AI & Tech
Patel predicts AI compute growth could hit a wall as leading-edge wafers shift
Dwarkesh Patel Podcast
Why smarter AI models could drive up compute prices 10x
"This is probably going to hit a wall by the end of next year when at the leading edge N3 nodes at TFMC AI will have gone from 60% to 86%. At some point you have just absorbed all leading edge way for capacity for AI and you can't keep increasing this number."
Dwarkesh Patel argues that sustaining roughly 3x year-on-year growth in AI compute may be difficult, citing constraints he attributes to three drivers: Moore’s law, new fab build-out, and wafer allocation shifting from consumer devices to AI. He predicts that at TSMC’s leading-edge N3 node, AI’s share could rise from 60% to 86% by the end of next year, after which further reallocation becomes much harder.
From this episode
Dwarkesh Patel Podcast