Apple has a long habit of handing last year’s brain to this year’s regular iPhone. The iPhone 15 shipped with the A16 chip while the 15 Pro got the shiny new A17 Pro, and nobody stormed Cupertino with pitchforks. It works because people shop for cameras and battery life, and nobody has ever asked a store clerk about die size. The iPhone 18 is about to run the same play with a twist. This time the regular model gets a brand new A20 chip, except it arrives months after everyone else’s.
The chip is the A20, and a leaked solder ball map analyzed on Weibo by user 白饭炒白米饭 (reported by Notebookcheck on October 1) suggests it is a slimmed down sibling of the A20 Pro. That Pro chip already powers the iPhone 18 Pro, starting at $1,199, and the iPhone Duo, starting at $1,999. Both reportedly use TSMC’s 2nm process, but the standard version supposedly keeps the AI hardware and trims the GPU, cache, and clock speeds. The base iPhone 18 is expected in spring 2027, and Apple has announced no price or date. So what does Apple cut when it wants a cheaper 2nm chip, and what does it refuse to touch?
Designer: Demon’s Tech
A Smaller Die Than Last Year’s Chip, on a Pricier 2nm Node
The standard A20’s bare die is estimated at 70 to 80 mm², which would make it smaller than the A19 at 82.8 mm² and well under the A20 Pro’s 98.8 mm². A die is the actual slab of silicon, and every square millimeter you shave off means more chips per wafer and a lower bill. On 2nm the savings count for more, since Notebookcheck’s report notes those wafers cost significantly more than earlier nodes. Our own look at the base iPhone 18 pegged its body as basically an iPhone 17 in disguise, so there is no fresh industrial design to pay for either. Apple looks to be spending the budget on silicon instead.
Apple reportedly gets there by halving the cache. The standard A20 is expected to carry 8 MB of L2 cache for its two super cores and 4 MB for its four efficiency cores, which is 50% less than the A20 Pro. Cache is the small, ultra fast memory sitting right next to the processor cores, so a smaller helping can mean slightly more trips out to slower memory under heavy load. It is also one of the most direct ways to shrink a die, because SRAM, the circuitry cache is built from, has scaled poorly on recent chip nodes. Honestly, it is the smartest corner to trim, since you will never see a cache spec on a phone box.
Why the A20 Pro Gets Fancy Packaging and the A20 Skips It
The A20 Pro reportedly uses WMCM, short for wafer-level multi-chip module, which places the processor and its memory side by side instead of stacking one on the other. That layout lets the chip connect straight to the vapor chamber, the cooling plate Apple introduced with the iPhone 17 Pro. Apple says the new Pro design delivers up to 40% better sustained performance than the 17 Pro. The standard A20 is expected to stay with traditional package-on-package, where the memory sits on top of the chip like a hat. A hat is cheap and compact, but it also traps heat right where you do not want it.
Packaging is how Apple reportedly gets the A20 Pro to 4.93 GHz, a figure from the leak analysis that Apple has never confirmed. The standard A20 is expected to clock considerably lower, so it likely does not need the same cooling tricks. Pair that with an expected 5-core GPU against the Pro’s 7 cores, and the gap could show up in a long gaming session. Past base chips trailed their Pro siblings by one GPU core, so a two-core deficit would be wider than usual. The vapor chamber stole the show in the iPhone 17 Pro teardown we covered, and the leak says nothing about the base model getting similar cooling.
The AI Hardware Apple Keeps, and What Spring 2027 Looks Like
The AI side of the chip reportedly survives the cuts largely intact. The leak says the A20 keeps the same dual 16-core Neural Engine as the Pro, the dedicated block that handles on-device AI tasks, along with its own 4 MB cache. The system-level cache, a shared pool of memory every part of the chip can tap, may grow from the A19’s 12 MB to 16 MB. The memory bus is also expected to widen to 96 bits, matching the Pro, which means more data can flow to the chip per second. In real life that means live translation and photo cleanup running on the phone instead of waiting on a server.
The iPhone Duo (the one with the open and close animation we gushed over) runs the same A20 Pro, so Apple’s fall lineup now shares one chip. The base model is expected to follow in spring 2027 with a lighter version of that platform. Apple has confirmed no date, price, or chip for it, so every number here is a leak. The iPhone 17 started at $799 and the Pros just rose $100, so the base model’s price is the number to watch. Back in the A16 days the hand-me-down chip was a consolation prize, but this time the regular iPhone gets a fresh 2nm brain with a little muscle trimmed, a deal I would happily take.
