Arbe Robotics

Arbe Robotics

Arbe Robotics was established in Tel Aviv in 2015 by Kobi Marenko, alongside CTO Noam Arkind and COO Dr. Ronen Sadeh, with a thesis that the missing piece in autonomous driving was not software but a fundamentally better sensor. While most of the industry chased lidar or struggled with the limitations of legacy 24 GHz radar, Arbe bet on building the first ultra-high-resolution 4D imaging radar chipset from the ground up. The company went public via a SPAC merger in October 2021, listing on the Nasdaq under the ticker ARBE. The firm's proprietary chipset, called Phoenix, generates roughly 100,000 detections per frame — a roughly 100x improvement over traditional radar — and can separate objects that legacy radar sees as a single blob. Its production strategy targets Tier 1 automotive suppliers, including a named collaboration with HiRain Technologies in China and development work with Weifu Group. Arbe's radar processes range, azimuth, elevation, and Doppler velocity simultaneously, enabling it to detect and classify objects at distances beyond 300 meters. Its engineering footprint spans Israel, with business development and sales offices established in Beijing and Munich to sit closer to both Asian and European automakers. The company deployed roughly $49 million in operating expenditures in 2023, as it pivoted from chip design into a production-readiness phase targeting model-year 2024–2025 vehicle programs. Arbe has publicly winnowed its near-term addressable market to the Chinese automotive supply chain, naming HiRain as a key partner progressing toward a pre-production radar system. In March 2024, the company announced a $35 million registered direct offering to extend its run rate, pending finalization of a Tier 1 production contract that it has characterized as imminent (per the firm's March 2024 SEC filing). Arbe also maintains an innovation lab concept called Radar-to-Data, aiming to generate AI-training datasets from its own sensor output. Arbe's structural wager is not merely a better part — it is a bet on sensor-fusion architecture. Where Tesla's pure-vision approach dispenses with radar entirely and lidar makers argue for a new class of sensor, Arbe claims its radar can replace lidar in core perception tasks while preserving the all-weather reliability and cost profile of radar. If that claim survives mass-production validation, the company's chipset becomes a hard requirement across the automotive supply chain, not an option. This all-in bet on a single product category — with no diversified product line to cushion a delay — defines its risk profile for any institutional allocator evaluating the autonomous-mobility exposure.