Three years is a long time for a “new” watch to run on old silicon. That’s roughly how long the Apple Watch has been stuck at S9-level performance — through the Series 10, the Series 11, the Ultra 2, and the Ultra 3 — and 9to5Mac reported on August 21 that the plateau is finally ending. The Apple Watch Series 12 and Ultra 4 are getting a genuine chip upgrade, the first one since the S9 landed in 2023, and it’s arriving at the exact moment watchOS 27 needs the horsepower to run Siri AI on your wrist.

That timing isn’t a coincidence. It’s the whole story.

Quick Verdict

	Details
What’s new	New S-class chip with real CPU gains, first since the S9 (2023)
What it replaces	The S10, which shipped unchanged in both Ultra 3 and Series 11 last year
Why now	watchOS 27’s on-device AI — a dedicated Siri app, redesigned coaching features — needs chip headroom the S10 doesn’t have
Expected event	September 8 or 9, 2026, alongside the iPhone 18 Pro lineup
Rumored pricing	Series 12 starting around $429, up from Series 11’s $399
Source	9to5Mac

Best for: Anyone weighing a new Apple Watch against waiting for the Ultra 4 who wants the actual performance story, not just the sensor rumors. Skip if: You need confirmed pricing or a ship date. None of this is official until Apple says so on stage.

What’s Actually New in the Series 12 and Ultra 4 Chip?

Here’s the condensed version, since this is what people are typing into search right now:

1. A new S-class chip lands in both models — Series 12 and Ultra 4 — with real CPU gains, not just a clock-speed bump dressed up as a new generation.
2. It’s the first true performance jump since the S9, which debuted with the Series 9 back in 2023.
3. The S10 — used in Series 10, then carried into Ultra 3 and Series 11 last year — didn’t meaningfully move performance at all. Three consecutive Watch launches, same ceiling.
4. The upgrade is timed to watchOS 27, which ships next month with Siri AI as its headline feature.
5. Nothing is priced or dated by Apple yet. This is pre-release reporting, not a keynote.

That third point is the one worth sitting with, because it’s a bigger admission than it sounds.

The S10 Was a Placeholder, Not an Upgrade

We flagged the S10 back in April as one of the Ultra 3’s selling points — “faster processing, better battery management,” which is the framing Apple itself leaned on at launch. Turns out that framing did a lot of work. 9to5Mac’s reporting says the S10 offered no real performance gains over the S9 that preceded it. Same ceiling, new label.

So the actual chip timeline looks like this: S9 ships in 2023 with genuine CPU gains. Series 10 gets the S10 in 2024. Ultra 3 and Series 11 both ship in 2025 with that same S10, unchanged. That’s three straight product cycles running on essentially 2023-era silicon, dressed up in three different generational names. Our own coverage of the Ultra 3 called it “a skip year” back in April, before this chip reporting existed — turns out the instinct was right for a reason we didn’t have data on yet.

Apple can get away with that kind of stall when a smartwatch chip mostly needs to run a UI and track a workout. It gets a lot harder to get away with when the chip is also supposed to run a language model.

Why watchOS 27 Is the Real Trigger

watchOS 27’s public beta shipped in July with three features that lean directly on processing power in a way past watchOS updates haven’t:

• A dedicated on-watch Siri app. It sits permanently in the center of a new Dynamic App Grid — you open it by pressing the Digital Crown — and it keeps a running log of your Siri conversations instead of treating every request as disposable.
• A single-tap gesture for controlling the Smart Stack one-handed, an evolution of the Double Tap gesture that requires the watch to recognize a subtler motion in real time.
• The Dynamic App Grid itself, which rotates five of its app slots based on time of day and usage patterns instead of showing a static layout.

None of that is dramatic on paper. All of it is inference happening on-device, continuously, on a chip that was never built for the job. 9to5Mac’s own beta testing flagged that Siri “has been a bit less responsive than I’d hope” running on current hardware — which is a polite way of saying the S10 is the bottleneck, not the software.

That’s the piece that connects back to Workout Buddy, the AI coaching feature we tested extensively when watchOS 26 shipped it in March. Workout Buddy already runs Apple Intelligence on-wrist to generate spoken coaching mid-session, and it’s genuinely good at that one job. watchOS 27 is the version where Apple stops keeping AI contained to a single workout feature and starts building it into the operating system’s core interaction model — Siri, the app grid, how you navigate the watch face itself. You can’t do that on a chip that’s been flat since 2023 and call it a serious product decision.

