23 Daily — According to leiphone.com, domestic third-party AI infrastructure (AI Infra) companies in China completed 20 rounds of financing in the first 10 months of 2026, demonstrating a capital concentration towards leading enterprises. Despite global skepticism around financing models of new cloud providers, China's strong demand for computing power orchestration and inference services continues to attract substantial investment.
Capital accelerates flow to leading players, forming a “1-9 effect”
Data collected by Leifeng Network shows that eight domestic AI Infra companies completed 20 financing rounds in the first ten months of this year, with significantly increased frequency. Investor He Dong noted that the industry’s "head effect" is more pronounced this year, with approximately 90% of funds flowing to 10% of top enterprises, forming what is called a "1-9 effect." Matrix Technology CEO Huang Chaobo stated this is a normal phenomenon as industry directions clarify and funds concentrate on the strongest competitors, though competition intensifies. Yan Jiacheng, co-founder of IO Capital, analyzed that China's push to catch up with the U.S. through infrastructure development is a key driver sustaining investment enthusiasm in domestic AI Infra.
Technical and market challenges for third-party AI Infra companies
The selected eight third-party AI Infra companies cover computing power scheduling, model deployment, inference optimization, and token services. Their core advantage lies in integrating computing power resources and providing efficient technical services. They do not produce chips nor necessarily build their own data centers, focusing instead on helping customers optimize computing power usage. However, compared to large tech companies, these third parties face significant substitution risks, as giants possess abundant computing resources, models, and customer ecosystems. Secondary market investor Chang Le believes that current opportunities stem from resource shortages and capability gaps, but some services may eventually be replaced by large firms or open-source tools.
Capital divergence alongside market opportunities
Investors hold divergent views on the long-term value of third-party AI Infra; some think they will ultimately be supplanted by large companies, while others believe they maintain positions among SMEs and certain industry ecologies. Wang Ming points out that large model companies and independent inference infrastructure operate with different business logic and sales models, indicating possible complementary spaces. Yan Jiacheng adds that decoupled industrial clusters and provincial computing power centers, due to dispersed resources and heterogeneous suppliers, require cross-provider unified management and scheduling systems, leaving room for third-party enterprises. Moreover, large cloud providers allocate high-end GPU resources preferentially to major clients, forcing small and medium enterprises to turn to third-party providers, reflecting current uneven market resource distribution.
Editor's take: Challenges and validation paths behind the investment boom
The third-party AI infrastructure market’s capital enthusiasm reflects high expectations for computing power orchestration and inference services but comes with onerous operating costs and competitive pressures from tech giants. For entrepreneurs, key success factors include improving operational efficiency and cost control under tight resources, as well as building technical and service moats through partnerships up and down the supply chain. Validation of these opportunities can be pursued by monitoring corporate client composition, token consumption volumes, and post-financing industrial layout shifts, analyzing whether companies can continuously acquire and reliably serve customers, especially SMEs and region-specific computing needs. Additionally, keeping an eye on discrepancies between financing valuations and actual profitability helps guide investment and entrepreneurial strategies rationally.