Cerence and SiMa.ai have agreed on a strategic partnership to make intelligent voice interaction available directly in the vehicle. The core of the collaboration is the integration of Cerence CaLLM Edge, a Small Language Model (SLM) specially developed for automotive applications, into SiMa.ai’s MLSoC platform Modalix. This allows conversational AI functions to be operated directly at the edge with low latency and low energy consumption – regardless of cloud connection or network quality.
The joint solution will be presented at the IAA Mobility 2025 in Munich. There, Cerence will be demonstrating its agent-based assistance platform Cerence xUI with CaLLM Edge on SiMa.ai hardware.
CaLLM Edge, Cerence AI’s compact voice model for the automotive sector, will run on SiMa.ai’s Modalix MLSoC platform, enabling intelligent and energy-efficient voice interactions directly in the vehicle. (© SiMa.ai)
Cerence’s hybrid AI architecture allows vehicle users to switch seamlessly between edge and cloud processing. This ensures stable voice control with high response speed, even if there is no connectivity. At the same time, data protection is maintained as sensitive voice data is processed in the vehicle. „With the combination of CaLLM Edge and Modalix, we are achieving a new quality of voice control in the car – powerful, energy-efficient and ready for series production,“ says Harald Kröger, President Automotive at SiMa.ai. „Together, we enable OEMs to provide multimodal interactions and real-time experiences in their vehicles – without the dependency on pure cloud solutions.“ (oe)
