In 2024 Chipchain carried out 6 nm AI full-custom chip design for a top AI customer. The company is authorised for TSMC 7 nm FinFET design.

AI chips spend most of their energy in a few blocks repeated many times over: MAC arrays, multipliers, adder trees and the pipeline registers between them. Chipchain applies its full-custom techniques to exactly those blocks: custom cells, library characterisation and re-characterisation, hand-crafted netlists, manual place and route and custom clock trees.

The 6 nm work followed the company's 7 nm design studies of 2019 and 2020, which covered standard-cell trade-offs, custom flip-flop design and placement trials.

With it, Chipchain's design work spans four FinFET process nodes: 16 nm, 12 nm, 7 nm and 6 nm.

Every project runs under a mutual NDA, with access limited to the named project team.