Introduction

Every autumn, Apple walks on stage and calls its newest chip the fastest one it has ever built. Most years, that claim is true on paper and forgettable in practice. The Apple A20 chip is shaping up to be different.

Reports describe it as Apple’s first processor built on a 2-nanometer manufacturing process, paired with a new packaging method and a larger Neural Engine. This article walks through what is known and reported about the A20 so far, section by section, without the marketing gloss.

What Is the Apple A20 Chip

The A20 is the next generation of Apple’s custom smartphone processor, expected to follow the A19 and A19 Pro. Reports point to two versions again this cycle: a standard A20 and a higher-tier A20 Pro, mirroring the split Apple has used in recent years.

The A20 Pro is expected to arrive first, inside the iPhone 18 Pro, iPhone 18 Pro Max, and a foldable iPhone this September. The regular A20 is reportedly planned for the standard iPhone 18 and iPhone 18e, expected in the first half of 2027.

Apple has not confirmed the chip or its name. Everything here comes from industry reporting and supply-chain sources rather than an official announcement.

The Move to a 2-Nanometer Process

Every A-series chip since the A17 has used TSMC’s 3-nanometer family of processes. The A20 generation is reported to shift to TSMC’s N2 node, the chipmaker’s first commercial 2-nanometer process.

This is not a small tweak. Current Apple chips rely on FinFET transistors, a structure that has carried several generations of gains. TSMC’s N2 introduces gate-all-around nanosheet transistors, which give the manufacturer tighter control over current flow through each transistor.

Better current control tends to reduce electrical leakage. In practice, that can translate into either more performance at the same power draw, or similar performance at lower power. TSMC itself has said N2 can offer roughly 10 to 15 percent more performance at equal power, or 25 to 30 percent less power at equal performance, compared with its previous node.

There is a cost side to this story too. First-generation nanosheet manufacturing reportedly has lower yields than mature FinFET production, and the new node requires new materials and tighter tolerances. That combination is expected to push up production costs for Apple.

New Packaging: What WMCM Changes

Alongside the process shrink, the A20 is reported to introduce a new packaging approach called Wafer-Level Multi-Chip Module, or WMCM. This would replace the InFO packaging Apple has used in recent chip generations.

The idea behind WMCM is to combine separate dies, such as the CPU, GPU, and Neural Engine, onto a single package rather than fabricating them as one monolithic chip. Each die can then draw power based on what it is actually doing at a given moment, instead of sharing a single power budget as one large piece of silicon.

Reports also mention new capacitors in the power delivery system, described as super-high-performance metal-insulator-metal capacitors. These are said to offer more than double the capacitance density of the previous generation, while cutting resistance in the circuit by around half.

Longer term, this modular packaging approach is expected to give Apple more flexibility. The same building blocks used for the A20 and A20 Pro could, in theory, scale up toward future M-series chips for Mac.

CPU and Performance Expectations

Specific CPU core counts and clock speeds for the A20 have not been confirmed. Benchmark predictions circulating ahead of launch point to Geekbench single-core scores around 4,200, with multi-core scores above 10,000.

For context, the current A19 Pro scores in the high 3,800s for single-core and around 10,000 for multi-core on Geekbench 6, depending on the source. If the predicted A20 numbers hold up, the generational jump would be incremental by Apple’s historical standards rather than dramatic, even with the new node underneath it.

That framing matters. A 2-nanometer process is a genuine manufacturing milestone, but a smaller transistor does not automatically mean a huge leap in everyday benchmark scores. Much of the real-world benefit may show up in efficiency and sustained performance rather than peak numbers alone.

GPU and Gaming Improvements

Graphics is one of the areas reportedly getting real attention this generation. Predictions point to results above 50 frames per second on the demanding 3DMark Solar Bay Unlimited benchmark, which would represent more than double the real-world graphics performance Apple offered three years earlier.

Some of that gain is expected to come from the new node allowing higher sustained clock speeds without hitting thermal limits as quickly. The A19 Pro already improved on this front compared with the A18 Pro, thanks in part to a vapor chamber cooling system inside the iPhone 17 Pro.

Whether the A20 Pro repeats a large GPU jump, or settles into a smaller yearly bump like most recent generations, remains to be seen until Apple confirms details.

