The thesis at the Shanghai forum
Artificial intelligence is no longer merely the engine of global semiconductor demand: according to industry figures and scientists gathered Friday in Shanghai at a major scientific forum, it is also entering the ways in which chips are designed and built. The South China Morning Post reports this in its Tech section, citing remarks at the event.
The most unequivocal statement is attributed to Chen Yunji, co-founder of AI chipmaker Cambricon Technologies. Chen predicted that the integration of artificial intelligence and chip design will over time lead to fully automating the design process. This is a prediction, not an already achieved result, and should be read as such.
What RSI means and why it matters
Chen pointed to the RSI technique, recursive self-improvement, as the most important development of the next decade. He describes it as the mechanism by which artificial intelligence models improve other artificial intelligence models, and AI chips improve other AI chips.
This is what is new compared with the usual picture: AI is not described only as a market that absorbs semiconductors, but as a tool that acts on the production cycle of those same semiconductors. If the dynamic indicated by Chen were to materialize, the consequence supported by the material is an acceleration within the sector, in which design and silicon feed each other. It remains, however, a ten-year prediction, with no data, intermediate timelines, or concrete examples provided in the source.
Demand and supply intertwine
The material highlights two distinct but connected aspects. On one hand, AI continues to drive demand for semiconductors; on the other, it is entering manufacturing. According to forum participants, the two phenomena should not be read separately: it is the very growth in demand that makes the way chips are made relevant.
The picture that emerges is therefore that of a sector in which the chain between those who design AI systems and those who produce chips tends to tighten. The source does not quantify this integration or indicate specific companies, timelines, or investments beyond Chen's prediction. Cambricon's role is also cited only through its co-founder, without details on products or projects.
What remains to be verified
The statements reported are remarks by an industry entrepreneur at a forum, not verified results or published studies. There are insufficient elements to assess how close full automation of design is, what technical obstacles remain, and how the industrial supply chain is preparing. The material offers no independent corroboration or responses from other experts.
For the reader, the news should therefore be placed at the level of stated prospects: it indicates a direction discussed among industry professionals, but does not demonstrate that it is already under way. Further verification would require data on concrete projects, adoption timelines, and third-party assessments, which the source does not currently contain.
