Infinigence AI Closes RMB 2.3 Billion Series C: Cross-Cluster Inference Cuts Costs 37.5%, Confidentially Files for HKEX IPO
Introduction
On October 12, 2026, Shanghai-based AI infrastructure company Infinigence AI (Wuwen Xinqiong) successfully closed a RMB 2.3 billion (approximately USD 344 million) Series C funding round, bringing its total cumulative financing to over RMB 4.3 billion. This funding milestone not only highlights the strong appeal of China's AI inference infrastructure sector but has also attracted significant market attention due to the company's simultaneous confidential filing for an IPO on the Hong Kong Stock Exchange (HKEX).
Funding Details
Investor Lineup
This Series C round attracted a diverse group of investors:
Government-backed investors:
- Hangzhou High-tech Financial Investment
- Social Security Zhongguancun Fund
- Zhuhai Technology Industry Group
- Xuhui Capital
Corporate and venture capital investors:
- Tencent
- Baidu
- Lenovo
- Sequoia China
- Qiming Venture Partners
- Shunwei Capital
The participation of tech giants like Tencent, Baidu, and Lenovo brings not just capital but also signals that Infinigence AI's technology platform has gained recognition from industry leaders, carrying significant strategic partnership value.
IPO Plans
According to reports, Infinigence AI confidentially filed for an IPO on the Hong Kong Stock Exchange in early October 2026, with an objective to list as early as the first half of 2027. If successful, this would make Infinigence AI one of the first companies in China's AI inference infrastructure space to list in Hong Kong.
Core Technology: Heterogeneous Chip Inference Platform
Technical Background
Infinigence AI was founded in 2023 by Tsinghua University professor Wang Yu and his students Xia Lixue (CEO) and Dai Guohao. The company's core mission is to solve the deployment efficiency problem of AI models in heterogeneous computing environments—specifically, how to make AI models run efficiently across chips of different brands and architectures.
Cross-Cluster Inference Architecture
The company's flagship product is an infrastructure platform capable of coordinating 37,000 petaflops of computing capacity across 16 mainstream chip types. Key performance metrics include:
- 51.5% reduction in time-to-first-token: Significantly improving user experience
- 37.5% reduction in cost per token: Dramatically lowering enterprise AI operating costs
- 40-fold increase in daily token usage: From December 2025 to July 2026
Main Product Lines
| Product | Function |
|---|---|
| GenStudio | Provides APIs for various language and vision models |
| AgentWorks | Facilitates the configuration and operation of AI agents |
| ComfyStack | Supports hosted workflows built with ComfyUI |
Technical Strategy for Navigating Export Controls
Against the backdrop of U.S. export controls on AI chips to China, Infinigence AI's heterogeneous computing platform carries special strategic value. By optimizing the software layer to coordinate mixed workloads between domestic chips (such as Huawei Ascend) and international chips, the company provides Chinese enterprises with a viable path to maintaining AI competitiveness in a restricted hardware environment.
Macro Context: China's AI Inference Market
Strategic Shift from Training to Inference
By 2026, China's AI industry has completed a strategic shift from a "parameter race" to "industrial deployment." Inference computing now accounts for 80% to 90% of model lifecycle costs, becoming the primary driver of AI budgets.
According to data from the National Data Administration, China's daily AI token usage reached 140 trillion in March 2026, representing a 1,000-fold increase since the start of 2024. This explosive growth directly drives enormous demand for efficient inference infrastructure.
The "Token Factory" Economic Model
China's AI industry is forming a unique "token factory" economic model, treating AI inference as a commodity infrastructure similar to water or electricity. Infinigence AI's platform is a core infrastructure provider for this model.
Government Support and Cost Advantages
Central and local government policies provide subsidized electricity for data centers, particularly in provinces like Gansu and Guizhou. This government support, combined with Infinigence AI's technical optimization, gives Chinese AI inference services significant cost competitiveness in the global market.
Asia-Pacific Investment Implications
Insights for Asia-Pacific AI Infrastructure Investment
Infinigence AI's funding case offers important insights for AI infrastructure investment in the Asia-Pacific region:
- Investment value of inference infrastructure: As AI applications scale, inference efficiency becomes a key competitive factor, making related infrastructure investments valuable long-term
- Market opportunities in heterogeneous computing: In environments with chip supply constraints, software platforms that can optimize heterogeneous computing hold unique market positions
- HKEX as a destination for AI company listings: Infinigence AI's choice of HKEX reflects Chinese AI companies' confidence in Hong Kong's capital markets
October 2026 AI Investment Landscape
According to market data, the AI sector recorded at least 105 funding deals totaling over $50.6 billion in October 2026. AI-driven startups captured 64% of all global venture capital in Q3 2026, totaling $102 billion.
In this context, while Infinigence AI's RMB 2.3 billion funding is relatively modest in scale, its strategic positioning in the critical inference infrastructure sector makes it one of the most closely watched funding cases.
Competitive Landscape Analysis
Key Competitors
In China's AI inference infrastructure market, Infinigence AI faces competition from:
- Cloud service providers: Alibaba Cloud, Tencent Cloud, Baidu AI Cloud
- Specialized inference optimization companies: Approaching.AI (Token-as-a-Service platform)
- Chip manufacturer proprietary platforms: Huawei Ascend ecosystem
Differentiated Competitive Advantages
Infinigence AI's core differentiation lies in its unified scheduling capability across chip types. Unlike competitors relying on a single chip architecture, Infinigence AI can dynamically allocate workloads across 16 different chip types, maximizing overall computing efficiency.
Conclusion
Infinigence AI's completion of a RMB 2.3 billion Series C round and confidential HKEX IPO filing marks a new development phase for China's AI inference infrastructure sector. Against the backdrop of scaling AI applications, efficient, low-cost inference infrastructure will become a core pillar of enterprise AI competitiveness.
For Asia-Pacific investors and enterprises, Infinigence AI's case provides an important reference: in the AI industry chain, investment opportunities at the infrastructure layer often carry more certainty and long-term value than those at the application layer.
Sources: technode.com (October 12, 2026), thesaasnews.com (October 12, 2026), panews.io (October 12, 2026)
