-
Chinese founders are more like top-tier executors on “known problems” than people who come up with questions nobody has thought of before. The most successful Chinese founders in history (such as Jack Ma, Ren Zhengfei, and Liu Qiangdong) were actually on the “atypical” path, but the current system is screening out people like them.
-
Shenzhen’s hardware “underground factories” are astonishing. Engineers take apart Western premium products, reverse-engineer them, and the supply chain is highly concentrated (more than 70% of hardware inputs come from the Greater Bay Area). The iteration speed is far beyond the West. A lot of companies follow the DJI path—start with niche consumer hardware (electric wheelchairs, lawn-mowing robots, etc.), get to hundreds of millions of dollars in revenue, and then expand horizontally.
-
Open-source models are impressive, but closed-source large language models still lag far behind the West, and the gap may widen (GPU constraints, distillation being blocked by the West, and a huge CapEx gap).
-
Software startups are mostly former ByteDance PMs/researchers, copying Western agentic/ambient consumer software and targeting overseas markets, but they can very easily be replaced with one click by large language model labs.
-
Almost all founders are using Claude, following the Dwarkesh podcast, and know the Silicon Valley ecosystem inside out, while prioritizing the global market (especially the West) rather than starting with the Chinese market.
-
The founders really worth investing in are the “outlier” founders who do not take the big-company, degree-driven path, and local VCs are currently barely digging for people like this.
