Apple aims to replace Broadcom, Qualcomm wireless chips with its own

Apple is said to be working to replace key wireless components in its devices with its own chips, a move that could see the Cupertino giant controlling most of the technology inside the smartphones and other mobile devices it makes. The iPhone maker is reported to be preparing internally produced cellular modem chips for the …

[Ali Dodd]

Walled Garden or Prison

Bluetooth and Wifi and Cellular built by Apple? First sign of issues and they'll be 'You are configuring your wifi/headphones/phone network wrong". We know they don't play well with others so this sounds very dangerous, Another reason for me to avoid fruity phones (your milage/koolaid consuption might vary)

[Knightlie]

Re: Walled Garden or Prison

"We know." Maybe provide a link to back up your paranoid FUD? Bonus points for finding one without the tired "kool-aid" insult.

[Richard 12]

Re: Walled Garden or Prison

How about the multiple times they've broken USB?

Sure, the El Capitan USB stack was technically permitted to change the things it did, but it wasn't what the authors of the standard intended and Apple were rules-lawyering when they insisted they weren't going to fix it.

Thousands of models of USB devices, including some really quite common professional USB sound devices ones suddenly stopped working with macOS.

A few of them got firmware updates to work around the change, but most did not - and even the ones that did required a Windows or Linux PC to install it, as macOS refused to see the devices at all.

More recently, both HDMI video and USB audio have had nasty issues.

[James O'Shea]

Re: Walled Garden or Prison

As expected. The anti Apple crowd here really is quite predictable.

[Captain Scarlet]

Re: Walled Garden or Prison

You rang?

I don't see a problem really, problem is Samsung make their own chips and use them. For me its no worse than what anyone else is doing, sad to say but I can't really be anti Apple here.

[Marty McFly]

Re: Walled Garden or Prison

Sheeesh. Next thing you know Apple will dump USB-C from their phones and force everyone to use a proprietary connecter.

Smarmy comment aside... You are right. If Apple doesn't make sure their chips are 100% compatible with the rest of the world, then it sets up an us vs. them compatibility play. However, it could be a very lucrative play if Apple is somehow able to do WiFi & Bluetooth better then that creates a market opportunity.

Either way it plays out, one thing is for certain.... It will cost the consumers more.

[DS999]

Re: Walled Garden or Prison

Apple can't make wifi that's not compatible with the rest of the world. What router are you going to connect to? Even if Apple started selling wireless routers again you only control the brand of router in your house, not at work, in airports, etc. etc.

Apple doesn't go their own way to be bastards, they go their own way when the existing standards suck. That's why Lightning was created, microUSB was a POS. USB-C is much better, but it came several years after Lightning. There would be no benefit to creating their "own" wifi, wifi already is overengineered as a standard for mobile devices. Who needs more speed than wifi 7 is already capable of going to a smartphone?

I'm guessing Apple wants to take wifi/bluetooth in house not because they want to "extend" them with proprietary stuff, or because they think they will save money. They want to do it to save power - or get better insight into and control over how much power it is using millisecond to millisecond. Currently they can precisely control the power the SoC and what's connected to it uses, but don't have as much control over cellular and wifi/bluetooth radios. While they have cost reasons to want to bring cellular in house, they don't have a problem with Broadcom's pricing like they do Qualcomm's.

Having wifi/bluetooth hardware and software under their complete control will allow it to be under the same global power management as the SoC. It will probably be integrated onto the SoC in its 2nd or 3rd iteration.

[hoola]

Re: Walled Garden or Prison

Wfif yes but Bluetooth is becomes very easy to force you to only use Apple accessories.

Given what we have seen in the past anything is possible, whether it is desirable is another thing but their customer base appears pretty resilient when it comes to being pushed to only buy Apple products with an appropriate premium on the price.

[Justthefacts]

Re: Walled Garden or Prison

Half true….whats the benefit to Apple to make the Bluetooth chip in the accessory? If all they want is lock-in, they can put anything over the top at the protocol level. A Bluetooth chip costs 50cents wholesale, but is internally complex and lots of R&D if you’re doing it from scratch - at least a $30M job to make something competitive.

If on the other hand, you’re suggesting a genuinely innovative Bluetooth replacement, the problem is that there isn’t a real drive to do so. The entire technical innovation in Bluetooth is about how to make really really cheap components do anything useful. It’s cost-engineering from top to bottom. That $30M program to develop a BT chip for 50 cents? If you want me to design you something that do the same job, even higher performance, but cost a dollar, I can do that for you for under $1M. At $4 pricepoint, I can design a POC on your desk for $50k in a few months, probably based on 6lowpan. But….that will never sell a single unit, because who is going to pay $4 for a $0.50 part?

