In the ever-evolving world of semiconductor manufacturing, where size matters – and matters a lot – Taiwan Semiconductor Manufacturing Company (TSMC) seems to be taking a cautious approach in adopting new technologies. According to sources from production tool manufacturers, TSMC has not placed orders for high-numeric aperture extreme ultraviolet (High-NA EUV) tools and is unlikely to use this technology for the 2nm and 1.4nm (A14) process manufacturing.
Exploring the Future of Chip Manufacturing
The website DigiTimes Asia, seeking to delve into claims made by industry analysts in early 2024, spoke with informants within the manufacturing tool industry. Forecasts from SemiAnalysis and China Renaissance suggest that TSMC is unlikely to adopt High-NA EUV production techniques within a five-year period. This new approach to non-upgrade in the coming years has generated significant interest in the technological sphere.
Intel Foundry Services (IFS) is mentioned as one of the first semiconductor manufacturers to roll out the most advanced machinery from ASML, although with no confirmed timelines. On the other hand, TSMC, the rival (and occasional business partner) of Team Blue, seems content with its existing infrastructure, although industry analysts indicate that cost considerations are key factors behind TSMC’s caution in its planning for the next decade.
TSMC’s Wait-and-See Strategy
Internal sources from DigiTimes firmly believe that TSMC will not make major changes until at least 2029, possibly coinciding with a 1nm production node. Analysts from China Renaissance suggest that High-NA EUV machines could be delivered in the future when facilities are prepared for a process with the codename “A10.” TSMC has an ambitious product roadmap in terms of “transistor count” for the coming years, aiming for the first 1nm products to be launched in 2030, although this timeline could change due to unforeseen circumstances.
Intel plans to “incorporate” its collection of more sophisticated ASML equipment once the 18A process becomes obsolete, with a feasible timeframe of 2026 to 2027, according to Tom’s Hardware.
Why the Caution?
TSMC’s strategy may seem counterintuitive in an industry known for its rapid innovation. However, there are several key factors to understand this apparent reluctance. First, the implementation of High-NA EUV technology requires massive financial investment, not only in terms of acquiring the new ASML tools but also in upgrading facilities to support this advanced technology.
Additionally, introducing a new manufacturing technology carries significant risk. Adjusting complex chip production processes to optimize new technology takes time and can result in delays and even reduced production efficiency during transition periods.
Finally, market demand also plays a crucial role. While the race to reach smaller production nodes is relentless, so is the cost-effectiveness analysis of implementing these improvements. TSMC could be carefully evaluating whether early adoption of High-NA EUV will offer a competitive advantage sufficient to justify its huge cost.
The Path to the Future
TSMC’s decision to possibly delay the adoption of High-NA EUV technology until 2029 or beyond reflects a cautious stance in an industry constantly balancing on the edge of innovation and financial prudence. However, the implications for the future of semiconductor manufacturing are profound. As companies like Intel move forward with new technology, the question remains: will TSMC take the risk at the right time to stay ahead of the competition, or will their caution become a disadvantage in the technological race?
Only time will tell how these strategies will unfold and whether TSMC’s wait to adopt High-NA EUV will prove to be a masterstroke or a missed opportunity. What is certain is that the semiconductor sector will continue to be a hotbed of innovation, with every strategic player seeking to secure its place in the future of technology.