Catch up on stories from the past week (and beyond) at the Slashdot story archive

 



Forgot your password?
typodupeerror
×
Communications Apple

Apple Starts Work on Its Own Cellular Modem, Chip Chief Says (bloomberg.com) 51

Apple has started building its own cellular modem for future devices, a move that would replace components from Qualcomm, Apple's top chip executive told staff on Thursday. From a report: Johny Srouji, Apple's senior vice president of hardware technologies, made the disclosure in a town hall meeting with Apple employees, according to people familiar with the comments. "This year, we kicked off the development of our first internal cellular modem which will enable another key strategic transition," he said. "Long-term strategic investments like these are a critical part of enabling our products and making sure we have a rich pipeline of innovative technologies for our future." A cellular modem is one of the most important parts of a smartphone, enabling phone calls and connection to the internet via cellular networks. Srouji said the $1 billion acquisition of Intel's modem business in 2019 helped Apple build a team of hardware and software engineers to develop its own cellular modem.
This discussion has been archived. No new comments can be posted.

Apple Starts Work on Its Own Cellular Modem, Chip Chief Says

Comments Filter:
  • Intel Was Elop'ed (Score:5, Interesting)

    by organgtool ( 966989 ) on Friday December 11, 2020 @03:27PM (#60820102)
    Due to their patent disputes with Qualcomm, Apple contracted Intel to create cellular modems for the iPhone. Apple wasn't satisfied with the quality of those modems, so they subsequently went back to Qualcomm. During that time, Intel wasn't able to acquire any other significant customers for their modems, so they were basically forced out of the modem business. Apple then bought the remains of Intel's modem business for pennies on the dollar to get access to Intel's R&D and probably so that they could also get the patents to force cross-licensing deals with the other manufacturers of cellular modems. While I don't believe it was Apple's plan from day one, they effectively got Intel to pay huge upfront costs of R&D for cellular modems, ditched them shortly thereafter, and bought the remains of the business at a steeply discounted price. That sounds a hell of a lot like how Microsoft pushed to get Stephen Elop in charge at Nokia, ran it into the ground in spectacular fashion, and paved the way for Microsoft to buy Nokia's smartphone division for an insanely cheap price. The main difference is that Microsoft's actions seemed more premeditated whereas Apple's actions seemed slightly more genuine upfront, although they may have planned on buying Intel's modem business as a contingency plan from the outset if the business relationship didn't pan out.
    • Then the crux would be whether Apple's negative judgment and rejection of Intel's modem chip was genuine, or contrived to cause the failure of that business. If Intel's agreement with Apple was not exclusive yet Intel failed to win other customers I guess that is our answer. I just hate to see Intel go downhill. Apple's M1 seems like quite an achievement but it doesn't help anybody who doesn't want to buy into their walled garden.
      • I agree with just about everything you've said, although Apple Silicon may actually help with people such as me (and presumably you) that don't want to be in a walled garden - it should put even more pressure on existing chip manufacturers to create chips that consume far less power. Who knows, one of the big chip manufacturers may even end up eventually supporting RISC-V as an alternative to avoid paying for ARM licenses. Yes, I know that is a pipe dream but any potential for even a slight increase in RI
      • Re:Intel Was Elop'ed (Score:5, Informative)

        by bill_mcgonigle ( 4333 ) * on Friday December 11, 2020 @04:16PM (#60820254) Homepage Journal

        > whether Apple's negative judgment and rejection of Intel's modem chip was genuine, or contrived

        It was genuine. Qualcomm smacked them with a ridiculous patent that IIRC came down to pipelining "ON A CELLULAR MODEM" - totally obvious to anybody who's been paying attention since the 80's.

        So Intel's modem had to wait for an ack before sending more data or something like that and the performance was just terrible.

        My memory is fuzzy on the implementation details but regardless Apple knows about this so they must intend to challenge the patent or hit Qualcomm so hard with other patents that their usorous heads will spin.

        Don't make me cheer for Apple but Qualcomm is just the worst.

      • Then the crux would be whether Apple's negative judgment and rejection of Intel's modem chip was genuine, or contrived to cause the failure of that business.

