Verkko Robotics launch VOLTAIC, a cost and energy saving Spiking Neural Inference Model (Sponsored)
# Verkko Robotics Launches Energy-Efficient AI Model for Edge Computing
London-based Verkko Robotics has released VOLTAIC, an artificial intelligence system designed to operate with significantly lower energy and computational demands than current large language models. The system uses a spiking neural network architecture—a technology that mimics how biological brains process information—to perform inference tasks while consuming less power and computing resources. According to the announcement, the model can continue learning from new data after deployment and retain previously learned information without full retraining.
For energy sector applications, this development is relevant to distributed systems that require AI decision-making at the edge—such as smart inverters, battery management systems, EV charging stations, or local grid controllers. Lower computational demands could enable more devices to run intelligent algorithms locally without relying on cloud connectivity or centralized processing, reducing latency and dependency on external servers. This matters for real-time grid coordination, where millisecond-level decisions about power distribution, storage dispatch, or load balancing currently depend on processing power and network infrastructure.
The practical question for operators remains whether VOLTAIC's real-world performance in energy applications matches its technical specifications. Spiking neural networks have theoretical efficiency advantages, but deployment success depends on validation in specific use cases—whether managing solar production variability, coordinating multiple heat pumps, or automating EV charging schedules. Field testing in operational energy systems will determine whether the efficiency gains translate to meaningful cost or carbon reductions at scale.