMC; I doubt they want to start that over for no real reason.
Qualcomm should be broken up.
Intel has focused on manufacturing efficiency for years now. Their innovation abilities have been lacking. A combination of Qualcomm and Intel will be a powerhouse. Intel as an entity will disappear but Qualcomm will be the stronger for it. I doubt Intel will go for it but I hope it happens.
> Their innovation abilities have been lacking
And Qualcomm's haven't? What did they really design besides the Snapdragon X Elite in the last 10+ years?
> but I hope it happens
So more industry concentration and even less competition would somehow be a good thing?
> What did they really design besides the Snapdragon X Elite in the last 10+ years?
The Snapdragon 8 series have been the flagship SoCs for non-Apple phones for years, that's why people had high hopes for the Snapdragon X on PCs
Snapdragons have mostly been reference ARM designs with Adreno bolted on being the big selling point, since it blows away Mali or whatever the reference design is today. They got that from AMD of all places, hence the name being an anagram of Radeon.
> The Snapdragon 8 series
They "just" are just using standard ARM cores which is hardly comparable to what Intel and Apple are doing (and they have been lagging behind Apple by a few years since forever; arguably Intel is closer to the M series than Qualcomm is to A series).
> Snapdragon X
I thought it's because they finally designed their own core, since ARM has been completely ignoring the laptop market and couldn't really offer anything?
> What did they really design besides the Snapdragon X Elite in the last 10+ years?
Lots of 5G modems
It's pretty easy to design the best 5G modems when you're are almost (effectively) the only company that's legally allowed to make them. They are basically a monopoly...
Intel would be doing much better as well if they didn't have to share x86 with AMD and could sue ARM into oblivion...
I imagine Apple would disagree.
We'll never really know though, will we? Or are you claiming that the endless patent related lawsuits (going both ways) had no impact on Intel's/Apple's efforts?
Is there a patent or something Qualcomm is trying to get from Intel? Seems like an odd acquisition otherwise.
As far as chip manufacturing they target different markets, I’d bet most IP isn’t transferable between the orgs anyway, especially since Apple bought Intels modem patents already.
A... single patent? No, there are licensing agreements for that.
China will block this deal from ever happening.
How exactly would a foreign power block two US companies from merging when it has no significant stock holding in either?
Same way Europe exercises outsized antitrust influence in the US: threatening to fine US entities, and if necessary deny them access to the European market.
It would be tricky, though, because my sense is that China still needs Intel/Qualcomm more than they need China. At the same time, it would be pretty deadly to be denied access to that market and your products subject to excess tariffs if imported by others.
To the poster, Europe does not exercise outsized antitrust influence in the US. Many of these companies have their tax residency in the EU. There is no need to "deny" anything. If the company gets fined and it refuses to pay the fine, EU seizes the money in one of many bank accounts in Europe.
Europe absolutely exercises significant antitrust influence upon US firms.
In practice, yes, as you point out: US firms must have assets in Europe to compete effectively.
But even if they didn't, Europe could deny access to the European market. So there is no reason to try and minimize surface in Europe. e.g. Apple has to comply with European antitrust rulings about app store access, even if Apple were to just sell their product to third party distributors in Europe and not have any presence in Europe.
> Europe could deny access to the European market
Not really, Apple would just get customers the same way they’re sold in any other part of the world that doesn’t officially have iPhones (i.e. Russia), the EU doesn’t have the authority to seize shipments purely based on a violation of the DMA.
Same way China blocked Intel from acquiring Tower Semiconductor.
Chinese regulators have a say because both companies do business in China.
Major markets like EU and China often do have a major influence on mergers even though the companies merging could be based in the US and most shareholding could be US.
I think it goes like this: The major market regulator could say directly/indirectly: Hey you can merge in the US ... but good luck operating in our geography in a frictionless manner if we are against your merger. As a regulator we can make life hell for you if you don't obey our anticompetitive laws. Since you derive a high percentage of your revenue/profits you must listen to us !
It all depends on the percentage of sales in the foreign geography. With EU/China it can be quite high -- especially for tech companies.
So yes, foreign powers can and often do block companies from merging.
And I think it ends like this:
CEO to board: We want to acquire $LESSER_COMPANY, but China opposes. We can go through, but we'll lose easy access to that market, and in net terms our valuation will drop.
Board: Forget it.
Like how UK blocked Microsoft's purchase of Blizzard.
What would motivate this
What’s bad for US semiconductor manufacturing (I.e. a poor and cash starved Intel) buys China more time to catch up to frontier fab tech. If Qualcomm buys Intel, the optimistic scenario (for the US) is a stronger domestic player.
> stronger domestic player
Is it though? Qualcomm is more likely to just strip Intel for parts than to turn it around and we'll just end up with more market concentration and less competition.
I did say “the optimistic scenario”, not “the only/likely scenario”
If that were true they’d denial-of-service every possible merger
Geopolitically you can't ixnay everything, you've got to pick your battles.
Yes. This would also give more volume for the existing Intel factories.
"So, why is Qualcomm under so much pressure? China is one significant reason. China is one of their critical worldwide markets and both the US and China governments seem to be increasing pressure on various US companies, including Apple, Qualcomm and Google."
https://www.rcrwireless.com/20230911/uncategorized/kagan-how...
From some perspective, governments of the big markets are like gangsters…
>Is this how arm vs x86 ends?
No... Intel isn't the only one that makes x86 processors and Qualcomm isn't the only one that makes arm.
Separately, Intel should sell them the flagging chips business and keep the fledgling foundries business separate.
> flagging chips business
Is it "flagging", though? Intel still seems to be pretty good at designing chips and their next gen laptop chips (made at TSMC) are allegedly more power efficient than the Snapdragon Elite (of course remains to be seen). It's the foundry that's dragging down.
If you look at the financial statements, it's quite the opposite however?
Their chips made on TSMC process are doing quite well and IFS has failed to secure worthwhile external customers and is losing money in their expansion hand over first.
IFS has just announced Amazon as a customer with a design on 18A. Microsoft is also expected to tape out one design. They’re not going to challenge TSMC this decade, but becoming the #2 fab 2030 is achievable.
The federal government is clearly ok with supporting a TSMC transition to the states. Something tells me though that they are willing to throw a lot of money at Intel if Intel is willing to fill the same niche that TSMC does currently.
That the chips currently produce more return than foundries is expected - it’s an established business. The foundries require much more up front investment. However the chips side of business has recently begun to show some cracks.
The foundries side of the business is in a different phase of life. It currently needs some TLC but has the potential to be totally ascendant at some point in the future. Assuming snapdragon is more interested in a chips business than a foundries business… it would just make sense to split them. There is tension with both under one roof as it is.
good joke
Yeah, Intel is really cheap. It would be a steal.
[flagged]