TechSpot is turning 25 years old! TechSpot means tech analysis and advice you can trust.
Prepare to be surprised: a 2nm-capable fab could carry a price tag of $28 billion. The cost of 2nm development and production will be high due to the complexity of the tools and the necessity of highly skilled talent. There is hope, however, in the form of AI-enabled EDA tools that could reduce costs by streamlining processes.
According to consulting firm IBS, a 2nm-capable semiconductor fabrication plant with a capacity of 50,000 wafer starts per month will cost about $28 billion. This is $8 billion more than the cost for a 3nm fab, illustrating the significantly higher costs that companies will face with the next generation of chips.
Costs for 2nm chips are expected to rise by around 50% compared to 3nm processors, according to IBS, which would mean companies like Apple could potentially spend $30,000 to process a single 300mm wafer using TSMC’s N2 fabrication process. However, there may be some flexibility in these numbers that could help bring down the high cost of these chips.
Different approaches that semiconductor companies take during the pre-construction, construction, and operations phases can significantly alter the final cost of the fab.
There are also sky-high costs associated with chip development, with software development alone accounting for $314 million and verification adding another $154 million, according to IBS. Additionally, designing chips at the 2nm node requires specialized, hard-to-find talent. Furthermore, the increased use of photolithography, a process used to create chip patterns, adds to the cost.
The advanced process of photolithography contributes to the increased cost of equipment and materials used in the fabrication process, especially for the 2nm node. The $8 billion cost differential for a 2nm-capable fab is due to the requirement of more EUV litho tools to maintain a 50,000 WSPM capacity.
However, IBS recognizes that there are complexities behind these figures and chip design is continually evolving. The firm has estimated that it could cost a company $725 million to build a sizable 2nm chip from scratch. But for many semiconductor companies, particularly startups, this may not be the case as they tend to pursue more efficient strategies.
IBS also emphasizes the increasing importance of AI-enabled EDA tools in chip design, as they streamline processes and reduce costs by automating complex design processes and optimizing chip performance.