        Let’s put it this way, not long after Apple ditched Qualcomm, they also killed the field test mode where you could actually see the received signal strength in dBm and cut off any developer access to it as well. The problem was the intel chips suck and phones would perform around 3dBm worse than Qualcomm ones, with some outliers even being 5+. Thus you could show beyond a doubt the newer phones were worse than the old ones so it was quickly axed.

      • Since Intel couldn't find anyone at all to buy their modems, it seems reasonable to conclude apple's rejection was on the merits too.

        • by AmiMoJo ( 196126 )

          Apple actually crippled the Qualcomm modems to make them as slow as the Intel ones, so that they could fit both to iPhones and people wouldn't complain that they got the slow model. Of course there were power consumption differences too, which they denied.

      • The M1 is used ATM in MacBook Airs and MacBook Pros, and later probably in Desktop Macs: which have no walled harden and never will have one

        • The fact that you can only get the M1 in Apple products is part of the wall in itself. If Intel died but American chip design lived on in Apple and I could buy any of a wide variety of laptops or desktops with an M1, the loss of Intel wouldn't be so bad.
    • âoeIntel needs no help running something into the ground.â
      - Will.i.am straddling the Intel OCC bike

    • I think they paid what the division was worth; in fact I think they overpaid.

      Apple's patent fight with Qualcomm was about breaking Qualcomm's monopolistic licensing practices. It's not correct to say that Intel couldn't find buyers for their modems; their modems didn't work. They would delay the release of Apple's 5G iPhones at least a year behind any Android ones particularly Samsung to get the tech right, and they couldn't have that. Intel wasn't going to acquire any other significant customers for

  • Must be at least a 7G one.
  • by mveloso ( 325617 ) on Friday December 11, 2020 @04:00PM (#60820196)

    If Apple is as good as they seem to be, their modem chip will be three times as fast as the Qualcomm one and use an order of magnitude less power.

    • Well, it’s going to take some serious dedication to make it worse than the intel ones.
    • And then drop to around half the effectiveness of a Qualcomm chip after 4 minutes of usage...
    • Apple will figure out how to integrate the redundant parts of the modem into their main CPU bus. Since 5G means the data will be a rush like a mainline injection of heroin not a puff on a bong, 5G is going to get bottlnecked exiting the radio.

      SO what's going to happen is Qualcom is going to start integrating CPUs into their Radios for the same reason. And thus Qualcom wins the whole match, and apple has to forget the M1 arms and go with whatever qualcom cpu does. ... or they integrate the radio into their

      • SO what's going to happen is Qualcom is going to start integrating CPUs into their Radios for the same reason.

        My guess is Apple can get a good modem in the CPU before Qualcomm can get a good CPU into the modem...

    • by AmiMoJo ( 196126 )

      Highly unlikely. And Apple isn't that good, they haven't produced any magical orders of magnitude improvements. Maybe an order of magnitude more hype.

      The power consumption of models is down to two things. There is the DSP and software stack, but the bulk is the RF side. And the RF side can only really be optimized by having better antennas or more efficient electronics at the silicon level. Antennas are high end physics and already close to the known limits of science, and efficient electronics are really d

  • Well of course the modem is one of the most important parts of a phone. Better modems with superior signal to noise are what let manufacturers hide substandard low gain antennas in the phone for a simultaneous cost savings and “sleek” look. I mean, think of the disaster needing a cell phone to actually receive the best range - massive clunker gross antennas that stick out! Thank god my phone is sle... *end carrier*
    • by bn-7bc ( 909819 )
      Or rgey could, combine a slightky better performing radio with a better antenna and claim better 5G performance then 95% (allways good to have s margin of error) of premium smartphones, pr loves that kind if thing
      • As you can basically get anything conductive to resonate in the cellular bandwidth you want by appropriate application and locations(s) of inductance and capacitance then for any particular design a larger area means more captured energy and thus a higher overall gain. This is true even if it’s a phased array of antennas distributed through a phone.
  • You wants the short but memorable relations. We will have fun this night! I'm waiting >> https://kutt.it/RILV47 [kutt.it]
  • This seems to be meant to pressure Qualcomm (stock etc.) more than reality. If true, Apple would simply do it in Apple fashion -- not say an effing word about it until it's ready and then drop it like a bomb on Qualcomm. This seems more intended -- ie. more or less "publicly" announced -- to gain some kind of concessions from Qualcomm.

"One lawyer can steal more than a hundred men with guns." -- The Godfather

Working...