Apple execs tell that they have painstakingly built a chip advantage as they are able to scale one architecture across its entire product ecosystem

Updated on: Sep 11, 2026, 07:01:04 IST

Prefer HTon Google

Share via
Copy link

It is no mean feat that Apple, days apart, has marked significant generational progress for its in-house silicon efforts across the iPhone, Mac and the Apple Watch, and is pushing the envelope of chip innovation. It might seem as if there’s a concerted effort to converge mobile and desktop silicon at Apple, but this trajectory has actually now reversed.

The A20 Pro chip which powers the new iPhone 18 Pro and the iPhone Duo, is a significant step forward in terms of performance, efficiency and on-device artificial intelligence (AI) compute capabilities, having taken inspiration from the M6 chip. Days earlier, Apple announced the M6, M5 Max and M5 Ultra—which fuses two M5 Max chips—arriving with the Mac mini and Mac Studio later this month.

This is an interesting moment in time for Apple, something Sri Santhanam, VP of Silicon Engineering Group at Apple calls a “historical rewind”. He points out that Apple started by building chips for the iPhone and then the iPad. “We could say the M series chips were inspired by the architecture we built for the iPhones and iPads,” he says.

The 2-nanometer similarities

The M6 is Apple’s first 2-nanometer architecture chip, while the M5 Ultra is its most powerful desktop chip yet. Then there’s the S11 chip that powers the Apple Watch Series 12 and Watch Ultra 4, also on the efficient 2nm architecture, with an advanced security component that provides the foundation for Apple Intelligence.

“It is about the power of integration between the silicon hardware and software. It is about the power of integration between the silicon hardware and software. We set a goal for increasing sustained performance, and the product design and systems architecture teams worked to figure out how to solve this at a holistic level,” explains Kaiann Drance, VP of Worldwide iPhone & Apple Watch Product Marketing, in a conversation with .

Overall power consumption of 2nm architecture is lesser compared to previous generation 3-nanometer chips, alongside a performance uplift. The A20 Pro marks two milestones—first 2nm chip in smartphones and Apple’s first 2nm silicon for iPhone. The A20 Pro chip has 2 neural engines, unheard of in smartphones. This means twice the AI compute power compared to the A19 Pro chip it succeeds.

“There’s the move to 2nm, which has smaller transistors, and is more compact. So there’s less switching capacitance, and that saves power. A lot of effort went into implementation and design techniques to further reduce power,” says Santhanam.

Apple explains this chip can perform as much as 40% better sustained performance than the iPhone 17 Pro, particularly in the iPhone Duo. This performance is achieved through a combination of the vapour chamber (which now has more coverage as part of the chassis), the use of graphite, and the industry-first nano-twin copper material to dissipate heat.

“We’re building upon the innovations from last year, and we’re doing as much as we can within the size of the envelope we have, given the footprint of the iPhone,” says Drance. “There are multiple facets that go into our energy efficiency, and we’re tackling that at every possible angle. And this time, changing to a new custom packaging is one of the reasons why we have industry-leading performance per watt.”

If we hazard a guess, the A20 Pro silicon is great news for a future generation MacBook Neo, as this chip will eventually be used there.

Synergy of convergence

The A20 Pro’s development was inspired by the M-series chips, particularly with the dual neural engines, and thermal management techniques. There is a 50% increase in memory bandwidth, which marks serious performance gains.

“It’s also a real testament to the scalable architecture. All the way from the form factor of an iPhone to a Mac Studio which has tremendously higher thermal capacity, you can just watch the scalability of Apple Silicon sing because it’s such a great architecture,” says Tom Boger, VP of Worldwide Mac & iPad Product Marketing.

Drance explains that the advantages of the dual Apple Neural Engine architecture transcend across workflows and usage scenarios.

“There are a lot of things that we do for image generation, for natural intelligence, and for camera capabilities. We’re also using that to improve the quality of photos from the camera, which gives us more to work with during processing,” she says, illustrating how having 32 cores provides the foundation for powerful on-device AI models for diverse workloads and gaming.

The A20 Pro has dual Neural Engines, while the Apple Watch’s S11 has four.

The M-series inspiration also comes from how the A20 Pro connects directly to a custom vapour chamber, dissipating heat that’s often causes performance fluctuations during multitasking, gaming, and complex AI workloads.

“As we develop each chip, if there’s a great new technology, we propagate it to the other chips as well. Apple Silicon had a tremendous focus on performance per watt, and we’ve never let go of that discipline,” Boger says.

The M-series inspiration also comes in the form of how the A20 Pro connects directly to a custom vapour chamber, dissipating heat that’s often the reason for performance fluctuations during multitasking, gaming, and complex AI workloads. Key has been a significant change in the packaging—the DRAM (Dynamic Random-Access Memory) volatile memory that is usually stacked atop the System on Chip (SoC) has been moved to sit side by side. For years, Apple researched ways to reduce the chip’s footprint.

“We’ve even increased the size of our caches, our system cache, or even our efficiency core’s cache sizes have all gone up, and that saves power. We don’t have to access DRAM as much because we keep data on board. So there’s architectural goodness in reducing power and improving performance,” says Santhanam.

He explains that in addition to the advantages of the 2nm technology, the packaging changes of the A20 Pro chip allowed Apple to rework memory bandwidth and channels, which saves power as well. “It certainly creates a challenge because when you take a DRAM from the top of the SoC (or system-on-chip) and move it to the side, the footprint grows, taking away valuable real estate,” Santhanam explains.

Apple first implemented this packaging change on the M6 chip for Macs, and now it has been replicated on the iPhone’s A20 Pro.

Does this mean Apple views this moment as one of possible convergence of mobile and desktop silicon over the next few years? Santhanam calls this a “historical rewind”, pointing out that Apple started by building chips for the iPhone and then the iPad. “We could say the M series chips were inspired by the architecture we built for the iPhones and iPads,” he says.

“As we develop each chip, if there’s a great new technology, we propagate it to the other chips as well. Apple Silicon had a tremendous focus on performance per watt, and we’ve never let go of that tremendous discipline. Every chip, no matter where it goes—even in a Mac Studio—maintains that relentless focus on performance per watt; it’s a common characteristic on every chip we develop,” Boger illustrated the vision.

Powerful compute, on your wrist

The Apple Watch Series 12 and the Watch Ultra 4 are powered by the new S11 chip, also on the 2-nanometer architecture. There was a need for significant chip upgrades, considering Apple is making a big pitch for Apple Intelligence as a private, yet personal intelligent hub on the iPhone and across all Apple devices. The feature set is dynamic across iPhone, Watch, and AirPods, for instance.

On the Watch for instance, will enable the new Siri AI as well as Audio Intelligence features including Live Rewind of conversations. Drance points out that the S11 chip’s Secure Exclave security component is the reason why Apple has been able to build the audio intelligence features for both Live Rewind and Siri Recap. “Privacy was fundamental,” she says.

See Less