Serenity, the small-cap stock-picker whose reconstructed positions are tracked here, spent a quiet Saturday session sharpening a familiar critique: that the modern US IPO has stopped being a way to fund growth and become a way for early private investors to cash out at inflated marks.

The Anthropic yardstick. His starting point was the eye-watering valuations attached to private AI leaders. He questioned why any public-market buyer would be enthusiastic about backing an AI lab like Anthropic at a rumored $2 trillion valuation — the implication being that by the time such a name reaches public hands, the upside has already been harvested privately. It is a consistent theme for Serenity, who builds his own book around small and micro-cap names he believes are mispriced, not richly valued mega-listings.

China as the counter-example. He pointed to recent and prospective Chinese listings as, in his view, genuine "price discovery" events — moments where the market actually sets a value rather than ratifying a pre-negotiated one. He cited Unitree, the Chinese humanoid-robotics maker, and ChangXin Memory Technologies (CXMT), China's leading domestic DRAM manufacturer. CXMT is the more directly relevant to his portfolio: it plays into his memory supercycle thesis — the idea that AI demand is driving a multi-year supply squeeze in DRAM, HBM and NAND. A newly-public, transparently-priced Chinese memory maker matters to that framework because it offers a fresh read on how the market values raw memory capacity, the same tailwind behind his core memory longs.

The takeaway. Nothing in today's commentary signaled a trade. It was a market-structure observation, reinforcing his preference for hunting overlooked upstream names over paying up for crowded, late-stage stories. For a reader new to Serenity, it underlines his core discipline: value is captured early, and by IPO day the easy money is usually gone.

*This is derived commentary, not investment advice.*