Qujing Technology completes Series A funding led by Henan Investment Group Huirong Fund, raising ove...

Qujing Technology completes Series A funding led by Henan Investment Group Huirong Fund, raising over RMB 1 billion (~$138M) in six months. The startup specializes in producing high-efficiency AI inference tokens, positioning itself as a critical supplier to China’s AI infrastructure layer.

Why it matters: This round exemplifies the capital-compression arc playing out in China’s AI infrastructure segment, where investors are placing large bets on token-production layer players rather than just model labs. Qujing operates in the AI Infrastructure segment, provisioning the compute-to-token conversion stack that foundation model developers and application companies consume. The size and speed of the raise—over RMB 1B in just six months—signals that Chinese state-backed capital is actively consolidating around a small set of infrastructure providers, mirroring the hyperscaler-distribution pattern already seen in the US market with companies like CoreWeave.

The significance: Qujing is not a model lab but a “token factory”—a pure-play infrastructure vendor whose success depends on inference efficiency and throughput, not frontier model breakthroughs. This funding is structurally analogous to the capital waves flowing to US AI infrastructure plays, but with a distinct China-specific dynamic: the lead investor is Henan Investment Group Huirong Fund, a provincial state-owned enterprise vehicle. The involvement of local government capital reinforces the pattern of provincial-level sovereign AI investment, where each region backs select AI infrastructure champions. For the broader substrate, this raises the open question of whether token-production companies can achieve the same scale and moat as hyperscaler-owned clouds, or whether they remain commodity intermediaries squeezed by price compression.