By Jeremy Laird

Copyright pcgamer

Skip to main content

Close main menu

THE GLOBAL AUTHORITY ON PC GAMES

View Profile

Search PC Gamer

PC Gaming Show

Movies & TV

Affiliate links

Meet the team

Community guidelines

About PC Gamer

PC Gamer Magazine Subscription

Why subscribe?

Subscribe to the world’s #1 PC gaming mag
Try a single issue or save on a subscription
Issues delivered straight to your door or device

From£35.99Subscribe now

Borderlands 4
Essential Hardware

Battlefield 6

Don’t miss these

We’re still waiting for the first 2nm chips but TSMC is accelerating its plans for 1.4nm silicon manufacturing, starting in 2027

If this report about 2nm silicon from Japanese chip manufacturing start-up Rapidus beating TSMC is true, there might just be hope for GPU prices

Graphics Cards
AMD reckons its next-gen GPUs will beat Nvidia at ‘any sort of AI workload’ and we’re praying that rubs off on the company’s gaming graphics cards

Intel claims 18A, the node Pat bet the company on, is either 25% faster or 38% more efficient than Intel 3. Though that’s a node Intel didn’t have enough faith in to release for desktops or laptops

Graphics Cards
The latest AMD RDNA 5 rumours are complicated but it looks like there really is going to be a high-end next-gen GPU to take on Nvidia’s best graphics card

Graphics Cards
AMD just revealed that its current RDNA 4 GPUs have a clever ‘modular’ architecture—here’s hoping it signals a broader range of next-gen graphics cards

Graphics Cards
AMD’s rumoured to be plotting a new ultra high-end gaming GPU, plus a $550 graphics card with RTX 5080 performance, but sadly we probably won’t see either until 2027

Intel’s ex-CEO said he bet the company on the 18A node but now a new report claims Intel is pushing customers to next-gen 14A instead

Graphics Cards
The RTX 50-series has delivered a record-breaking $4.28B in gaming revenue for Nvidia… no matter what you think about VRAM levels, launch pricing, and availability

Intel’s new Nova Lake CPU is reportedly being made on TSMC N2 right now, pointing to a hybrid 18A node and late 2026 launch

New chip industry roadmap predicts true 10 nm silicon won’t arrive until 2039 and yet Moore’s Law is, actually, alive and kicking

Tech firms in China are reportedly still patiently waiting to buy Nvidia’s AI chips, even though the Chinese government would prefer they didn’t

Graphics Cards
If you want to know why graphics cards are still so painfully pricey, you might want to take a look at TSMC’s record-high profits which are heading for $50 billion this year

Further details of Nvidia’s GB10 ‘superchip’ shine light on the complex collaboration with MediaTek

Graphics Cards
AMD CEO Dr. Lisa Su says she can see AI ‘helping us design the next-generation GPU much faster and more reliably’

Nvidia is rumoured to be first in line to use TSMC’s ultra-advanced A16 chip node, although it’s AI GPUs that’ll likely see the benefit first

Jeremy Laird

16 September 2025

Yup, it’s AI driving the move.

When you purchase through links on our site, we may earn an affiliate commission. Here’s how it works.

(Image credit: Bridget Bennett/Bloomberg via Getty Images)

Somewhat surprisingly, Nvidia doesn’t have much of a track record when it comes to using cutting-edge silicon. But a new report claims that’s about to change dramatically, with Nvidia planning to be the very first to use TSMC’s next-next gen A16 node.

Taiwanese outlet CTEE claims that Nvidia will be the first to adopt TSMC’s A16 node due to competition from AMD, in part for what it calls HPC or high performance chips built on TSMC’s upcoming N2 node and due out this year.
The only N2 chips AMD has planned for next year, as far as we know, are server CPUs, not GPUs. AMD’s next major AI GPU release, MI450, is expected to be on TSMC N3 silicon.

