With no real immediate competition, TSMC is a behemoth in semiconductor etching. But beware...
No leaks this time: the new TSMC roadmap that spread across the web was released by TSMC itself. The Taiwanese company took advantage of the North American Technology Symposium 2026 held on April 22 in Santa Clara, California, to dot all the i's and cross all the t's, reminding the entire industry that the number one foundry is the Taiwan Semiconductor Manufacturing Company (TSMC). Behind them, the world's number two and three - Samsung and Intel - share the rest. Oh, they're nice leftovers, but it's a sure sign that when AMD, Apple and NVIDIA want to have their best chips produced, it's to TSMC that they turn.
TSMC's plans currently stretch to 2029 © TSMC
At the North American Technology Symposium 2026, TSMC took stock of its future, both near and far, as 2 nm (N2) technology begins to be well established. A future that looks to the infinitely small. So, the name of the next engraving node, the A16, isn't there to pay homage to the freeway that links Paris to Belgium via the North Sea coast. No, A16 is there for 1.6 nm. Surprisingly, while TSMC used to talk about it for late 2026, A16 is now only mentioned for early 2027. Could TSMC be experiencing technical problems?
In reality, we need to distinguish between the development phase, which should have been completed by the end of 2026, and the production phase, which won't be effective until early next year. No, for the time being, there's no question of a delay at TSMC. "For the time being"? After the A16, TSMC is already aiming to bring the A14 to market in 2028, and the A13 and A12 in 2029. All would be well if such processes did not challenge the technical limits of the EUV machines currently used by TSMC. The Taiwanese company has declined to place an order for ASML's new High-NA EUV machines - which cost the princely sum of $400 million apiece - but could be in for a tough time of it. TSMC is counting on optimizing its current machines, while Intel has not hesitated to invest in equipping its Oregon plant with High-NA EUV. Which strategy will be the right one? We'll probably find out in 2028/2029.
