Kailian Capital's 2021 bet on China's DRAM champion ChangXin pays off after IPO
What's the deal? Kailian Capital invested US$30.9M (≈$33.7 million) in ChangXin Memory Technologies, a Chinese DRAM chipmaker, in a late-stage venture round disclosed in February 2022. The deal valued ChangXin at US$14.8B post-money. The firm is now reviewing the bet after ChangXin went public.
Why now? ChangXin listed on Shanghai's STAR Market this year, at one point reaching a market value of US$412.5B. That milestone turned a long, loss-making investment into a visible return and prompted Kailian partner Yao Ningbo to reflect on the thesis publicly at the 2026 Financial Street Forum.
What's the endgame? ChangXin builds DRAM, now the biggest bottleneck in AI computing — memory accounts for 60% of the reference cost in Nvidia's designs. Kailian backed domestic substitution in a market dominated by Samsung, SK HynixDealroom has a profile for this one. Try Dealroom →, and Micron. In 2021, ChangXin was still deep in heavy capital spending and losses.
The investor's logic: "Semiconductors are a cyclical industry, so you have to invest at the bottom," Yao said (translated from Chinese). "If it's all ablaze like it is now, investing only gets you burned." Yao, a former Intel employee, said chip startups require patience across multiple industry cycles.
Why it worked: The sector hit a trough around 2023, with heavy capital spending squeezing domestic makers while Samsung and SK Hynix cut output to defend prices. Memory prices only recovered in early 2025, pulled up by AI demand, and shortages peaked in 2026. Kailian has since backed AI chip designer Qingwei Intelligence and smartphone SoC maker Unisoc, both in STAR Market IPO coaching.
What could go wrong? Yao warned that current shortages and high prices should not be mistaken for durable competitive strength. ChangXin must keep spending to catch up on technology and scale, and returns still hinge on holding its ground once the cycle turns down. He also flagged overheated valuations in hard tech, where some companies hit billions on little more than an idea.
The signal: The ChangXin story captures the patient-capital case for Chinese semiconductors — buy during losses and downturns, bet on localisation, and wait out the cycle. With DRAM now central to AI infrastructure and global supply tightening, the strategic rationale has only sharpened. But the lesson Kailian itself draws is caution: cyclical upswings are not the same as lasting competitiveness.
Read more: sohu.com
Image credit: Generated with Gemini
