OpenESSAgent is seeking 3 million yuan in angel round financing to enter the vertical large model track for energy storage engineering.

The domestic energy storage installed capacity continues to grow at a high speed. By the end of 2025, the cumulative installed capacity of new-type energy storage will reach 136 million kilowatts, with a year-on-year growth of over 80%. Independent energy storage and industrial & commercial energy storage have become the main force of the market, and industry competition is shifting from hardware sales to the competition of engineering delivery efficiency and software platform capabilities. However, behind the rapid expansion, the energy storage engineering delivery sector has long faced structural pain points: the training cycle for senior energy storage engineers is as long as 6-12 months, and the growth rate of supply is far behind the growth of projects. More than 30% of projects have rework and communication costs. It takes 7-15 days to manually complete a single project plan. The BMS and PCS protocols of different brands are not unified, which leads to long on-site joint commissioning time, and parameter configuration errors easily cause project rework and delay. Industry experience is scattered among individual engineers, making it difficult to form standardized assets. Among the existing products on the market, general large models lack the support of energy storage electrical mechanism and grid connection rules, and the output results cannot be used for commercial purposes; traditional energy storage software only has process management capabilities, cannot generate engineering results to reduce core labor costs; self-developed systems of leading manufacturers only serve internal teams, with concerns of closed ecology and lack of neutrality; pure algorithm teams lack engineering implementation experience and are difficult to solve practical industrial problems. The market is in urgent need of a productivity tool that can directly generate commercial engineering results, and the OpenESSAgent vertical large model project for energy storage engineering came into being to enter this blank track.

The OpenESSAgent project has been launched for nearly a year. Its founder Lu Jixiong has R&D experience in Guoxuan Hi-Tech's mobile energy storage projects, and is responsible for the overall technical architecture, R&D of the vertical large model for energy storage and product iteration. The co-founder Xie Yingjie has experience in consecutive entrepreneurship and government-enterprise business delivery. The team covers full-chain capabilities including hardware engineering, communication protocols, AI algorithms and commercialization. In the early stage of R&D, the team faced the problem of data and protocol standardization caused by the fragmentation of equipment in the energy storage industry. Instead of simply relying on large models to generate content, relying on the communication engineering experience accumulated by the founder, the team took the lead in establishing a layered device protocol library to precipitate the protocols, boundary conditions and industry experience of avoiding common pitfalls of mainstream devices. At the same time, the vertical large model and the protocol engine are designed to be decoupled: the model is responsible for scheme generation, and the dedicated engine completes protocol verification, adaptation and conversion. By adding an engineering verification layer to suppress large model hallucinations, the core Demo development has been gradually completed and certified, which can realize full-process function demonstration.

As a vertical AI agent for energy storage engineering delivery scenarios, OpenESSAgent has a number of unique technical advantages: it adopts a three-layer decoupled AI agent architecture and adds an independent engineering verification layer, which effectively suppresses large model hallucinations and ensures the reliability of commercial engineering output; it realizes cross-domain integration of electrical engineering and communication, and can complete energy storage system design and multi-brand equipment protocol adaptation at the same time; it builds a dual knowledge base of engineering case library and private protocol library of mainstream equipment, forming a continuously iterating data flywheel; it realizes the collaboration of computing and electricity, completes the integration of design, simulation and scheme output, and solves the pain point of switching between multiple software in the industry; core capabilities are supported in SDK form, which can be called by third-party platforms, laying a foundation for diversified monetization. Compared with similar products on the market, OpenESSAgent is not a shell based on general large models, it can directly output implementable engineering results, adheres to the positioning of a third-party neutral player, does not bind its own hardware, can adapt to energy storage equipment of all brands, and enjoys higher industry acceptance. Its unique protocol adaptive engine directly solves the stubborn industry problem of multi-brand equipment joint commissioning, covering the full-link closed loop from scheme design to commissioning and delivery, which is different from the single-point function layout of competitors. Multiple monetization paths also provide more space for commercial expansion. This platform can compress the original 7-15 days of manual scheme output to the hour level, helping customers increase delivery efficiency by more than 60%, reduce on-site commissioning manpower by 90%, automatically output directly usable drawings, BOM lists and equipment adaptation communication files, and realize full-link automated delivery of energy storage engineering.

In terms of market space, the current energy storage industry is moving from informatization and data intelligence to the stage of engineering generative AI. The involution of informatization has ended, and engineering generative AI has opened the second growth curve of energy storage digitalization, with clear track dividends. According to industry forecasts, the market size of the energy storage engineering delivery track reaches the 100-billion-yuan level. With the continuous growth of installed capacity, the demand for cost reduction and efficiency improvement will be further released. At this stage, OpenESSAgent takes small and medium-sized energy storage EPC integrators and energy storage engineering design service providers as its core SaaS subscription customers, expands SDK interface authorization cooperation with PCS/BMS energy storage hardware manufacturers in the medium term, and covers new energy operation and maintenance and integrated energy platform ecological customers in the long term. At present, it has connected with the first batch of pilot customers.

Feedback from the first batch of pilot customers shows that OpenESSAgent's automatic generation capability for schemes, drawings and BOM has significantly shortened the production cycle of pre-bidding schemes for energy storage projects, effectively alleviated the pressure of insufficient engineers of integrators, and the cross-brand equipment protocol adaptive function directly hits the core pain point of multi-hardware mixed commissioning in the EPC industry, with strong customer demand. In response to the improvement demands put forward by pilot customers such as CAD export adaptation, coverage of local grid connection specifications, and compatibility of niche hardware models, the team has clarified the optimization direction, will gradually add the CAD format export function, expand the engineering knowledge base to include local grid connection and electricity price rules, and continuously expand the device protocol library.

OpenESSAgent adopts a composite business model combining subscription, project and authorization, based on SaaS annual fee charging by account, charges project service fees for projects of different complexity, and at the same time opens SDK interface authorization to charge authorization fees, with value-added services such as simulation verification and commissioning guidance, forming a high-frequency renewable large-scale monetization path. At present, the project is seeking angel round financing of 3 million yuan, releasing 12%-15% equity. The funds will be mainly used for product iteration R&D, benchmark customer market expansion, core team expansion and enterprise operation reserve. It plans to land paid benchmark customers in 6 months, form stable revenue in 12 months, and connect to the next round of financing after completing the commercialization verification closed loop.

Talking about entrepreneurial insights, core members of the team said that the pain points in the energy storage engineering delivery field are real problems from the industrial end. The structural gap that pure AI practitioners do not understand the industry and traditional industry practitioners do not understand AI has given vertical entrepreneurial teams exclusive opportunities. The team will continue to take root in industrial scenarios, build engineering and data barriers, aim to become the AI infrastructure entry in the energy storage engineering field, reconstruct the operation mode of energy storage engineering, enable every energy storage project to quickly move from demand to designability, deliverability and maintainability, and finally reduce cost and improve efficiency for the whole industry chain.