Bluetooth is a remarkably crap spec, but it hits a market niche perfectly. There’s just no need to replace it.

[DoctorNine]

Re: Walled Garden or Prison

For most things that's true, but not high fidelity audio or virtual reality appliances. The limits of current BT data transmission will spark innovation here eventually. As others have said, Apple prefers NOT to just accept crap spec standards. They tend to consider alternatives which are cheaper, technically superior, or help with vertical integration in their ecosystem. Nothing to see here.

[Richard 12]

Re: Walled Garden or Prison

Apple prefers NOT to just accept crap spec standards. They tend to consider alternatives which are cheaper, technically superior, or help with vertical integration in their ecosystem.

That's the concern people have. Apple have spent a great deal of effort keeping users in their walled garden, and very little on interoperability.

[DS999]

Re: Walled Garden or Prison

Apple is willing to let its proprietary stuff become part of standards, which they would not do if they were interested in lock in.

They offered Lightning's physical spec to the USB standards org when they were designing USB-C. The physical design for their magnetic connection to align wireless chargers is part of the Qi-2 standard.

If they wanted lock in, they would keep stuff like that proprietary. They want stuff that's better.

I could easily see them extending bluetooth for better audio quality if the standard isn't already doing that itself. It would still be "bluetooth" and operate like bluetooth when you use an iPhone with non Apple earphones or use AirPods with a non Apple device, but could be better when used as a pair. If they did so, they would most likely offer that enhancement to the BT standards org.

[Justthefacts]

Re: Walled Garden or Prison

High-fidelity audio: depends what bitrate you are thinking about. If by “high-fidelity” you can still consider below 1 Mbps, the Bluetooth lower layers can do it, and all we are talking about is whether the codec is standardised and available as a hard macro in the chip. There are already good codecs, and aptx HD @576kbps is pretty close to what fits inside Bluetooth bandwidth.

Apple has no unique IP there, and if they did there’s no advantage against just dropping it into the standard and letting others implement it at zero cost to Apple.

If you want actually good audio quality however, the obvious answer is WiFi audio, at higher bitrate. The downside is higher power consumption and cost, but that’s always going to happen with the higher bitrate.

Bluetooth limited bitrate isn’t a flaw or technology limitation. What it actually is, is a good dividing marker that deliberately constrains applications. If you are really power and cost sensitive, then about 500kps is a good line in the sand for the absolute max to size your application at. It forms the target spec for the codec designers. If you want more than that, you are going to have to accept that power and cost aren’t your primary drivers, and some low-power WiFi variants are the technology you need.I don’t see a market niche between.

In completely other contexts, industrial control buses are usually limited to 1Mbps, eg CAN or 1553. There’s no market for 2Mbps or 5Mbps control buses, because broadly speaking you should be sizing your dataflow to live within 1Mbps. And if you can’t, there’s a good standard solution for you, and it’s called Ethernet.

[sreynolds]

Let me get my Kleenex

I don't feel any sympathy for the one trick CDMA pony nor Broadcomm. Both, in my opinion don't have that good a reputation. Unfortunately Apple is just slightly too small to be called a monopoly but there might soon be a compelling case to have it broken up. For how much longer is it going to be supplying the oil and the gas stations is anyone's guess.

[Justthefacts]

Re: Walled Garden or Prison

A market opportunity for….what? Apple selling WiFi access points? They used to do that, if you remember. The issue is that there’s no money in it. Access points are a commodity product, cost £100, you only sell one per household, and they last at least twice as long as the phone . Remember, Apple are choosing to exit the market for smartphones under £500. Worst of all, there’s lots of really really clever stuff that you have to master to make a good WiFi chip. So….high technical risk, large investment, ongoing (because after WiFi 7 there’ll be WiFi 8), low-margin return. Doesn’t sound like a great investment for Apple to me.

[Johnb89]

Re: Walled Garden or Prison

Given that existing bluetooth implementations everywhere are pretty shit, could this be worse? Yes, but anyway, the bar is low.

[Justthefacts]

Re: Walled Garden or Prison

I used to think Bluetooth was shit, until I really sat down and engineered with it. What BT actually is, is the right solution for cost-engineered and power-critical uses eg powered by a coin cell battery. It fits the niche.

If you want something better, you don’t want something that’s just a bit better. You’re playing in a different part of requirements space, and likely low-power WiFi is the correct solution for you. A WiFi chip is “only” about 8x more expensive ($4 vs 50cent). Scaling linearly on Mbps, just how keen are you to engineer for the incredibly specific design space between say 2Mbps and 10Mbps costing $1-2?

