Ant Digital Technologies upgrades EnergyTS to 3.0, pairs it with Agentar Energy agent platform
The AMW Read
Incremental vertical product update: Ant packages energy-ops agents with a specialized time-series model upgrade; sub-segment signal for enterprise energy AI, not a substrate-wide shift.
Ant Digital Technologies upgrades EnergyTS to 3.0, pairs it with Agentar Energy agent platform
At the Bund Summit on September 10, Ant Digital Technologies launched Agentar Energy Edition, an integrated platform for building, deploying, coordinating, and governing AI agents for energy and carbon enterprises. The same release upgraded EnergyTS, the firm’s energy time-series model, to version 3.0. Agentar packages industry workflows as reusable skill packs and digital expert agents that can orchestrate sub-agents with vertical SaaS systems. Cited use cases include power-trading loops spanning forecast, sensing, analysis, decision, bidding, risk control, and review, plus new-energy mobility workflows across charging-site selection, operations, battery services, and predictive maintenance. The company says it already fields digital workers across more than 20 scenarios spanning generation, grid, load, storage, and carbon, and lists customers including GCL Technology and Linyang Energy.
EnergyTS 3.0 is positioned as a core engine for those agents, with stronger multi-factor learning, historical-condition retrieval, probabilistic forecasting, and long-horizon adaptation. Training data is said to exceed 100 million power-trading records, with prediction-interval coverage above 95 percent. The model covers photovoltaic, wind, load, and electricity-price tasks, and weather-element forecast error is claimed to lead peers on multiple public benchmarks, with a maximum horizon of 45 days. In a market where enterprise buyers are moving from agent connectivity to production reliability, a domain time-series model wired into governed agent workflows is a clearer commercial path than generic chat assistants for trading desks and grid operators.
For builders and investors, the signal is that Chinese enterprise AI vendors are packaging vertical agents around specialized forecasting models rather than selling either layer alone. Watch whether Ant Digital can turn the claimed forecast coverage and customer footprint into measurable trading or dispatch gains that competitors without energy-native models cannot match.
#AntDigitalTechnologies #EnergyTS #Agentar #EnergyAI #TimeSeriesAI #EnterpriseAgents