Baidu is moving forward with plans to spin off its AI chip unit, Kunlunxin, and list it in Hong Kong, highlighting the company's push for semiconductor self-sufficiency. On January 1, 2025, Baidu said Kunlunxin confidentially filed a listing application with the Hong Kong Stock Exchange, formally setting the stage for a potential initial public offering, though key details such as the offering size, structure, and timing remain undecided.
Kunlunxin was founded in 2012 as an internal Baidu unit focused on developing AI chips to support the company's search, cloud, and AI model ambitions. Over time, it has evolved into an independently operated business, with Baidu retaining a controlling stake of around 59%. Reuters previously reported that Kunlunxin was valued at around CNY21 billion (US$3 billion) following a recent fundraising round.
The proposed spin-off reflects Baidu's effort to promote Kunlunxin's standalone value and attract investors focused specifically on AI semiconductors. Baidu has been under pressure from slowing ad growth and has stepped up investment in AI, chips, and autonomous driving as new growth engines. A separate listing would also give Kunlunxin more flexibility to raise capital directly and strengthen its market profile with customers, suppliers, and partners. Even after a listing, Kunlunxin will remain a subsidiary.
The timing of the filing is also noteworthy, considering the increase in AI- and chip-related IPO activity in Hong Kong. Beijing is pushing for self-sufficiency in chips in response to escalating US export controls on advanced semiconductors. According to Reuters, LSEG data shows that Hong Kong raised more than US$36 billion from new listings in 2025, its strongest year since 2021.
The Kunlunxin spin-off is significant for the broader industry because it reflects how Chinese technology companies are increasingly separating chip design units from their parent platforms to unlock capital, fine-tune strategic focus, and accelerate product development. While domestic chips like Kunlunxin's cannot yet fully replace leading alternatives, they are gaining traction in inference workloads, government projects, and telecom and cloud deployments where cost, supply stability, and local support matter. A successful listing could provide Kunlunxin with the financial resources and market visibility to scale faster, strengthen software ecosystems, and compete at a higher level.
Article edited by Jerry Chen
