Eshu Marneedi

Mark Gurman, leaking the timeline for Apple’s custom modems at Bloomberg:

Apple Inc. is preparing to finally bring one of its most ambitious projects to market: a series of cellular modem chips that will replace components from longtime partner — and adversary — Qualcomm Inc.

More than half a decade in the making, Apple’s in-house modem system will debut next spring, according to people familiar with the matter. The technology is slated to be part of the iPhone SE, the company’s entry-level smartphone, which will be updated next year for the first time since 2022…

For now, the modem won’t be used in Apple’s higher-end products. It’s set to come to a new mid-tier iPhone later next year, code-named D23, that features a far-thinner design than current models. The chip will also start rolling out as early as 2025 in Apple’s lower-end iPads…

In 2026, Apple looks to get closer to Qualcomm’s capabilities with its second-generation modem, which will start appearing in higher-end products. This chip, Ganymede, is expected to go into the iPhone 18 line that year, as well as upscale iPads by 2027…

In 2027, Apple aims to roll out its third modem, code-named Prometheus. The company hopes to top Qualcomm with that component’s performance and artificial intelligence features by that point. It will also build in support for next-generation satellite networks.

In the middle of this timeline — which, alas, isn’t written in a nice bulleted or ordered list like Axios, but in Bloomberg’s house style — Gurman slips in this very Bloomberg detail:

Qualcomm has long been preparing for Apple to switch away from its modems, but the company still receives more than 20% of its revenue from the iPhone maker, according to data compiled by Bloomberg. Its stock fell as much as 2% to a session low after Bloomberg News reported on Apple’s plans Friday. It closed at $159.51 in New York trading, down less than 1%.

I’ve attributed most of Intel’s post-2020 slump to the loss of Apple as a partner. People like to claim Apple wasn’t an important or large customer because the number of Mac units Apple sells each year pales in comparison to Intel’s other clients, but the number of end-user units is irrelevant. It’s undoubtedly true that Apple paid Intel lots of money and was one of its most important customers. Apple was always reliable: it wanted the latest Intel processors each year in Macs and wanted them quick. When Intel was behind or underwater, it could always have confidence that Apple would be a reliable, recurring source of income. In 2020, that changed, and now the company is doing so poorly that it fired Pat Gelsinger, its chief executive since 2021, as a vote of no confidence, per se.

It’s not wrong to argue that the primary reason for Intel’s latest downfall is that it never developed processors for smartphones, ceding that ground to Qualcomm and Apple, but I have a feeling Intel would’ve been fine if it still had Apple as a partner. It lamented the loss of Apple — sourly¹ — because it realized how bad it was right then that it lost such a reliable buyer. Partners come and go all the time, but if Intel felt it wouldn’t hurt after Apple’s departure, it wouldn’t cook up attack ads featuring Jason Long, who famously played the Mac in Apple’s clever “Get a Mac” marketing campaign. That was a move born out of sheer desperation; Intel has been desperate since 2021.

Now, back to Qualcomm. Before this story, I was under the assumption that Qualcomm made the vast majority of its revenue from its mobile processor business — the popular Snapdragon chip line. That majorly composes Qualcomm’s business, but it isn’t the vast majority. Either way, I severely underestimated how much it would hurt Qualcomm to lose Apple as a partner. Qualcomm makes more than 20 percent of its total revenue from just one company, one trading partner. Because of that, I think I’m ready to make a rather bold prediction: 2026 will be to Qualcomm what 2020 was to Intel. Once Apple starts shipping its own modems in the standard and Pro-model iPhones, it’s game over for Qualcomm. Apple wasn’t Intel’s biggest customer, but it was strategically the most important, and I feel the same is true for Qualcomm.

But clearly, Apple believes building modems is much harder than designing Arm-based microprocessors, as evidenced by how long it’s taken Apple to build its own modems. Apple has been trying to compete with Qualcomm since the two companies got into a spat back in 2018 when a Chinese court ruled Apple infringed on Qualcomm’s patents. Whereas Intel and Apple have always historically been friends, the same can’t be said for Qualcomm — the two companies have been in fierce competition since that kerfuffle, and it’s going to come to a head in just a few months when Apple launches its first modem, ideally not even to much fanfare. If the next-generation iPhone SE is just as reliable as previous models, Apple has a winner, and Qualcomm will inevitably sweat.

To make matters even worse, Qualcomm is currently embroiled in a lawsuit with Arm, which licenses its designs to Qualcomm, which then modifies them and fabricates (makes) them with Taiwan Semiconductor Manufacturing Company. Arm has already canceled Qualcomm’s license to produce chips with Arm designs, and if it wins in court this month, that cancelation will be set in stone. The reaction to this problem has mostly been tame — tamer than I believe it should be — because the industry is sure that Arm is shooting itself in the foot by making enemies with arguably its most important customer, but this is bad for Qualcomm, too. It’ll probably switch over to using the RISC-V (pronounced “risk-five”) instruction set, but that’s a drastic change. Add this Apple deal to the mix, and the company is in deep trouble.

It’s possible Qualcomm weathers the impending storm better than Intel because it’s arguable that Qualcomm is in a much better position financially. Qualcomm chips aren’t behind — they’re competitive with the very best iPhone-grade Apple silicon, and they’re popular amongst flagship Android manufacturers. The same couldn’t be said for Intel back in 2020, which was slipping on its latest processors and had fierce competition from Advanced Micro Devices. But the relatively recent talk about Qualcomm potentially buying Intel seems almost nonsensical after Gurman’s Friday report, and the chip design market seems more volatile than it ever has in recent history.

Also from Gurman today:

Apple Inc.’s effort to build its own modem technology will set the stage for a range of new devices, starting with slimmer iPhones and potentially leading to cellular-connected Macs and headsets.

According to this report, Apple’s main concern for bringing cellular connectivity to the Mac is space, and that’s addressed with its own modems. Initially, this struck me as unbelievable since Mac laptops ought to have tons of room inside for a tiny modem that fits even in the Apple Watch, but perhaps an iPhone-caliber modem isn’t powerful enough to handle the networking requirements of a Mac? I’m really unsure, but a bit of me still believes it’s feasible to stuff a Qualcomm modem in a MacBook Pro, at least. In any event, I’m a fan of this development, even as someone who doesn’t use their Mac outside, in the wild, very often. When I do, however, I typically rely on iPhone tethering, and that’s just a mess of data caps and slow speeds. I’d love it if I could tack on a cheap addition to my existing iPhone cellular plan for a reasonable amount of data on my Mac each month.

I understand the appeal of a cellular-connected Apple Vision Pro less, but if it works, it works. Either way, Qualcomm is screwed since not only is it not receiving the mountain of reliable cash that comes with an iPhone deal, but it’s also not able to profit from Apple’s new cellular ventures.