On June 3, Ant Digital officially launched the "Power Trading Agent" and the "New Energy Operations Agent" at the 2026 Shanghai SNEC Photovoltaic Exhibition. Both products are developed based on its self-developed intelligent agent platform DTClaw, which helps enterprises to build AI employees with industry expert capabilities in bulk. Currently, Ant Digital has completed scenario implementation in collaboration with Linyang Zhivei and Gintech New Energy. It is reported that this is the first time Ant Digital has applied its self-developed intelligent agents in the energy industry.

image.png

New energy power stations have long relied on manual decision-making, which has clear limitations in high-frequency trading markets. Ant Digital, using the DTClaw platform and the EnergyTS3.0 time series large model as the foundation, has collaborated with Linyang Zhivei to implement the first power trading agent case. By integrating Linyang's real operation data from 30 GW of power stations, it has established a complete workflow from prediction, strategy generation, automatic declaration, to risk control and review, achieving unattended autonomous trading around the clock, reducing labor costs by more than 60%, and increasing the speed of analysis and strategy generation by more than 20 times.

At the same time, Ant Digital has implemented the first photovoltaic scenario use case of the operations agent in collaboration with Gintech New Energy. AI is embedded into the maintenance process, connecting the entire workflow from monitoring, alerts, inspection, diagnosis, work orders, archiving, to prediction. In Gintech New Energy's operations, key anomaly identification and feedback time is compressed to within 1 minute, end-to-end inspection and defect elimination efficiency is improved by 50%-70%, and fault location and decision-making time is reduced by 30%-60%, significantly improving the speed and accuracy of maintenance response.

In April this year, Ant Digital launched the professional intelligent agent platform DTClaw for enterprise scenarios, which includes over 100 core expert-level Skills. Based on this, the "Power Trading Agent" and the "New Energy Operations Agent" have been deeply optimized for the complex business scenarios in the energy industry, forming a team of "AI employees" covering multiple scenarios, allowing enterprise customers to "use out of the box". In addition, thanks to the iteration of the self-developed EnergyTS3.0 energy time series large model, the product has reached industry-leading levels in four scenarios: photovoltaic power generation forecasting, wind power generation forecasting, electricity load forecasting, and day-ahead electricity price forecasting. In multiple provinces, the highest power forecasting accuracy rate has reached over 94%.

Reporters noted that this implementation in power trading and operations scenarios represents further expansion of Ant Digital's AI solutions into the energy industry. As a B2B technology company, Ant Digital has been focusing on cutting-edge AI technologies, accelerating the global entry of small and medium-sized enterprises into the AI era. Its LingDT series models, DTMaaS platform, DTClaw intelligent agent, and AI security products ZOLOZ have achieved large-scale applications in multiple industries. In the financial sector, Ant Digital has cooperated with financial institutions to launch over 100 scenario applications. In the energy sector, Ant Digital has already implemented scenarios such as energy asset management, carbon certificate traceability, and power trading.