Related Articles

We’re still waiting for the first 2nm chips but TSMC is accelerating its plans for 1.4nm silicon manufacturing, starting in 2027

If this report about 2nm silicon from Japanese chip manufacturing start-up Rapidus beating TSMC is true, there might just be hope for GPU prices

AMD reckons its next-gen GPUs will beat Nvidia at ‘any sort of AI workload’ and we’re praying that rubs off on the company’s gaming graphics cards

Nvidia currently uses a version of TSMC’s N4 node for all its GPUs. N4 is actually a refinement of N5, which dates back to 2020. Meanwhile, the first chip made on TSMC’s N3 technology, the Apple A17 Pro SoC for iPhones, debuted back in 2023.

Consequently, Nvidia is now bumping up against what’s known as the reticle limit with its Blackwell generation of GPUs. That’s a physical limit on the size of an individual chip that can be etched onto a wafer.
That physical size in turn puts a hard limit on how many transistors can be squeezed into a single chip. Nvidia’s most powerful AI GPU, GB200, actually uses a pair of reticle-sized GPUs fused together on a single interposer. Nvidia’s RTX 5090 GPU, codenamed GB202, is likewise up pretty close to TSMC’s N4 reticle limit at 750mm2.
Indeed, Nvidia essentially carried over N4 production technology from the previous RTX 4090 family to the current RTX 50 GPUs. That’s one important reason why the likes of the RTX 5070 and RTX 5080 are barely any better than their direct progenitors.

The biggest gaming news, reviews and hardware deals
Keep up to date with the most important stories and the best deals, as picked by the PC Gamer team.
Contact me with news and offers from other Future brandsReceive email from us on behalf of our trusted partners or sponsorsBy submitting your information you agree to the Terms & Conditions and Privacy Policy and are aged 16 or over.
For Nvidia’s next-gen Rubin generation, Nvidia will be using TSMC’s N3 node. But if this news about TSMC A16 is correct, Nvidia could skip N2 altogether and jump straight to A16. Of course, the immediate question with all of this is how relevant it is to gaming. The idea that Nvidia might want to accelerate it’s AI roadmap onto A16 makes a lot of sense. But gaming GPUs?
The thing is, Nvidia’s approach for multiple generations now has been to used a shared GPU architecture across both AI and gaming and produce all chips on the same node. Obviously an AI GPU is very different to a gaming GPU. But certain elements of the chips, such as at least parts of the Tensor cores, are like the same on all chips. That makes it cheaper and faster to engineer GPUn designs for a given node.

One reason why the RTX 5070 is disappointing is because it uses an old TSMC node, not cutting-edge silicon. (Image credit: Future)
Thus, Nvidia’s current Blackwell GPUs are all on N4 and its next-gen Rubin GPUs, including both AI and gaming variants, are expected to be on N3. This TSMC A16-related report, if accurate, would apply to the generation after Rubin, which is known as Feynman and is due out in 2028.
If it turns out that Nvidia is indeed skipping N2 and that includes gaming GPUs, too, then it could be very good news. Nvidia will have many more transistors to play with in a couple of generations’ time as a consequence. And you’d normally expect that to translate into significantly higher performance and perhaps new features.
Of course, Nvidia also has form for launching disappointingly incremental GPU upgrades, such as Blackwell. But as I said, part of the problem with Blackwell is the old TSMC node. So, this A16 rumour for the Feynman generation could be very positive news for PC gaming. As ever, we’ll ultimately have to wait and see.

Best graphics card 2025All our current recommendations

👉Check out our full guide👈
1. Best overall: AMD Radeon RX 9070
2. Best value: AMD Radeon RX 9060 XT 16 GB
3. Best budget: Intel Arc B570
4. Best mid-range: Nvidia GeForce RTX 5070 Ti
5. Best high-end: Nvidia GeForce RTX 5090

Jeremy Laird

Social Links Navigation
Hardware writer

Jeremy has been writing about technology and PCs since the 90nm Netburst era (Google it!) and enjoys nothing more than a serious dissertation on the finer points of monitor input lag and overshoot followed by a forensic examination of advanced lithography. Or maybe he just likes machines that go “ping!” He also has a thing for tennis and cars.

