Apple, in a first, reported no CPU gains for the A15, likely because of the engineer exodus to companies like Nuvia and Rivos
everything we know Michael Simon / Macworld : The rumor mill got a lot wrong about Apple's iPhone 13 event Tweets: Dylan Patel / @dylan522p : A15 is a 7.6% ST increase with 8% clock increase. CPU architecture looks mostly the same, but with new MMU and nested virtualization. https://browser.geekbench.com/ ... https://browser.geekbench.com/ ... @nerdtechgasm : Tech community not impressed by Apple's new iPhone due to lacklustre CPU gains on A15.. (same node, +GPU +ML perf) Me: Battery lasts 2x longer on a single charge? Impressive. Dylan Patel / @dylan522p : A15 in the iPhone 13 Pro has a 55% increase on GPU vs A14 in iPhone 12 Pro. CPU caches are unchanged, but clock went up to 3.23GHz from 3GHz. Doubt there's much of an IPC increase. The 32MB LLC and potential inclusion of LPDDR5 doing big work in GPU. https://browser.geekbench.com/ ... Dylan Patel / @dylan522p : @reneritchie ... You really can't read. Maybe don't subtweet when you don't understand the first thing about semiconductors. https://twitter.com/... @dcominottim : What a crappy announcement of the A15. Is Apple finally hitting strong the IPC wall? Brain drain after key top architects and designers left for NUVIA and others? #SiliconGang Alex Russell / @slightlylate : @CydeWeys If the continued increase in system cache size is true, I am keyed to think the hand-wringing about low net perf wins on A-series SoCs may be overwrought...Apple's always been able to win speed by moving data closer to the core. Ben Bajarin / @benbajarin : @dylan522p They put those resources in other areas of the design. And the meta point is its much less about CPU these days and more about the heterogenous solution. To say it is because the team is weakened is just not true. We can debate other reasons but that's not it. Dylan Patel / @dylan522p : @BenBajarin CPU engineers don't magically flip over to other teams and architectures. The article talks about heterogenous, yes. The team is certainly weakened due to Nuvia and not Rivos departures. Can't the >100 engineers who have left. Hussein Kanji / @hkanji : With Intel design teams starting to get back on track, AMD executing almost flawlessly, and Qualcomm coming in soon like a hammer with Nuvia cores, we aren't sure if Apple's lead will be sustained https://semianalysis.substack.com/ ... Ben Bajarin / @benbajarin : @dylan522p This was less about struggling and more about being second gen 5nm, essentially no process gains. The pattern when jumping to a new process is generally 20% perf gains so wouldn't expect much if staying on same process as they are. Dylan Patel / @dylan522p : @BenBajarin Process node isn't what dictates CPU architecture. It helps a bit on power and clocks, sure but it isn't the main reason. Apple's lead comes from architecture. They have the transistor budget to change the architecture and get more gains. They didn't use it on the CPU though. Rene Ritchie / @reneritchie : There are some hilariously bad A15 hot takes making the silicon sub stack rounds. I'm working on a video, but it'll take a while. In the meantime, I look forward to reading @BenBajarin @anandtech, and others who understand how transistor budgets can be spent across features... Andreas Proschofsky / @suka_hiroaki : Apple was really good in hiding it yesterday but it seems like the A15 in the iPhone 13 is not going to have any gains in CPU performance compared to the A14. Which is quite a surprise given Apple's steady progress during the last few years https://semianalysis.substack.com/ ... Michael Gartenberg / @gartenberg : “it makes me wonder if Apple is perhaps trying to soft-pedal a new chip that isn't much faster than the older model” Apple doesn't say (or omit) things without much thought https://www.macworld.com/... Gabriele / @lelesocho : Unprecedented in the whole history of iPhone, it wouldn't be concerning usually since they didn't change the process node... but that did happen in the past and still had CPU gains even on the same node. Apple chip supremacy potentially ending, ironic now that they are on Mac too. https://twitter.com/... @mistertechblog : “We believe #Apple had to delay the next generation CPU core due to all the personnel turnover Apple has been experiencing” https://semianalysis.substack.com/ ... Sascha Segan / @saschasegan : iPhone 13: A15 processor has 2 fast, 4 efficiency cores. Claim 50% faster than “competition” (aka Snapdragon 888.) 