China's AI IPO pipeline stays resilient as investors look beyond infrastructure

What happened

A Citi investment-banking executive said Chinese AI supply-chain IPOs are holding up despite broader concerns about technology valuations.

The banker said demand reflects a relatively balanced mix of domestic and international investors.

Citi expects the next phase of AI listings to include more model developers and AI application companies rather than being dominated by infrastructure suppliers.

Why it matters

Public markets impose a different test from venture funding: investors increasingly expect durable revenue, margins, retention, and a credible path to profitability.

A shift toward AI application IPOs would test whether companies above the infrastructure layer can build defensible software businesses.

Key numbers & facts

• 118 Hong Kong IPOs in the first nine months of 2026
• HK$388B raised
• Fundraising was reported as up 106% year over year

Practical implications

• For startup builders, focus on proprietary workflow integration, distribution, data advantages, reliability, and measurable customer ROI.
• Track gross margin and inference cost early instead of treating compute expense as an afterthought.
• Design products whose value survives model commoditization.

Likely future impact

Analysis, not a confirmed outcome.

• AI infrastructure companies may continue to be judged on capital efficiency while application companies are increasingly judged like mature SaaS businesses.
• Public-market scrutiny could reduce valuation premiums for companies whose only differentiation is access to the same underlying models as competitors.

Evidence

• Reuters reports Citi's market assessment and the Hong Kong IPO fundraising data.