The Financial Times has reported that China’s AI investment surge had become a correction by October 2026, leaving start-ups with less funding and investors reassessing monetisation timelines. The shift matters beyond venture capital because enterprise AI buyers depend on suppliers that can fund model development, compliance and support.
Funding and valuations
Venture capital deployment into Chinese AI companies fell from USD 18.4 billion in the first three quarters of 2025 to USD 6.1 billion in the same period of 2026. That is a 67 per cent year-on-year decline.
DeepSeek, Zhipu AI and MiniMax have seen valuations cut by 40 to 60 per cent in recent funding rounds. The stated pressure is commercial rather than technical alone. Many companies are still spending heavily while generating limited commercial returns.
Regulation and investor caution
China’s regulatory environment has added to the funding pressure. New data governance rules introduced in early 2026 require compliance infrastructure, which increases operating costs for AI start-ups.
International investors have also pulled back from the Chinese AI market because of geopolitical tensions and export control uncertainty. That reduces another source of capital for companies that need large and continuing investment.
Large technology groups pull back
The correction is not confined to start-ups. Baidu announced a 23 per cent reduction in its AI division headcount in September 2026. Alibaba Cloud paused expansion of its model-as-a-service offerings.
Tencent shifted internal resources away from frontier model development towards AI applications in gaming and advertising. That points to a more selective phase, where commercial use cases are taking priority over broader model-building commitments.
Enterprise demand has lagged forecasts
Enterprise adoption in China has not reached earlier expectations. Large language model adoption remains below 12 per cent among Chinese businesses, compared with analyst expectations of 40 to 50 per cent penetration in 2026.
High deployment costs, integration work and data sovereignty concerns are listed as reasons for slower adoption. For technology buyers, those constraints are familiar. A model subscription is only one part of the cost; data preparation, security review, integration and operating responsibility decide whether a project reaches production.
Global comparison
The Chinese correction does not mean the same pattern has already played out elsewhere. U.S. AI start-ups raised USD 67 billion in the first half of 2026. The difference is that investor scrutiny over profit prospects has increased, and several American AI companies have delayed IPO plans.
For enterprise customers, the practical question is supplier durability. A vendor that depends on continued funding may change pricing, narrow its roadmap or reduce support if capital becomes harder to obtain. That risk belongs in procurement and architecture decisions, not only in finance discussions.
The fall in Chinese AI funding, lower start-up valuations and weaker-than-expected enterprise adoption are capital-allocation issues addressed in ECS's CxO + Strategy topic, alongside keynotes, peer panels and closed-door roundtables for decision-makers. At the conference, attendees can compare investment assumptions with practitioner experience, product group contact and sessions that connect AI plans to governance and operating models. Bring the AI business cases, supplier questions and deployment assumptions that need testing before budgets are committed.
This article was written by the ECS editorial team with the assistance of Amal Agent, run.events' AI assistant, which supported the research, drafting and proofreading.