4-core GPU, claim 30% faster than competition #AppleEvent Rich DeMuro / @richontech : “Frankly, the competition is still playing catch up to our chips. Not just from last year, but from two years ago” Says Apple as they introduce the new A15 Bionic chip #AppleEvent https://twitter.com/... @tomsguide : All the little details about the #A15Bionic from the presentation. Apple's comparing it vs the competition and not the iPhone 12, which is interesting. Wonder what the year vs year upgrades are. #AppleEvent https://twitter.com/... Stephen Shankland / @stshank : The new A15 Bionic processor has 15 billion transistors and gets new AI and graphics abilities for better performance shooting photos and running apps, Apple says at #AppleEvent. https://www.cnet.com/... Kevin C. Tofel / @kevinctofel : Think I'll be sticking with my iPhone 12. 15 billion transistors in the 13 sounds nice but I can get by with as few at 10 or 11 billion. https://www.apple.com/... Benedict Evans / @benedictevans : Transistors in the iPhone 13 SoC: 15bn Transistors in all the CPUs in all the Macs sold in 1984: ~25bn. Neil Cybart / @neilcybart : I like Apple's strategy of announcing the iPhone 13 as if it's the only flagship and then only afterwards do they say “Oh, you want something better? Here's the iPhone 13 Pro.” (Apple has done the same in previous years.) Adam Nash / @adamnash : The Apple A15 Bionic chip is just amazing. 15 billion transistors at 5nm. 15.8 trillion operations per second. During my 1st internship (@NASA) we were working on scaling software to take advantage of a future w/ 1 teraflop supercomputers... now you have 15x that in a phone. 😳 https://twitter.com/... Vadim Yuryev / @vadimyuryev : Low performance gain is likely due to the overheating issues that force Apple to throttle down the performance. I mentioned this in my previous leaks. If Apple couldn't bring vapor chamber cooling to the iPhone 13, performance would be limited again. We saw that with the (1/2) https://twitter.com/... John Garfield / @peresuslog : This problem can only be solved when Apple abandons ARM and adopts the superior RISC-V ISA. Only a matter of time... https://twitter.com/... Dylan Patel / @dylan522p : Apple's CPU team is struggling and delivered no CPU gains for the 1st time ever! The next major core architecture was likely delayed due to personnel overturn from Nuvia exodus With a new exodus to Rivos and stronger competition, future doesn't look bright https://semianalysis.substack.com/ ...
Context & Ripple Effects
Apple had established a steep CPU-performance trajectory with the A14, whose single-thread performance had nearly tripled over five years. Against that backdrop, the absence of a clear CPU uplift in the A15 presentation drew unusual attention to reports that more than 100 engineers had moved to Nuvia and Rivos.
The A15’s disclosed configuration emphasized a six-core CPU alongside a 16-core Neural Engine, while later testing found meaningful efficiency gains and workload-dependent speed changes. Separate coverage also recorded a 52% GPU gain, making the CPU narrative only one part of the chip’s generational change.
First-order effects
- Apple faces immediate scrutiny of its CPU-design execution and talent retention after the reported departures to Nuvia and Rivos, rather than being judged solely on the A15’s broader feature set.
- Nuvia and Rivos gain credibility as destinations for Apple silicon engineers, while Apple’s A15 case shifts attention toward efficiency, GPU performance, and its Neural Engine.
Second-order effects
- Intel and AMD, already facing an Apple silicon challenge after the A14’s large single-thread-performance advance, gain a clearer benchmark for assessing whether Apple can sustain that CPU trajectory.
- Apple’s chip messaging has more reason to emphasize heterogeneous performance: GPU and machine-learning accelerators can carry visible generational gains when CPU improvements are modest or uneven by workload.
Third-order effects
- If Apple’s CPU gains remain less predictable, custom-silicon competition will increasingly turn on execution depth across CPU, GPU, efficiency cores, and dedicated accelerators—not on a single headline CPU metric.
- The reported movement of specialist engineers to Nuvia and Rivos points to talent concentration as a strategic constraint in custom-chip roadmaps, alongside architecture and process choices.
The trend: Mobile silicon is moving toward heterogeneous performance gains, while retaining elite CPU-design talent becomes a core execution risk for chip teams.