Chipmaker TSMC's new A14 process will apparently offer a '15% speed improvement' but our GPUs won't be made on it for a while
TSMC, the world's biggest chipmaker, has just announced another process node which will almost certainly, amongst an undoubted slew of AI chips, be used to make some of our gaming GPUs and CPUs in the future. This next-gen process is 'A14', meaning 14 angstroms or 1.4 nanometres or really, really small.
This was announced yesterday at TSMC's North America Technology Symposium, and the company says the process "is designed to drive AI transformation forward by delivering faster computing and greater power efficiency."
TSMC was already the world's biggest chipmaker even before all this AI business started to really kick off, but ever since then, it's a company name, alongside Nvidia, that's on a ton of people's lips. Naturally, then, talk of its upcoming process nodes will be of interest to many people, but we PC gamers can throw our hats into that pool of interested onlookers, too.
That's because TSMC makes lots of the chips that end up in some of the best gaming CPUs and best graphics cards, whether from AMD, Nvidia, or even Intel. Currently, for instance, Nvidia RTX 50-series GPUs are made predominantly on TSMC's 4 nm node, and the same goes for AMD's Ryzen 9000-series processors. Intel's Arrow Lake chips look to now be made exclusively by TSMC, too, ever since Intel killed its 20A process last year.
The newly announced A14 node is planned for 2028. Compared with its upcoming N2 process (set for later this year), TSMC says "A14 will offer up to 15% speed improvement at the same power, or up to 30% power reduction at the same speed, along with more than 20% increase in logic density."
Although it's a few years away, I can't help but get a little excited about new processes. That's primarily because we've seen with the RTX 50-series GPUS just how unexciting a new GPU generation can be if it doesn't come off the back of a new process node (the RTX 50 series is on the same process as the RTX 40 series).
That being said, Nvidia doesn't usually use bleeding-edge nodes for its GPUs, and we'll be more likely to see AMD and Apple chips made using A14 to begin with.
A14 won't have backside power delivery until 2029, either, according to our colleagues at Tom's Hardware. Backside power delivery essentially moves power interconnects to the underside of the chip, reducing inference and the distance that power has to travel, thereby increasing efficiency and performance.
Your next upgrade
Best CPU for gaming: The top chips from Intel and AMD.Best gaming motherboard: The right boards.Best graphics card: Your perfect pixel-pusher awaits.Best SSD for gaming: Get into the game ahead of the rest.
We expect to see backside power delivery (AKA 'Super Power Rail') from TSMC first with A16 in 2026. Intel's ahead in this game, however, as its 18A process already has backside power delivery and is ready to go as of two months ago.
As for whether this A14 production will also make an appearance in the US, as well as from TSMC's Taiwan fabs, it seems like it might. I'm basing this on the company's recent earnings call, in which the company claimed that six fabs are planned in Arizona: "In that six fab, the 2-nanometer will be a major node, and that's what I say, 30% will be there. As time goes by, after the 2-nanometer will be 1.4 and 1.0, that has not been discussed yet."
This was in response to a question about what percentage of future leading nodes will come from the US vs from Taiwan, and to my ears it seems like TSMC is saying 1.4 and 1.0 will come from the US, but the percentage hasn't been discussed yet.
Whatever the case, here's to some healthy progress in process nodes across the board, whether from TSMC, Intel, or anyone else. Architectural and AI changes aside, raw performance increases are a direct result of transistor density, and we can all get behind that.
Hashtags
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
CNBC
22 minutes ago
- CNBC
Nvidia says its AI chips do not have a 'kill switch' after Chinese accusation
Nvidia on Tuesday rejected Chinese accusations that its data center GPUs for artificial intelligence include a hardware function that could remotely deactivate the chips, which is commonly called a "kill switch." "NVIDIA GPUs do not and should not have kill switches and backdoors," wrote Nvidia's Chief Security Officer David Reber in a blog post on Tuesday. The blog post comes after the Cyberspace Administration of China said last week that it needed Nvidia to provide documents about what it called security vulnerabilities in the H20, Nvidia's data center AI chip intended for the Chinese market. The regulator specifically mentioned "backdoor" security risks, according to the New York Times. The statement is an example of how Nvidia is navigating geopolitical conflict as its AI chips remain in high demand by countries and companies around the world. U.S. lawmakers have proposed legislation that would require AI chips under export regulations to be equipped with location-tracking systems. The U.S. has placed export controls on some Nvidia chips to China because of national security reasons, saying that the country could use the chips to gain an advantage in AI or for military purposes. Nvidia CEO Jensen Huang has argued that it is better for the U.S. if Nvidia's chips become the global standard for AI computers, especially among Chinese developers. The H20 generates billions in revenue per quarter for Nvidia in sales, although the company does not typically break out its revenue specifically. The chip was briefly banned from export to China in April. The company said its guidance would have been about $8 billion higher except for lost sales from a recent export restriction on its China-bound H20 chips. The Trump administration said in July that it would grant a waiver for the chips to resume sales. Silicon Valley technologists and security experts generally believe that backdoors — when a device has a hidden function that would allow a government or attacker to secretly take data from a computer or otherwise control it — are untenable in products. Apple, in particular, has publicly fought off government requests for what it calls "backdoors" in the past as well. Nvidia declined to comment further on its blog post. Reber argued in the blog post that secret backdoors are dangerous vulnerabilities that could be used by hackers, not just officials, and that they "violate the fundamental principles of cybersecurity." He also said that if a kill switch or backdoor were to be put in products like Nvidia GPUs, that they would harm U.S. national security interests. "Hardwiring a kill switch into a chip is something entirely different: a permanent flaw beyond user control, and an open invitation for disaster," Reber wrote. "It's like buying a car where the dealership keeps a remote control for the parking brake — just in case they decide you shouldn't be driving."
