Did China just beat Intel?

Summary

A teardown of Huawei’s Kirin 9030 reveals that SMIC’s 7nm N+3 process achieved EUV-class density without EUV tools, beating TSMC N6 on raw density and matching Intel 18A on metal pitch. However, power and performance still lag significantly behind leading-edge nodes from TSMC, Intel, and Samsung. Export controls changed but did not stop China’s progress, and SMIC is now licensing its technology to other domestic fabs, potentially shifting the chokepoint from a single fab to an entire ecosystem. The video concludes that China doesn’t have to be TSMC to matter; it just needs to be good enough to not need TSMC at all.

• SMIC's N+3 node achieves 113 MTr/mm², denser than TSMC N6 at 108 MTr/mm², using quadruple patterning and DTCO.
• The Kirin 9030’s minimum metal pitch (32.5 nm) is tighter than Intel 18A’s relaxed high-performance cells (36 nm).
• Performance and efficiency still lag: Huawei’s prime core runs around Arm Cortex-X2 levels, with Apple’s E-cores beating it at 1W vs 4.5W.
• Intel 18A retains overall transistor density and backside power delivery advantages despite the metal pitch headline.
• China cannot close the bleeding-edge gap on power/performance, but export controls failed to stop them; they changed the problem.
• SMIC is directed to license N+2/N+3 to other domestic fabs, turning a single sanctionable fab into an ecosystem.
• If process learning flows into AI accelerators, the choke point becomes a whole ecosystem, not one fab.
• The key takeaway: China only needs to be good enough to not need TSMC, not to beat it.