Beamr Imaging Ltd. (Nasdaq: BMR) launched a lossless compression feature for its ML-Safe video data compression product targeting autonomous vehicle test fleets, according to a press release dated Sept. 16, 2026.

The company said the new capability compresses 12-bit Bayer raw camera footage — the sensor output before it enters the AV processing pipeline — without any data loss. In Beamr's testing on 12-bit Bayer recordings from 8-megapixel cameras, compressed files were 47% smaller than raw files.

Beamr said the compression runs on the video encoder integrated in a GPU already present in the data logger, operating separately from the GPU compute cores so it does not draw on resources used by AI models. No new hardware is required, the company said.

The product addresses a storage constraint common to AV test fleets, where a single vehicle can run as many as 16 cameras and record tens of terabytes of footage in one shift. Previously, Beamr said, AV teams often could not compress 12-bit Bayer footage at the point of capture because encoders in test vehicles did not support that format.

"Through our joint work with AV and ADAS companies we identified their need for lossless compression at the edge, in the data-collection vehicle, supporting multiple camera streams in real time. So we built it," said Sharon Carmel, Beamr CEO.

The lossless option joins an existing lossy compression mode within the ML-Safe product, which the company said is validated against a customer's own ML model accuracy metrics. Beamr Chief Product Officer Dani Megrelishvili said the lossless addition completes the ML-Safe offering for teams that require bit-exact preservation of camera data.

Beamr plans to demonstrate the solution at AutoSens Europe in Barcelona, Sept. 22–24, at stand 131.