Yahoo
an hour ago
- Yahoo
Foxconn to sell former car factory in Ohio for $375m
Taiwan's Foxconn has finalised a deal to sell a former car factory in Lordstown, Ohio, for $375m, which includes the facility's machinery, according to a report by Reuters. Despite the sale, Foxconn plans to continue operations at the site, focusing on a broader range of products that align with its strategic priorities. The company, known for manufacturing data centre products for Nvidia and assembling iPhones for Apple, did not specify the exact products to be produced at the plant. However, it noted that its cloud and networking product business has experienced "significant growth." A source familiar with the situation informed Reuters that the Ohio facility is expected to support artificial intelligence (AI) data centres. The site spans over six million square feet, making it six times larger than a plant Foxconn is currently constructing in Houston for the production of Nvidia's GB300 AI servers. Foxconn acquired the Lordstown plant, previously a General Motors small-car factory, in 2022 for $230m from the now-bankrupt US electric vehicle startup Lordstown Motors Corp, as part of its strategy to expand into the electric vehicle sector. In 2022, the company began pre-production at the Lordstown Motors plant and had initially invested in the facility to produce electric pickup trucks in collaboration with Lordstown. However, the partnership deteriorated, leading to Lordstown Motors' bankruptcy and subsequent legal action against Foxconn. In its announcement, Foxconn stated that it sold the factory to an "existing business partner," but did not provide additional information. The company reaffirmed its commitment to automotive customers in the US, indicating its capability to quickly scale up automotive production to meet demand when necessary. Foxconn's recent activities reflect its ongoing expansion beyond its traditional role as an iPhone assembler. Last week, the company announced a strategic partnership with industrial motor manufacturer TECO Electric & Machinery to develop data centres. In June 2025, Elektrobit signed a joint development agreement with Foxconn to create a new AI-centric software platform intended to support software-defined electric vehicles. "Foxconn to sell former car factory in Ohio for $375m" was originally created and published by Just Auto, a GlobalData owned brand. The information on this site has been included in good faith for general informational purposes only. It is not intended to amount to advice on which you should rely, and we give no representation, warranty or guarantee, whether express or implied as to its accuracy or completeness. You must obtain professional or specialist advice before taking, or refraining from, any action on the basis of the content on our site.
Yahoo
an hour ago
- Yahoo
TSMC Rocked by Trade Secrets Scandal Tied to Next-Gen AI Chips
Taiwan authorities have arrested six individuals tied to a suspected theft of core chipmaking technology from Taiwan Semiconductor Manufacturing Co. (NYSE:TSM), a company that sits at the very center of the global tech supply chain. The arrests came after TSMC flagged former and current employees who may have illegally accessed sensitive information. Prosecutors searched homes from July 25 to 28, and the investigation now centers on whether any data may have leaked externally. TSMC said it caught the issue early through an internal probe, and has already begun legal and disciplinary action against those involved. Several reports suggest the stakes may be tied to TSMC's upcoming 2-nanometer process node, which is expected to enter mass production later this year. The Nikkei reported that multiple employees were fired for allegedly attempting to extract critical technical details. Meanwhile, Taiwan investigators also searched the local office of Tokyo Electron, a key Japanese supplier. That development adds another layer of complexity, as global players double down on chip R&D just as generative AI demand puts semiconductor capabilities under a microscope. For investors, the timing is hard to ignore. Capital expenditures in chipmaking are pushing past $30 billion annually for giants like TSMC and Samsung (SSNLF), while rivals in ChinaHuawei and SMICare still producing at 7nm. In the U.S., Intel is progressing on more advanced nodes. With companies like Meta and DeepSeek pouring billions into AI infrastructure, TSMC's ability to safeguard its IP may become just as important as delivering next-gen performance. What happens next could ripple across the entire AI supply chain. This article first appeared on GuruFocus.
