
Token Factory SiliconFlow's HK IPO Application Accepted for CSRC Filing: Over 10 Million Registered Users, Nearly 2.9 Billion Yuan Raised This Year
SiliconFlow's Hong Kong IPO application has been accepted for CSRC filing; it has over 10 million registered users, raised nearly 2.9 billion yuan this year, and positions itself as an independent ecosystem token supply platform.
Key Takeaways
- Key Highlight:SiliconFlow's Hong Kong IPO application has been accepted for CSRC filing; it has over 10 million registered users, raised nearly 2.9 billion yuan this year, and positions itself as an independent ecosystem token supply platform.
- Key Topics:Token, Factory, HK, IPO
SiliconFlow's Hong Kong IPO application has been accepted for filing by the China Securities Regulatory Commission (CSRC). Founded in 2023, the company positions itself as an independent ecosystem token supply platform, with over 10 million registered users and more than 13,000 enterprise clients. Its core business is not building its own large models or directly selling computing power; instead, through its self-developed inference engine, it uniformly schedules heterogeneous chips from vendors such as NVIDIA, Huawei Ascend, Muxi, and Moore Threads, enabling mainstream models like DeepSeek, Qwen, and Kimi to run on pooled computing power, and then provides services on a per-token basis.
The value of this model lies in decoupling computing power, models, and user demand. For model developers, it allows them to reach a broader user base without building expensive infrastructure themselves. For chip makers, heterogeneous chips can be uniformly scheduled, lowering the barrier to ecosystem adaptation. For users, paying per token is more flexible than purchasing GPUs outright or renting entire machines, making it especially suitable for scenarios with highly fluctuating inference demand.
From the perspective of fundraising pace, raising nearly 2.9 billion yuan this year indicates that the capital market has a relatively high level of recognition for the AI infrastructure services sector. However, it should be noted that such platforms are highly dependent on upstream chip supply and downstream model ecosystems, and face competition from cloud providers' self-developed inference services. The filing acceptance is only the first step; the company must still pass the Hong Kong Stock Exchange hearing. Whether its business model can sustain profitability will be a key focus for the market.
Overall, SiliconFlow's progress reflects the trend of increasing division of labor in the AI industry: computing power, models, and applications are accelerating their separation, and independent token supply platforms are expected to become key hubs connecting all parties. If it successfully lists, it will provide a valuation benchmark for similar companies and may also push more computing power service providers toward public markets.
This page provides an editorial summary based on publicly available information. It is not a republished article. Use the source link below for the original report.