Source: Finance Association
① TSMC said on Wednesday (April 24) that it is developing a new chip manufacturing technology called “A16,” which will be put into production in the second half of 2026; ② the company said that artificial intelligence chip makers may become the first adopters of this technology; ③ rival Intel previously announced that it is developing a new technology called “14A” to challenge TSMC.
$Taiwan Semiconductor (TSM.US)$On Wednesday (April 24), it was said that a new chip manufacturing technology called “A16” that the company is developing will be put into operation in the second half of 2026. At that time, TSMC will become a long-term competitor$Intel (INTC.US)$Start a “big duel” with the most advanced chips.
TSMC is the world's largest foundry manufacturer of advanced chips, and$NVIDIA (NVDA.US)$und$Apple (AAPL.US)$major partners.
The company announced the “A16” news at a conference in Santa Clara, California. TSMC executives said at the meeting that artificial intelligence chip companies may become the first adopters of this technology rather than smartphone manufacturers. However, AI chip companies need to optimize their designs to take advantage of the full performance of TSMC's manufacturing process.
According to reports, the A16 will combine TSMC's superrail structure with nanosheet transistors. Superrail technology can move the power supply network to the back of the wafer to free up more space on the front side of the wafer, thereby improving logic density and performance, making A16 suitable for high-efficiency computing (HPC) products with complex signal wiring and dense power supply networks.
According to media reports, compared with the N2P process, the A16 chip density is increased by up to 1.10 times. At the same operating voltage, the speed is increased by 8-10%; at the same speed, the power consumption is reduced by 15-20%.
In addition to A16, TSMC also announced that it will launch N4C technology. N4C continues N4P technology, reducing grain costs by up to 8.5% and has a low adoption threshold, and is expected to be mass-produced in 2025.
Regarding this latest announcement, analysts pointed out that TSMC's new technology may put a lot of pressure on Intel. The latter announced in February that it would replace TSMC with a new technology called “14A” to make the chip with the fastest computing power in the world.
The big duel with Intel
Kevin Zhang (Kevin Zhang), senior vice president of business development at TSMC, said that due to the demands of artificial intelligence chip companies, the company's newly developed A16 chip manufacturing process was faster than expected.
Zhang Xiaoqiang pointed out, “AI chip companies are eager to optimize their designs to take full advantage of the performance of our process.”
TSMC is confident enough about the A16. Zhang Xiaoqiang believes that there is no need to use Dutch chip equipment manufacturers$ASML Holding (ASML.US)$A new “high numerical aperture EUV” lithography machine is used to produce A16 chips. In contrast, Intel revealed last week that it plans to be the first company to use the Asmack machine to develop its 14A chip.
According to reports, Asmack's cost of each high-numerical aperture EUV is 373 million US dollars.
Speaking about Intel, Dan Hutcheson, vice chairman of analytics firm TechInsights, said, “Judging from some metrics, I don't think they're leading.”
Kevin Krewell, head of TIRIAS Research, believes that the technology being developed by Intel and TSMC will be several years away from mass production, and they need to prove that the final chip matches the technical capabilities they are promoting.
editor/tolk