CATL-Backed Consortium Leads Nine-Figure Funding Round in Chinese AI Compute Broker Gongji Technology
TREE NEWS reports: Beijing-based Gongji Technology (共绩科技) announced on November 3 that it has closed a new funding round totaling several hundred million yuan. The round drew participation from a consortium including the China Internet Investment Fund, Ningbo Wending (a wholly owned subsidiary of battery giant CATL), Puquan Capital, Yuanhe Puhua, Saimu Technology, Deshi Investment and Qiheng Gongchuang. Existing shareholders Chunhua Capital, Lihuang Venture Capital and Xinjin Venture Capital followed on, with Lighthouse Capital serving as exclusive financial advisor. The company did not disclose the specific round designation or post-money valuation.
Founded in 2023 by a core team from Tsinghua University—founder Fu Zhi studied under Chinese Academy of Engineering academician Jiang Yi—Gongji operates two platforms: Gongji Cloud, which supplies elastic GPU compute, cloud hosts and bare-metal services to model developers, trainers and AI application builders; and Gongji Compute-Network Federation, which provides resource access, monitoring, management and cross-node scheduling for compute owners. The model aggregates idle capacity from disparate holders to improve utilization and sell elastic compute to demand-side customers. Revenue split between the two businesses was not disclosed.
Why CATL Is Interested
The strategic logic extends beyond pure AI. Ningbo Wending is a wholly owned CATL subsidiary, and Puquan Capital shares equity ties with the battery maker. On September 15, CATL announced that Ningbo Wending would commit 1.12 billion yuan to a 5.768 billion yuan Xiamen-based venture fund targeting new-energy supply chains, artificial intelligence and embodied intelligence applications. Gongji’s business also reaches into compute-power coordination: its “compute-for-storage” concept schedules flexible workloads to periods and regions with abundant green power, and the company has built dispatch platforms in Qinghai, Hebei and the Greater Bay Area.
Policy tailwinds are real. Compute-power coordination was written into the government work report for the first time in 2026 and incorporated into the 15th Five-Year Plan outline. On October 30, the National Energy Administration projected that national compute electricity consumption would grow by more than 100 billion kWh per year during the 15th Five-Year Plan period, reaching roughly 800 billion kWh by 2030—about 6% of total social electricity use.
Market Implications
Equities: The deal reinforces a broadening narrative that China’s AI infrastructure buildout is spilling into the energy and industrial sectors. CATL’s participation signals that battery and power-management majors see data-center power demand as a strategic adjacency. Listed Chinese compute-leasing and IDC names could see sentiment support, though Gongji itself is private and the round’s size, while meaningful, is small relative to sector capex.
Commodities: The compute-power linkage is structurally bullish for electricity demand and, at the margin, for natural gas, coal and grid equipment in China. Green-power scheduling schemes could support renewable curtailment reduction, a modest positive for solar and wind operators.
Crypto and decentralized compute: Gongji’s federated scheduling model mirrors the pitch of decentralized GPU networks, but it is centralized and enterprise-focused. Its traction—annual recurring revenue above 200 million yuan, revenue growth exceeding 800%, and more than 12,000 enterprise customers including Zhipu, ModelBest, Tencent and China Mobile—suggests real demand for compute brokerage. That is a competitive data point for tokenized compute projects, though it does not directly validate on-chain settlement models.
Bonds and FX: Negligible direct impact. The read-through is indirect: continued policy support for AI and power infrastructure supports the case for Chinese credit issuance in the technology and utility sectors.
Key Takeaways for Investors
- CATL is using venture capital to position itself at the intersection of energy and AI compute—an adjacency worth monitoring across its supply chain.
- Gongji’s reported 800%+ revenue growth and 200 million yuan ARR are striking, but the company has not disclosed the comparison base, ARR calculation methodology, or audited revenue.
- The compute-brokerage model faces execution risk: heterogeneous hardware, variable availability and network conditions make quality-of-service guarantees difficult for continuous training and inference workloads.
- Policy support for compute-power coordination does not guarantee that independent scheduling platforms capture commensurate market share. Watch actual dispatchable capacity, utilization and recurring revenue durability.
- For crypto investors, the deal is a reminder that centralized, enterprise-grade compute marketplaces are scaling quickly—raising the bar for decentralized alternatives to prove cost and reliability advantages.