Apple’s broader AI health ambitions took a hit earlier this year when Project Mulberry got scaled back after Eddy Cue reportedly decided it didn’t clear the bar Oura and WHOOP had already set. The watch-based AI push — Workout Buddy, now a rebuilt Siri — is the version of that ambition Apple kept. A chip upgrade that exists specifically to support it is Apple putting hardware weight behind the strategy it didn’t abandon.

When and How Much

Apple’s fall hardware event is expected on September 8 or 9, 2026, with the Series 12 and Ultra 4 debuting alongside the iPhone 18 Pro lineup — the standard September pairing Apple has run for years now. Labor Day falls on September 7, and Apple has historically kept a few days of runway after the holiday, which is why Wednesday the 9th gets more weight in most of the reporting than the 8th. Nothing here is locked until Apple sends the invite.

Pricing is rumor-tier at this point, but the number showing up most consistently is around $429 for the Series 12, up from the Series 11’s $399. That tracks with comments Tim Cook made on a May earnings call about rising RAM and component costs pushing “an increasing impact” across Apple’s lineup — the Watch has dodged that pressure so far, and a real chip upgrade is exactly the kind of change that could end the exemption.

What This Changes About the Wait-or-Buy Math

Back in April, we ran the wait-or-buy calculus on the Ultra 4 based on a firmware teardown that surfaced Touch ID (codename “AppleMesa”) and a doubled sensor component count. That post didn’t have this chip story, the narrowing September date, or the AI framing to work with — the leaks at the time were about hardware, full stop.

This changes the calculus, and not subtly. Touch ID and better sensors are nice-to-haves you could reasonably skip if you don’t care about biometric authentication on your wrist. A chip that’s the difference between Siri actually working versus lagging behind your request is not a nice-to-have if you plan to use any of watchOS 27’s AI features regularly. If you bought an Ultra 3 or Series 11 in the last year expecting the “usual” annual refresh to be skippable, that assumption held for design and sensors. It doesn’t hold for performance anymore, not with an operating system built around features the old chip visibly struggles to run.

If you’re on a Series 9 or later: the AI features work today, and the chip isn’t the bottleneck for you yet. If you’re on an Ultra 3 or Series 11: you own the S10, which this reporting says never actually improved on the S9 you might have been comparing it against. You’re not behind by one generation. You’re behind by zero, because the “generation” in between didn’t move anything. If you’re shopping new right now: buying a Series 11 or Ultra 3 today means buying the chip Apple itself is reportedly replacing in about two weeks. That’s a bad time to buy, worse than the timing we flagged back in April.

Who Should Care

Current Ultra 3 and Series 11 owners: Nothing changes for you today, and nothing needs to. But go into watchOS 27 with realistic expectations — you’re running new AI features on chip hardware that, per this reporting, hasn’t actually improved since 2023.

Anyone shopping for a watch this month: Wait. Two weeks is a short window to avoid buying a chip that’s about to become last-generation the moment Apple confirms the event.

People deciding between an Apple Watch and a screenless alternative like Oura or WHOOP: This doesn’t change that calculus much. The chip story is about Apple catching its own software up to itself, not about competing with dedicated recovery hardware.

The Bottom Line

A flat chip for three straight product cycles was a strategy Apple could quietly get away with as long as the Watch’s job was tracking workouts and showing notifications. watchOS 27 makes that strategy visible, because Siri lagging on-device is the kind of thing regular people notice even if they’ve never heard of an S9 versus an S10. The chip upgrade reportedly landing in the Series 12 and Ultra 4 reads less like routine iteration and more like Apple finally building the hardware its own AI ambitions needed two years ago. Nothing here is official yet — pricing, exact specs, and the date all still carry “expected” and “rumored” qualifiers. But the pattern is different this time: hardware and software are finally moving together instead of one waiting on the other.

---

Details reflect reporting from 9to5Mac (August 21, 2026), cross-referenced with 9to5Mac’s watchOS 27 public beta coverage (July 13 and August 20, 2026), MacRumors’ September event reporting (August 13, 2026), and Macworld’s pricing rumor roundup. All chip specs, pricing, and event dates are pre-release reporting, not confirmed by Apple. This post will be updated once Apple announces anything on the record.