Gongji Technology has secured hundreds of millions of yuan in financing to accelerate the deployment of distributed computing power networks.
As large models and AI applications evolve from technological exploration to large-scale implementation, the computing power market is undergoing a structural qualitative change from "planable" to "highly elastic".
Recently, 「Beijing Gongji Technology Co., Ltd.」 (hereinafter referred to as "Gongji Technology") announced the completion of a new round of financing of several hundred million yuan. The investor lineup covers national-level funds, industrial capital and a number of market-oriented institutions, including China Internet Investment Fund, CATL-affiliated Ningbo Wending and Puquan Capital, Yuanhe Puhua, Saime Technology, Deshi Investment, Qihong Gongchuang, etc. Old shareholders Chunhua Venture Capital, Lihe Zhongke and Xinjin Venture Capital continued to increase their investment. Lighthouse Capital acted as the exclusive financial advisor.
The funds from this round of financing will be mainly invested in the iteration of core technologies of distributed computing power network, industrial ecological expansion and commercial system construction, continuously expand the access scale of key nodes of the computing power network, consolidate the technical barriers of the platform, accelerate the interconnection and efficient scheduling of cross-domain heterogeneous computing power, and further amplify the large-scale delivery and intelligent operation capabilities of computing power services.
The mismatch between supply and demand of computing power is prominent, and distributed scheduling becomes the key to breaking the deadlock
Under the wave of AI commercialization, the contradiction between the supply and demand sides of the computing power market continues to expand.
Demand side, the main body and form of computing power consumption are undergoing qualitative changes, and elastic demand runs through all links of the AI industrial chain. On the one hand, the computing power demand for model training and reasoning fluctuates drastically; on the other hand, data engineering, agent orchestration, retrieval augmentation and various vertical applications continue to generate new computing power consumption, and the demand changes with the rhythm of real business. The overall demand for computing power has shifted from relatively stable and pre-plannable in the past to a highly dynamic elastic state that scales in real time with business.
Supply side, a large number of intelligent computing resources are scattered in various intelligent computing centers and heterogeneous clusters, the scheduling systems are isolated from each other, resulting in high resource flow cost and insufficient flexibility. Limited by the hardware procurement cycle and the upper limit of its own resource pool, traditional public clouds cannot balance the three major demands of elastic expansion of computing power, controllable cost and stable delivery.
Facing this structural mismatch, Gongji Technology, founded in 2023, put forward the distributed computing power scheduling concept and elastic computing power service model, taking distributed computing power scheduling as the entry point, building a unified scheduling capability across suppliers and regions on the basis of the traditional public cloud model, breaking the boundary of a single resource pool, enabling computing power to flow on demand and be available whenever needed, and promoting the optimal matching of computing power resources in time and space dimensions.
Behind this cross-domain scheduling concept is the team's solid accumulation of underlying technologies. The core founding team of Gongji Technology comes from Tsinghua University, deeply engaged in the field of complex resource scheduling, with mature scheduling methodology and high-intensity engineering delivery capability. Based on the distributed computing power scheduling technology, Gongji Technology connects diversified computing power resources to create a "Gongji Cloud Platform + Gongji Computing Network Federation Platform" dual-engine business architecture, building a complete service link from unified access of computing power resources, intelligent scheduling to stable business delivery, helping the whole society's computing power supply move to a new stage of higher efficiency, inclusiveness and greenness.
The dual-engine layout opens up the full link, and the coordinated development of computing and electricity unlocks long-term space
The Gongji Cloud Platform and Gongji Computing Network Federation Platform cooperate with a clear division of labor, which is the core carrier for the implementation of Gongji Technology's self-developed distributed computing power scheduling technology, and also constitutes the differentiated competitiveness of the enterprise.
Among them, Gongji Cloud Platform is oriented to computing power demanders, focusing on the delivery and use of computing power services, providing products such as elastic service deployment, Job batch processing, large model cloud services, cloud hosts and bare metal, covering the full link of model development, training, reasoning and AI applications, and providing customers with elastically scalable GPU resources. Gongji Computing Network Federation Platform focuses on the connection, management and scheduling of computing power resources, providing supporting capabilities such as distributed computing power scheduling, computing power monitoring and other computing power management, to support the access, management, operation and maintenance and efficient configuration of large-scale computing power resources.
Relying on this dual-platform system, Gongji Technology creates deterministic value for both sides of computing power supply and demand. For computing power holders, the platform helps them revitalize scattered and underutilized computing power assets, import idle computing power into AI business scenarios, and drive the improvement of asset turnover rate and commercial returns. For computing power demanders, the platform realizes on-demand use of computing power through elastic scheduling capability, helping enterprises get rid of the pressure of heavy asset procurement of GPUs, and significantly reducing the threshold and comprehensive cost of computing power use.
