I'm questioning technical risks such as BEOL transistors and 2T DRAM cell structure, not the economics. Cerebras has already retired their technical risk.

It's risky, like landing a rocket, but not impossible.