[A Non e-mouse]

Qualcom & Broadcom have both been accused of price gouging and arrogance. If they weren't so greedy, Apple may not have decided to look at designing their own chips.

[Gene Cash]

price gouging and arrogance

Good thing Apple's never done any of that, eh?

[DS999]

I don't think I recall

Apple ever complaining about Broadcom's unfair pricing. They've had tons of problems with Qualcomm and have gone to court with them, but AFAIK they have never sued Broadcom or been sued by them over wifi/bluetooth chips.

[Anonymous Coward]

Re: I don't think I recall

Apples problem with Qualcomm is having to pay the price asked. I can see little that is unfair in Qualcomm providing the core underlying technology that makes a mobile phone/data terminal possible, and expecting a decent return. Ask yourself how much an iPhone without a radio would be worth, and how many dollars difference from what they sell for with it.

Apple simply does not want to give anyone anything except a tiny sliver of the pie that's barely enough to keep from starving.

Very few companies have the ability to stand up to them. Qualcomm spent a lot of money and time to build "defense in depth" against the likes of Apple.

Now if you are a Qualcomm competitor - yeah you have very good reason to be unhappy.

Apples model is to play competitors off against each other until they sell at, or often below cost. This often results in Apple (or it's captive supplier) picking up all the investment and capital expenditure for free. i.e. effectively stealing it.

The only defense suppliers have against that is create an effective monopoly - just like Apple.

[I have worked for two companies involved with Apple. One supplied a component for Apple, and it went as above - within 3 years they had to sell the factory, and it still makes parts for iPhones.

At the other one we declined to get involved as it was clear they would simply steal our tech after we had made no profit developing it for them. (Our competitor got sucked in to working with them, and was out of business in two years, their top management couldn't resist the pull of working for Apple, even though lower down expected it to go the way it did - Apple were pretty blatant)]

[doublelayer]

Re: I don't think I recall

"Apples model is to play competitors off against each other until they sell at, or often below cost."

Welcome to economics, where the method you use is to find the place that can give you what you want for the lowest cost to you. I'm not sure why you expected something else, but if you want to sell components and not below cost, stop turning the price down when Apple asks. You don't need a monopoly to avoid making stupid economic decisions with your own company, and many other component manufacturers manage to supply Apple or other large manufacturers and make a profit. If it turns out that your competitor can make a profit at a price where you can't, then that's your problem and you might want to do something about it.

[Anonymous Coward]

Re: I don't think I recall

The voice of no experience speaks.

[doublelayer]

Re: I don't think I recall

Maybe you'd care to tell me what my lack of experience, to which I confess, has failed to teach me? If you've got tales of how everybody in component manufacturing expects to lose money when they sell their products, then I'm happy to learn. In my experience of other things, there are a few reasons companies might choose to sell at a loss, and all of them are risky to the company and require them to have planned a lot for it. History is full of examples of companies who thought that being cheap would win in the long term only to find that they don't have enough money to take that as far as they need to. Unless you have a foolproof plan for why selling at a loss is good for you, don't do it, and if you think you have one, it still might not be one.

[hoola]

Re: I don't think I recall

Exactly what Apple do with all their products.

iStuff has always been priced for what the market will stand and like most other premium products, pricing is the most important aspect as it provides exclusivity or perceived quality.

[mevets]

Re: I don't think I recall

Is there some sort of collective that sell their electronic goods for cost plus a reasonable margin?

That is the business model of the Mennonite and Amish communities; but is there a technology equivalent?

Do Samsung, LGE, ... sell for a reasonable price because they are community minded, or is viability?

The problem with price is not the producer that sets it. If there is a problem, it lies with the consumer who accepts it.

[Snake]

Re: Apple's pricing problem with Qualcomm

Designing their own chips will help Apple with chip pricing but will not completely eliminate the problem, as Qualcomm owns the much of the technology of cellular itself. Apple will still have to license the technology from Qualcomm...as if Qualcomm will just roll over and deliver a license agreement to Apple that is as cheap as Apple thinks they will get.

Any loss in chip revenue will be happily rebalanced with a very lucrative, nicely padded, patent licensing agreement.

[DS999]

Re: Apple's pricing problem with Qualcomm

Sure Apple will have to pay Qualcomm's licensing but Apple has been building their own wireless patent portfolio so Qualcomm will have less leverage over time.

