TSMC is aiming to put pressure on Intel and Samsung at a time when the American and Korean companies seemed to be gaining ground against the Taiwanese firm. According to leaks, the company has already set its sights on its next milestone in its advanced roadmap, which was unveiled months ago. Information suggests the start of mass production for its A14 (1.4nm) chip in 2028, with another step added to the schedule: the commencement of A10 chip testing in 2029. This places TSMC’s advancements in direct competition with its rivals, theoretically utilizing EUV High-NA technology simultaneously.
The same reports indicate that TSMC is pushing for mass production of its 2nm process for the fourth quarter of 2026, with orders for this node already reportedly booked through 2028. Will TSMC be the first to market, or will Intel once again take the lead?
TSMC to Begin Mass Production of A14 in 2028 to Counter Intel’s 14A
TSMC officially introduced the A14 at its North American Technology Symposium on April 23, 2025, as its next-generation leading-edge logic process. At the event, several concrete details were shared. Compared to the current N2 process, the A14 promises up to a 15% performance increase at the same power consumption, or up to 30% reduced power consumption at the same speed, along with over 20% greater logic density. The company also announced the evolution of its NanoFlex architecture to NanoFlex Pro, aiming to enhance performance, energy efficiency, and design flexibility.
Therefore, knowing that production for this node is slated for 2028 provides perspective on how long the N2 and N2P nodes will remain in production – likely a short period – and how quickly they will become TSMC’s more affordable nodes sooner than we might expect.
The leak aligns perfectly with this outlook, reiterating the 2028 date for A14 and adding that TSMC plans to supply products to key clients like Apple. This means that by 2028, Apple could have a significant advantage over its competitors if Intel experiences even the slightest delay.
The A10 Remains an Enigma: Will it Ultimately Use EUV High-NA?
Details regarding the A10 are significantly scarcer, which is unfortunately common, as secrecy surrounding these developments is at its peak. The published information mentions testing in 2029, but TSMC has not yet released a public datasheet for the A10 comparable to that of the A14. This makes the lack of information about the A10 particularly noteworthy, as we are essentially starting from scratch. As for other aspects, the situation remains unchanged: there are no official figures on performance, power consumption, density, or mass production dates.
For now, the official and concrete information remains focused on the A14 and N2 nodes. The same leak also highlights two additional points: TSMC is accelerating investment to maintain its position in these processes, and yield will be crucial for ensuring stable production, profitability, and competitiveness in this new era.
While not surprising, it is interesting to note the increased investment, which reveals that TSMC must exert maximum effort because Intel and Samsung are proving to be more formidable competitors than they may wish to acknowledge.