Neural Engine and On-Device AI

Several reports describe the Neural Engine as the component receiving the biggest share of new transistor budget in the A20 generation, ahead of the CPU. This lines up with Apple’s broader push toward running more AI features directly on the device rather than in the cloud.

The current A19 and A19 Pro already pair a 16-core Neural Engine with neural accelerators built directly into the GPU cores, a combination Apple has used to boost on-device machine learning tasks like dictation, photo processing, and Siri’s newer voice.

An upgraded Neural Engine in the A20 would likely continue that direction, supporting more advanced AI features while keeping the processing on-device for privacy and speed reasons rather than routing requests to a server.

Memory, Modem, and Connectivity

RAM is expected to grow again with the A20 generation. Reports mention configurations in the range of 12 to 16GB of LPDDR5x memory for the higher-end models, up from the 12GB found in the current A19 Pro.

Extra memory matters for AI workloads specifically, since larger on-device language models need room to run without constantly swapping data in and out. iOS features tied to dictation and Siri already require devices with additional RAM, which suggests memory capacity will keep climbing alongside AI ambitions.

On modems, reports link the A20 Pro generation to Apple’s own C2 modem, continuing the company’s shift away from third-party cellular chips that began with the C1 and C1X modems in recent iPhone models.

Which iPhones Will Get the A20

Reports suggest the A20 Pro will debut first, inside the iPhone 18 Pro, iPhone 18 Pro Max, and Apple’s first foldable iPhone, all expected around the usual September launch window. Rumored display sizes for that trio include a 6.3-inch iPhone 18 Pro, a 6.9-inch iPhone 18 Pro Max, and a foldable device with a 5.3-inch outer display and a 7.8-inch inner display.

The standard A20 chip is reportedly reserved for the regular iPhone 18 and iPhone 18e, which are expected in the first half of 2027 rather than alongside the Pro models. A second-generation iPhone Air may also use an A20-series chip, though that device has reportedly been delayed from its original fall 2026 target.

None of this has been confirmed by Apple, and device names, timing, and specifications can still shift before launch.

Key Takeaways

• Process node: The A20 generation is reported to move to TSMC’s 2-nanometer N2 process, Apple’s first.
• Packaging: A new method called WMCM is expected to combine the CPU, GPU, and Neural Engine as separate dies on one package.
• CPU performance: Predicted Geekbench scores suggest an incremental generational gain rather than a dramatic one.
• GPU performance: Graphics benchmarks are predicted to more than double compared with results from three years earlier.
• AI focus: The Neural Engine is reported to receive the largest share of new transistor budget this generation.
• Memory: RAM is expected to rise toward 12 to 16GB on higher-end models.
• Rollout: The A20 Pro is expected first in Pro and foldable iPhones this September, with the standard A20 following in 2027.

Frequently Asked Questions

Is the Apple A20 chip officially announced?
No, as of this writing the A20 and A20 Pro are based on industry reporting and supply-chain sources rather than an official Apple announcement.

What manufacturing process will the A20 use?
Reports indicate the A20 will be built on TSMC’s N2 process, a 2-nanometer node using gate-all-around nanosheet transistors.

Which iPhones will get the A20 Pro first?
The A20 Pro is reportedly expected in the iPhone 18 Pro, iPhone 18 Pro Max, and a foldable iPhone launching around September.

Will the A20 have more RAM than the A19 Pro?
Reports suggest a range of 12 to 16GB of LPDDR5x memory, which would be an increase over the A19 Pro’s 12GB.

What is WMCM packaging?
WMCM, or Wafer-Level Multi-Chip Module, is a reported packaging method that combines separate CPU, GPU, and Neural Engine dies onto a single package instead of one monolithic chip.

Conclusion

The Apple A20 chip represents more than a routine annual update, at least based on what has been reported so far. A first-generation 2-nanometer process, a new packaging method, and a Neural Engine that reportedly gets the largest share of new transistor budget all point to a generation focused as much on AI and efficiency as on raw speed.

Predicted CPU gains look modest by Apple’s own historical standards, while GPU improvements appear more substantial on paper. Until Apple confirms the chip on stage, all of these details remain informed reporting rather than fact, and some specifics will likely change before the iPhone 18 lineup actually ships.