Liquid AI puts on-device generative AI into Mercedes-Benz cars
The startup says its flagship models are entering production vehicles in North America and power assistance without relying on the cloud.
Liquid AI says its flagship Liquid Foundation Models are rolling out in Mercedes-Benz vehicles in North America, where they run directly on the car’s hardware to power in-vehicle assistance.
The deployment moves generative AI from a cloud service into a production consumer device. Liquid AI co-founder and CEO Ramin Hasani said the companies are now in production and are working toward deployment at scale. The Wall Street Journal separately described the partnership as running AI models inside the car rather than in the cloud.
Neither company has publicly detailed the exact vehicle models, hardware configuration or rollout volume. The available claims therefore establish the production deployment, but not how many drivers have access or which assistant features use the models.
On-device inference can make an automotive assistant less dependent on connectivity, reduce round trips to remote servers and give the manufacturer tighter control over latency and data handling. It also demonstrates a practical market for smaller, hardware-efficient generative models: systems that must fit fixed compute and power budgets instead of assuming a data-center connection.
