SiliconFlow has completed the second phase of its Series B+ round and its Series C round, bringing the company's cumulative equity financing in 2026 to nearly 2.9 billion yuan, or about US$433 million. The funding attracted a broad group of national funds, central and local state-owned capital, industrial capital, and professional investment institutions. The announcement, confirmed by China Renaissance, underscores growing investor confidence in the company's token supply platform and AI infrastructure strategy.
Investor Composition and Strategic Backing
Participants in the second phase of the Series B+ round and the Series C round included the China Internet Investment Fund, China Reform Fund, China Mobile Chain Fund, China Orient Asset International, and regional funds such as the Beijing Energy Fund and the Shanghai Yidian Intelligent Computing Fund. Existing investors also increased their commitments. The resulting shareholder base brings together strategic direction, industrial coordination, and market-driven capital.
Funding Scale and Disclosure
China Renaissance confirmed that the company has not disclosed the individual size of either the second phase of the Series B+ round or the Series C round. The newly announced amount reflects total equity financing completed during 2026 rather than a single round. SiliconFlow said the proceeds will mainly support research and development in inference engines, heterogeneous computing power scheduling, and model and chip adaptation.
Market Position and Platform Growth
Frost and Sullivan data shows that SiliconFlow was China's largest independent ecosystem token supplier by annual token throughput in 2025, ranking among the top five token suppliers overall in the country. The company launched its serverless token service in May 2024, and the token factory model has since gained consistent market traction. By April 30, 2026, registered users exceeded 10 million, and by June 21, 2026, the company had served more than 13,000 enterprise clients.
Model and Chip Coverage
SiliconFlow has cumulatively supported more than 170 mainstream AI models and adapted to more than 10 chip types from vendors including Nvidia, Huawei Ascend, Muxi, and Moore Threads. The company has built deep cooperation with customers in AI, internet, energy, finance, and telecommunications, as well as with chipmakers and large model developers. This ecosystem reach has strengthened its role as an independent and open platform for AI inference.
Commercial Momentum and Financial Improvement
SiliconFlow's commercialization accelerated in 2026 as serverless token services, dedicated instances, and local deployment solutions all recorded faster growth. Revenue from enterprise clients and overseas business continued to increase. As business scale expanded and inference efficiency improved, the company reported steady progress in its revenue structure and gross margin performance.
Executive Perspective
SiliconFlow founder and CEO Yuan Jinhui said that AI technology is entering production environments at a faster pace, pushing customers to demand higher scale, efficiency, and stability from token supply. He added that the company will continue to pursue an independent and open platform positioning while converting its inference optimization and heterogeneous computing scheduling capabilities into more dependable and cost-efficient token services. Yuan also emphasized plans to accelerate global market expansion and work with partners to build open AI infrastructure.
SiliconFlow's latest financing round highlights the growing importance of token supply infrastructure as AI adoption moves into enterprise production. The participation of state-backed and industrial investors provides both capital and strategic alignment for future growth. With strengthened research capabilities and an expanding global footprint, the company is positioned to reinforce its role in the next phase of AI infrastructure development.