Apple also won't have to pay to license patents they don't need. For instance the current chipsets include support for Qualcomm's proprietary 3G CDMA technology that's being phased out in the US - and which isn't a standard so FRAND rates for patents don't apply. Apple's modem will most likely be LTE and 5G only (yeah I know 3G and even 2G are still "in use" in Europe but that's mainly as support for legacy devices like alarms so most spectrum has been or is in the process of being refarmed for LTE/5G so the chance of being in a place where 2G/3G is the only connection available is slim and becoming slimmer by the day)

Not having to license patents that apply only to Qualcomm proprietary stuff or older tech will reduce licensing fees.

[A Non e-mouse]

Re: Apple's pricing problem with Qualcomm

Weren't Qualcom insisting that if you wanted their latest 5G chips you had to buy their CPU too?

[Anonymous Coward]

By greedy you mean "making a profit Apple could be making"

Apple has a quite unusual business problem. They don't make many products. They have already maxed out the prices they can charge.

So how do they invest to increase profit?

That is why they seek to own every part of the value chain i.e. all the chips and technologies, store retailing margins etc

It is actually a very disciplined and self aware business that understands what it isn't.

It does however suck to be an Apple supplier who has any IP value add at all, because it is only a matter of time until Apple wants it for themselves

[mevets]

Don't make many products?

That is a key insight. They only make:

- computers 4 flavours

- laptops 8 flavours

- tablets,12 flavours

- cell phones, 16 flavours

- direct ear buds, 2 flavours

- bluetooth ear buds, 2 flavours

- bluetooth other head phones (beats), ? flavours

- bluetooth can head phones

- home speakers, multiple flavours

- tv streaming box, multiple flavours

- wheels for big ass computers (???)

- monitors

- stands for monitors

It is pretty dramatic, compared to say 1980:

- computer

or 1990:

- computer, 2 flavours

or 2000:

- computer, 3 flavours

or 2010:

- computer, 4 flavours

- mp3 player

- cellphone

,,,,

There is a lot to slag apple for; maybe most of all for having too many products.

[Oglethorpe]

Re: Don't make many products?

16 'flavours' of phone? Really?

[doublelayer]

Re: Don't make many products?

I'm not sure where you got your numbers, any of them, from. You say they have eight flavors of laptop. They have five models that they make, and one of those is an old version which they still manufacture. Likewise, they only have eight iPhone models that they still sell, three of which are older versions. In 1980, not only did they have three computer models (2 plus, 3, and 3 plus), but they also had a display and a printer. By 1990, they had lots of models of Macs and peripherals.

The distinction is between Apple and comparable companies. How many models of Mac laptop can you buy? Air, Pro 14 inch, and Pro 16 inch, with a new model of each one every year or so. How many models of Dell laptop can you buy? I don't know because they seem to have an infinite supply with names like Inspiron 15 5000 4I2SNZ7, but I can guarantee it's more than three. This can bring positives, such as it being easier for Apple to standardize on parts, and it can bring negatives, like not having a choice of laptop at any price level. This is something where Apple has taken a different approach to some other large computer manufacturers, though I know of a few small companies that similarly aim for a small set of models.

[mevets]

Re: Don't make many products?

Yeah, sorry, I didn't realize there was a difference between "make" and "manufacture". But from my reading, I can order a:

MacBook Air M1

MacBook Air M2

MacBook Pro 13" M2

MacBook Pro 14" M1-Pro/8+14

MacBook Pro 14" M1-Pro/10+16

MacBook Pro 16" M1-Pro/10+16

MacBook Pro 16" M1-Max/10+32

MacBook Pro 16" M1-Max/10+32 (*)

IMac M1 M1/8+7

IMac M1 M1/8+8

IMac M1 M1/8+8 (*)

Mac Mini M1/8+8

Mac Mini M1/8+8 (*)

Mac Studio M1-Max/10/24/16

Mac Studio M1-Max/20/48/32

Mac Pro 8086/8-28 + 1.5TB

Mac Pro 8086/8-28 + 1.5TB (*)

Which is actually 17. Sorry, ran out of fingers. The ones with (*)'s seemingly have similar base specs, but are separate modes because < who cares >; they are separate devices.

I can only imagine what segregates one dell laptop from another; perhaps the box, or the colour coded audio and video connectors?

Older versions? So it only counts if a company makes a new product in the last 1, 3, 6, 12 months? In that case Dell is shit out of luck. It is still selling ancient shit with whichever tray it can beg off of intel. It arguably has one product.

Of course it is terrible of apple to have some sort of standardization. I suppose that is what happens when you control manufacturing quality instead of slapping a badge on whatever piece of whitebox crap you can buy from alibaba.

[Oglethorpe]

Re: Don't make many products?

