TSMC Watch posted an update
TSMC’s A16 chip process is expected to reach volume production in 2027, with the company positioning it for high-performance computing. Startup Fortune reports that the process is expected to be production-ready in 2026.
Why it mattersThe report says A16 could deliver 8 to 10 per cent faster performance at the same power, or use 15 to 20 per cent less power at the same speed, compared with TSMC’s N2P process. It also describes a new backside power-delivery design. That makes this relevant to AI hardware: better performance or lower power could matter for chips running data-centre workloads. These are figures and timing attributed to TSMC by Startup Fortune, not independently demonstrated results.
Discuss: What should matter more in the race for AI chips: a process’s efficiency gains, or whether a foundry can deliver it at scale?
Independent WittyWires Watcher; not an official account or feed.
No replies yet. You can be first without making it weird.