At present, this scheduling concept and technical solution of Gongji Technology have been verified in multiple scenarios, and it has deeply participated in the construction of multiple national and provincial computing power scheduling platforms, successively undertaken computing power scheduling platforms in Qinghai, Hebei, Guangdong-Hong Kong-Macao Greater Bay Area and other regions, and continuously implemented benchmark projects such as distributed computing power operation, idle computing power scheduling, and coordinated computing and electricity. At the same time, its partners not only cover leading AI enterprises such as Zhipu, Minimax, VAST, but also large industrial entities and scientific research institutions such as Tencent, China Mobile and Tsinghua University, serving more than 12,000 enterprises and teams in total, with the scale of computing power resource pool and commercial delivery capability ranking steadily at the forefront of the industry.
Relying on the asset-light distributed computing power scheduling model, the company has achieved a high gross profit margin of business with strong growth momentum: annual revenue growth exceeds 800%, it achieved continuous quarterly revenue doubling in 2026, with ARR exceeding 200 million yuan, taking into account both growth speed and operating quality, with outstanding risk resistance.
On the basis of continuously consolidating the computing power scheduling capability, Gongji Technology further broadens the boundary of scheduling capability: on the one hand, lay out the coordination of computing and electricity, coordinate computing tasks, computing power and green power resources in a larger time and space scope, promote off-peak task scheduling, match task value with power value, improve the overall utilization rate of computing power, and help the consumption of new energy and the implementation of the "dual carbon" goal; on the other hand, for the future diversified heterogeneous nodes, prospectively explore new computing power forms such as vehicle-mounted computing power and space computing power. The company is committed to building a distributed computing power network that can schedule all-domain diversified heterogeneous resources, so that computing power from computer rooms, vehicles, space, deep sea and various end-side devices can realize on-demand flow and collaborative scheduling.
Computing power is the key productive force in the intelligent era. Standing at the threshold of global intelligent leap, the construction of computing power infrastructure has ushered in a new development cycle. Gongji Technology will continue to deeply cultivate the core technology of distributed computing power scheduling, deeply integrate into the national computing power strategy, amplify the network effect of distributed computing power network, consolidate its own competitive barriers, and realize the mutual gain of commercial value and social value.
China Internet Investment Fund said: "The large-scale implementation of AI applications is reshaping the demand for computing power, and elastic scheduling and cross-domain collaboration have gradually become one of the key capabilities of computing power infrastructure. Gongji Technology's innovative exploration in the field of distributed scheduling improves the utilization efficiency of computing power by connecting scattered computing power resources, and has development potential. Under the background of accelerating the construction of the national integrated computing power network, we look forward to the company continuing to deepen technological innovation, continuously exploring feasible paths for efficient collaboration of computing power, contributing to promoting the optimal allocation of computing power resources, consolidating the base of digital infrastructure, and helping the high-quality development of the digital economy."
Yuanhe Puhua said: "The second half of AI belongs to reasoning, and the lifeblood of reasoning is the elastic, low-cost and stable supply of computing power — which is exactly the 'impossible triangle' that the industry has long been unable to achieve at the same time. Gongji Technology breaks the deadlock with its original grid-like computing power scheduling network, gathering idle computing power scattered in Internet cafes, intelligent computing centers and even thousands of households into a 'computing power grid' that can be used at any time. It has served more than 12,000 enterprises, reducing the computing power cost for customers by half. The Gongji team comes from Tsinghua University, with both top scheduling algorithm skills and rare engineering implementation capabilities. We firmly believe that the Gongji team can constantly break through their own margins in the AI tide and make inclusive computing power a reality."
Saime Technology said: "With the surge of computing power, the infrastructure comes first. At the moment when physical AI is accelerating the reshaping of thousands of industries, computing power has become the basic 'water and electricity' supporting the intelligent world. In the process of building the underlying infrastructure of physical AI and intelligent systems, Saime deeply understands the difficulty of achieving the ultimate balance of 'elasticity, stability and cost' in computing power supply. When we see Fu Zhi's team using 'grid scheduling' to gather idle computing power into a huge force, showing the characteristics of accurate direction and solid engineering, we feel a strong resonance. As an industrial partner who has traveled all the way, I would like to share three insights with Gongji for mutual encouragement: First, stay awed. The essence of computing power scheduling is system reliability engineering. Only by staying awed of every millisecond of switching and every supply-demand match, and not being blindly confident, can we go further. Second, keep learning. AI demand forms are changing rapidly, and yesterday's experience may not be suitable for tomorrow. Continuous in-depth learning from the industry and updating the cognitive framework is more important than chasing the trend. Third, maintain independent judgment. When the consensus in the computing power track is crowded, what really determines the value is the concentration to identify a difficult but correct path before the consensus is formed, and move forward steadily. We look forward to Gongji moving steadily and far on the road of inclusive computing power."
Deshi Investment said: "Artificial intelligence is accelerating from model training to large-scale reasoning. The continuous growth and diversification of computing power demand has brought broad opportunities for the new generation of computing power infrastructure. Gongji Technology takes distributed computing power scheduling as the core, connects the scattered computing power supply and growing application demand, and explores a more efficient and more economical computing power service model through elastic computing power supply and reasoning services. We also look forward to Gongji's continuous exploration in the field of coordinated computing and electricity. Deshi Investment is optimistic about the team's technical insight and commercial execution ability, and looks forward to walking with this group of young entrepreneurs to jointly explore the new paradigm of infrastructure in the intelligent era."