Key Takeaways
- • Beijing-based Gongji Technology announced on Nov. 3, 2026 that it completed a funding round worth hundreds of millions of yuan, with the exact round designation and post-money valuation undisclosed.
- • CATL-linked capital entered the round through Ningbo Wending, a wholly owned CATL subsidiary, and Puquan Capital, which has equity ties to CATL.
- • The company reports annual revenue growth above 800%, ARR above 200 million yuan (approximately $30.0 million), and more than 12,000 enterprises and teams served, though the comparison base and ARR methodology were not disclosed.
- • Proceeds will fund distributed computing network iteration, node access expansion and commercialization; scheduling platforms operate in Qinghai, Hebei and the Guangdong-Hong Kong-Macao Greater Bay Area, with scale and revenue not separately disclosed.
- • Policy tailwinds include computing-power coordination entering the 2026 government work report and the 15th Five-Year Plan, with the National Energy Administration projecting national computing-power electricity use near 800 billion kWh by 2030.
Summary
Beijing Gongji Technology Co. announced on Nov. 3 that it has completed a new funding round totaling hundreds of millions of yuan. Investors include the China Internet Investment Fund, Ningbo Wending, Puquan Capital and Yuanhe Puha, while existing shareholders Primavera Capital, Lihe Venture Capital and Xinjin Venture Capital increased their investments. Simou Technology, Deshi Investment and Qiheng Gongchuang were also named among the investors. Lighthouse Capital served as exclusive financial adviser. According to the company's disclosures, the round's designation and post-money valuation were not revealed.
Founded roughly three years ago, Gongji positions itself as a matchmaker between computing-power supply and demand. Its Gongji Cloud platform offers elastic GPU computing, cloud hosts and bare-metal services for model development, training, inference and AI applications, while its Gongji Computing Network Federation platform provides resource holders with access, monitoring, management and cross-node scheduling to put dispersed idle capacity to use. The company cites partners including Zhipu, ModelBest, Tencent and China Mobile.
The company says annual revenue growth exceeded 800%, quarterly revenue continued to double through 2026, and ARR topped 200 million yuan (approximately $30.0 million). It has cumulatively served more than 12,000 enterprises and teams. However, the revenue comparison base, the ARR calculation methodology and recognized revenue scale have not been disclosed, and these figures should be read as company-reported rather than audited.
The CATL-linked participation is a notable element. Ningbo Wending is a wholly owned CATL subsidiary, and Puquan Capital has equity ties to the group. On Sept. 15, CATL announced that Ningbo Wending committed 1.12 billion yuan (approximately $167.9 million) to the Xiamen Shidai Yiyuan Venture Capital Fund, a fund with total scale of 5.768 billion yuan (approximately $864.7 million) focused on the new energy supply chain and AI and embodied intelligence. Another CATL subsidiary holds 45% of the partnership interest in the fund's manager, Xiamen Puquan. Gongji's proposed "computing as storage" model would schedule flexible computing tasks to periods and regions with greater green-power supply, though the scale and revenue contribution of its scheduling projects remain undisclosed.
Policy support provides context. Computing-power coordination was written into the 2026 government work report for the first time and incorporated into the 15th Five-Year Plan outline. On Oct. 30, the National Energy Administration projected that national computing-power electricity consumption will grow by more than 100 billion kWh annually on average during the plan period, reaching about 800 billion kWh by 2030, or roughly 6% of total societal electricity use. Still, demand growth does not automatically translate into share for independent scheduling platforms, and resources drawn from varied sources raise open questions about performance, continuity and data security for long-running training and inference workloads. Key indicators to monitor include actual schedulable capacity, utilization, recurring revenue sustainability and whether coordination projects become stable commercial revenue.
Founded roughly three years ago, Gongji positions itself as a matchmaker between computing-power supply and demand. Its Gongji Cloud platform offers elastic GPU computing, cloud hosts and bare-metal services for model development, training, inference and AI applications, while its Gongji Computing Network Federation platform provides resource holders with access, monitoring, management and cross-node scheduling to put dispersed idle capacity to use. The company cites partners including Zhipu, ModelBest, Tencent and China Mobile.
The company says annual revenue growth exceeded 800%, quarterly revenue continued to double through 2026, and ARR topped 200 million yuan (approximately $30.0 million). It has cumulatively served more than 12,000 enterprises and teams. However, the revenue comparison base, the ARR calculation methodology and recognized revenue scale have not been disclosed, and these figures should be read as company-reported rather than audited.
The CATL-linked participation is a notable element. Ningbo Wending is a wholly owned CATL subsidiary, and Puquan Capital has equity ties to the group. On Sept. 15, CATL announced that Ningbo Wending committed 1.12 billion yuan (approximately $167.9 million) to the Xiamen Shidai Yiyuan Venture Capital Fund, a fund with total scale of 5.768 billion yuan (approximately $864.7 million) focused on the new energy supply chain and AI and embodied intelligence. Another CATL subsidiary holds 45% of the partnership interest in the fund's manager, Xiamen Puquan. Gongji's proposed "computing as storage" model would schedule flexible computing tasks to periods and regions with greater green-power supply, though the scale and revenue contribution of its scheduling projects remain undisclosed.
Policy support provides context. Computing-power coordination was written into the 2026 government work report for the first time and incorporated into the 15th Five-Year Plan outline. On Oct. 30, the National Energy Administration projected that national computing-power electricity consumption will grow by more than 100 billion kWh annually on average during the plan period, reaching about 800 billion kWh by 2030, or roughly 6% of total societal electricity use. Still, demand growth does not automatically translate into share for independent scheduling platforms, and resources drawn from varied sources raise open questions about performance, continuity and data security for long-running training and inference workloads. Key indicators to monitor include actual schedulable capacity, utilization, recurring revenue sustainability and whether coordination projects become stable commercial revenue.
Go deeper than the headline
Research every firm in this story, track the ones that matter and get the daily briefing — free for 7 days.