I'd say if your new product is a clear replacement for the old, it doesn't constitute multiple, parallel products. Lenovo have become sloppy with their naming as of late but, going back, the X series Thinkpads weren't twice as diverse in their offerings when the X280 came out but the X270 was still available.

On that note, I think you counting configuration options really shows just how locked in you are with Apple products, especially the new CPU models, with their soldered storage. By the same measure, Lenovo makes 1536 'flavours' of the gen 9 X1 Carbon (ignoring software configurations) and it would be a heck of a lot more if they hadn't embraced the devilry of soldered RAM and integrated batteries.

[43300]

Re: Don't make many products?

A different CPU / RAM / storage provision really doesn't count as a separate model - it's just a variation.

For example if I want to buy a Dell Latitude 5430 I can buy it with various options as regards the above. But it's still a 5430.

I make your list above as 9 actual models.

[43300]

Re: Don't make many products?

Indeed! Compare them to say, Dell. Ok, they don't make phones and only a limited number of tablets, but they make far more computers (actually different models, whereas Apple are a few variations on the same basic model). They also do servers, switches, storage, etc. In the computer market they also work at all price levels, whereas Apple basically only do the upper price bands.

[anthonyhegedus]

Apple make hardly any products at all. Just all the Apple ones and fuck-all else

[Ace2]

They bought the remnants of an Intel group…

If designing such a chip were a sure thing, Intel wouldn’t have failed miserably at it. I’ll believe it when I see it.

[karlkarl]

Currently macOS Ventura is the first macOS that you can't finish installing from scratch without first going online and connecting to Apples servers.

Apple states that this isn't DRM but instead is required in order to grab 3rd party firmware that can't be distributed with the OS.

Hopefully by having "all Apple" licensed firmware, we can solve this bug because frankly this internet requirement is sodding ridiculous.

[Anonymous Coward]

Funny how Linux and Windows don't do that, yet support far more hardware out of the box.

[Anonymous Coward]

*cough* Windows Activation *cough*

[Oglethorpe]

Still usable. If the activation servers went permanently offline, you'd be nagged and unable to change wallpaper but, otherwise, have a functional system.

[43300]

Many computers have the Windows key in the UEFI, so you can install it and activate Windows 10 at least without an internet connection. If a device isn't supported by the generic Windows installer (as is the case when clean-installing some of the latest Dell laptops - SSD not recognised), the driver can normally be loaded in from a USB stick at the appropriate stage of the intstaller.

I've not tested the latest version of Windows 11, but I believe the same applies to at least the Pro version (not sure about the Home one).

[Jim Willsher]

and so Qualcomm is expected to continue to supply all the 5G chipsets for iPhones in 2022.

2023?

[DS999]

Rumor has it

The first iPhone with the Apple designed cellular will be the next iPhone SE coming out in spring 2024. Presumably they want to "test" it on a lower volume/less important product first, and if that goes well it would be used in the 2024 iPhone 16.

[mevets]

Cutting off nose to spite face

I am sure Qualcomm considered cutting of their supply of money in order to spite Apple for daring to cut off their supply of money. That way, the court could end up giving all of Qualcomm's IP to Apple for free.

[DS999]

Re: Cutting off nose to spite face

Qualcomm could refuse to sell modems to Apple but cannot refuse to license their patents used in cellular standards to Apple. Their shareholders would probably revolt though, giving up billions a year in revenue just because your customer is making changes that will reduce the amount of revenue you make from them is generally a pretty poor business strategy.

If there were no other companies in the world that could provide a modem to Apple maybe they'd consider it as a way of squeezing them, but Apple could call Samsung, HiSilicon, Mediatek or various other companies that sell cellular modems. Each of them has its own problems (i.e. Apple prefers not to do business with Samsung where it can be avoided, difficulty in getting modems from Chinese companies like HiSilicon in the current trade war, etc.) but they would find a way to make it work if Qualcomm decided to go thermonuclear on them.

[mevets]

Re: Cutting off nose to spite face

Yeah, sorry I refuse to augment sarcasm indicators for the humour impaired.

[systemBuilder22]

It's hopeless.

I am sure that Apple has absolutely NO IDEA what its doing here. It takes 1000 hours of drive-around testing in the worst RF environment(s) in the world to optimize the cell phone performance of a Qualcomm phone, and that is done in San Diego, the #2 worst environment in the world after Hong Kong, and it is done when the cellular system is designed. Qualcomm never discloses the control parameters of their phone algorithms which govern handoff. And that, folks, is why nobody makes phones that work as well as Qualcomm, not even Apple ...