General Motors

Solution

General Motors describes an in-house AI stack used by its motorsports operation from the Charlotte Technical Center in North Carolina. The implementation ingests three live race-day inputs: onboard driver radio, car-to-pit telemetry where series rules allow it, and trackside photographs captured as cars pass. GM says off-the-shelf transcription software could not handle race noise and specialized terminology, so it built proprietary speech AI that transcribes radio instantly and lets analysts follow multiple drivers and topics at once. Separate AI models process telemetry to monitor fuel consumption and run tire-wear simulations that support pit-stop timing. Computer-vision workflows tag incoming images in seconds and alert engineers when crash damage or mechanical issues are detected, or confirm when a car appears undamaged. The outputs are served directly to engineers and strategists in the command room for second-by-second decisions; the source does not identify a named storage platform, only immediate analyst-facing workflows.

Data flow

During a race, driver radio audio, permitted car telemetry, and trackside photos flow to GM's Charlotte Technical Center. Proprietary AI transcribes speech, analyzes fuel and tire signals, and tags images for visible damage or mechanical issues. The processed outputs are surfaced to engineers and strategists as near-real-time text, alerts, and decision support used to adjust cars and time pit stops.