Vellex Computing

Vellex Computing, a Stanford spinout backed by the U.S. National Science Foundation and Department of Energy, is an analog semiconductor company making AI training radically more energy-efficient. Today, training AI models can require megawatts of power and massive cloud compute costs, putting it out of reach for most edge applications. We solve the optimization problem at the heart of AI training by using analog circuits and our analog compute architecture rather than brute-force digital computation. This proprietary architecture has demonstrated 17,000x speedup over conventional techniques while operating at milliwatt power levels. Starting with edge AI in energy, industrial IoT, robotics, and remote sensing, Vellex enables devices to learn continuously and securely in real-time, without any cloud reliance. With active commercial pilots and a growing patent portfolio, Vellex is unlocking the next generation of adaptive, cloud-free industrial intelligence.

Key Technology Highlights

Physics-Based Analog Compute: Maps AI optimization problems directly onto analog circuits, which settle to a solution through physical dynamics rather than iterating through discrete digital steps. The result is an AI training engine that operates within the power budget of embedded and battery-powered systems.
Zero Data Movement: Computes in-situ where data resides, eliminating the repeated transfers between memory and compute units that dominate power consumption in conventional digital training.
Energy Efficiency and Speed: Delivers up to 17,000x processing speedup over conventional approaches while operating at milliwatt power levels.
On-Device Continuous Training: Enables battery-operated devices, including autonomous drones, satellites, IoT sensors, and humanoid robots, to train continuously on live data without cloud connectivity. Models adapt in real time, eliminating the cost and latency of cloud retraining pipelines.
On-Device Data Security: Field data stays on the device. No transmission to external servers means no exposure to network-based risks and no dependency on internet connectivity.