You must confirm your public display name before commenting

Please logout and then login again, you will then be prompted to enter your display name.

We’re still waiting for the first 2nm chips but TSMC is accelerating its plans for 1.4nm silicon manufacturing, starting in 2027

If this report about 2nm silicon from Japanese chip manufacturing start-up Rapidus beating TSMC is true, there might just be hope for GPU prices

AMD reckons its next-gen GPUs will beat Nvidia at ‘any sort of AI workload’ and we’re praying that rubs off on the company’s gaming graphics cards

Intel claims 18A, the node Pat bet the company on, is either 25% faster or 38% more efficient than Intel 3. Though that’s a node Intel didn’t have enough faith in to release for desktops or laptops

The latest AMD RDNA 5 rumours are complicated but it looks like there really is going to be a high-end next-gen GPU to take on Nvidia’s best graphics card

AMD just revealed that its current RDNA 4 GPUs have a clever ‘modular’ architecture—here’s hoping it signals a broader range of next-gen graphics cards

Latest in Hardware

This spicy leverless DIY kit has tiny circuit boards under each button so you can create your own rapid trigger Hitbox

I don’t want to give horror game devs any ideas but this chilling ‘sphere o’ mouths’ lip-syncing robot is haunting my dreams and I want it to stop

This entire website is hosted on a disposable vape so think twice before throwing one in the trash

If you bought Corsair PC memory after 2018 you might be entitled to a share of $5.5 million from a class action over advertised DDR4 and DDR5 speeds

AI cloud companies have guzzled up so much HDD storage that they might be coming for our SSDs next

Are high-end GPUs even worth buying anymore? I’m starting to have my doubts

Latest in News

MindsEye’s response was so mindblowingly negative its star thought ‘I might never work in another game again’

The creator of X-Com isn’t worried about old fans not vibing with his new game: ‘It’s successful for people who have maybe not played any strategy at all’

Nvidia is rumoured to be first in line to use TSMC’s ultra-advanced A16 chip node, although it’s AI GPUs that’ll likely see the benefit first

Witcher creator Andrzej Sapkowski gets gold Polish culture medal, accepts it with face like he’s just had his Switch 2 pre-order cancelled

Palworld is going quiet as Pocketpair takes an iron to its ‘quirks and jank,’ with full release planned for 2026

Those years of Silksong memes were no joke: Even Borderlands 4 can’t escape its shadow after their first weekends on Steam

HARDWARE BUYING GUIDES
LATEST GAME REVIEWS

Best gaming laptop in 2025: I’ve tested a ton of notebooks this generation and these are the best in every category

Best Hall effect keyboards in 2025: the fastest, most customizable keyboards for competitive gaming

Best PCIe 5.0 SSD for gaming in 2025: the only Gen 5 drives I will allow in my PC

Best graphics cards in 2025: I’ve tested pretty much every AMD and Nvidia GPU of the past 20 years and these are today’s top cards

Best gaming chair in 2025: I’ve tested a ton of gaming chairs and these are the seats I’d suggest for any PC gamer

Humanscale Freedom Chair with Headrest review

Strange Antiquities review

Medion Erazer Scout 15 E1 review

Hollow Knight: Silksong review

Asus ProArt Case PA401 review

PC Gamer is part of Future plc, an international media group and leading digital publisher. Visit our corporate site.

Contact Future’s experts

Terms and conditions

Privacy policy

Cookies policy

Advertise with us

Accessibility Statement

Future Publishing Limited Quay House, The Ambury,

BA1 1UA. All rights reserved. England and Wales company registration number 2008885.

Please login or signup to comment

Please wait…