Emerson and SiMa.ai announced a collaboration on May 26 to embed SiMa.ai's MLSoC (Machine Learning System on Chip) inside Emerson's next-generation rugged industrial PCs, bringing AI inference for quality inspection, predictive maintenance, and anomaly detection directly onto the factory floor without requiring cloud connectivity.
The MLSoC processes images, video, audio, text, and sensor data simultaneously in real time while maintaining a low thermal footprint, allowing the IPC to handle AI inference without a separate analytics server. What's new here is that detection and response tasks now execute on the same edge device as core process-control logic, keeping proprietary data fully on-premise. "As operations teams leverage Physical AI at the edge, they move beyond simple monitoring to closed-loop autonomy where they can adjust processes in real time, minimizing product defects early in the production phase, reducing waste and increasing production efficiency," said Krishna Rangasayee, chief executive officer of SiMa.ai.
For inline quality inspection, that closed-loop architecture means the system detects defects and adjusts process parameters in the same on-device inference pass. The platform targets a range of applications across discrete and process industries, including compressed air optimization, packaging, automotive machine efficiency, semiconductor manufacturing, and oil and gas monitoring such as wellhead management and flare monitoring.
The IPCs are rated for operating temperatures from -40 degrees F to 140 degrees F and are built to withstand high vibration and shock loads. They integrate with Emerson's existing PLC stack and IIoT-ready SCADA/HMI platform, which the company positions as a unified industrial edge intelligence layer.
The partnership comes as IOT Analytics projects the global industrial AI market to grow from $43.6 billion in 2024 at a 23 percent compound annual rate through 2030, reaching $153.9 billion. Both companies also highlight air-gapped installations, covering nuclear, power, and water facilities, as a target segment where on-premise AI inference removes the connectivity requirement that has blocked adoption in those environments.
