High-end computing power prices rise, leading firms explore 'Token Factory' model
The high-end computing power sector is experiencing significant price increases, with leading companies like Xingyun Technology announcing not only order increases and price hikes for previously signed large contracts but also a shift in their charging model. The new model introduces a "fixed service fee related to Token (token) revenue." This development follows a trend observed throughout the year, as over ten listed companies, including Runjian Co., Ltd., Hongxin Electronics, Xingyun Technology, Chaoxun Technology, and Nanwei Software, have announced their exploration of the "Token Factory" business. Industry insiders note a substantial change in the business model within the computing power sector, moving from a focus on hardware accumulation to selling Tokens. This transition signifies a shift from hardware leasing to Token-based delivery, with the "Token Factory" model at its core. The competitive focus in the industry is evolving from GPU ownership to TaaS (Token-as-a-Service), where billing is based on Tokens. Consequently, the commercial model is upgrading from selling resources to selling services.
The computing power industry's pivot from hardware-centric sales to a Token-as-a-Service (TaaS) model reflects a maturing market seeking higher-value revenue streams. This shift from selling raw capacity to selling metered output (Tokens) suggests a response to increasing demand and the high cost of specialized hardware like GPUs. By charging based on Token generation, companies aim to capture more value from their infrastructure investments and align costs more directly with customer usage. This evolution could lead to more dynamic pricing and service offerings, potentially benefiting end-users who require specific computational outputs rather than just hardware access. However, it also introduces complexity in managing and verifying Token generation, which will be crucial for maintaining trust and market stability in the long term.
AI-generated to prompt reflection — not editorial opinion, not advice, not a statement of fact. How this works.