6D Color LiDAR: Hesai Picasso vs Ouster REV8

Introduction: The missing link in autonomous perception

Since the advent of autonomous perception systems, a fundamental trade-off existed: LiDAR provides accurate 3D geometry without color, cameras provide color without reliable depth. The standard solution — sensor fusion (calibration, registration, synchronization) — is an integration nightmare: thermal drift, temporal misalignment, computational cost.

In 2026, this trade-off disappears. Hesai and Ouster deliver the first sensors that combine depth and color at the pixel level, on the same chip. The result is a directly colorized 3D point cloud — no calibration, no fusion, no additional latency.

But while the destination is the same, the paths taken by Hesai and Ouster are radically different.

1. Hesai Picasso 6D: Color through ASIC integration

The Picasso SPAD-SoC is a custom integrated circuit (ASIC) that embeds the SPAD sensor AND color processing on the same chip. Each Picasso pixel measures both Time-of-Flight (ToF) for depth and spectral intensity across three channels (R, G, B).

Where a classic SPAD sensor only detects photon arrival (timing → distance), the Picasso analyzes the wavelength of the received photon. Result: each cloud point contains 6 dimensions — X, Y, Z, intensity, and two chromatic components.

ETX series key specs: up to 4,320 channels, 4K full-color resolution, max range 600 m (object detection 15×25 cm at 150 m), range @10% reflectivity ~250 m, production S2 2026 (samples available). Target: automotive, robotics, smart city.

ASIC approach advantages: high spectral resolution (3 channels per pixel without Bayer mosaic), ultra-high sensitivity (116 dB dynamic range), scalability (1,080 to 4,320 channels on the same chip), and industrial scalability leveraging Hesai's 4 million LiDAR/year production.

2. Ouster REV8: Color through native flash

Ouster's REV8 family relies on a different innovation: each SPAD pixel captures both the ToF signal (depth) and the ambient RGB light reflected by the LiDAR's infrared illumination. In a single laser shot, the REV8 records time-of-flight → distance, IR return intensity, and scene color in the visible spectrum.

Color is not painted onto the point cloud afterwards — it is captured simultaneously by the same sensor, on the same chip. Ouster calls this native color LiDAR.

REV8 family key specs: 48-bit color resolution (116 dB dynamic range), up to 256 channels, doubled range vs REV7 (OS0: ~50m, OS1: ~150m, OS2: ~300m), in production since May 2026, BABA certification for US public markets. Target: robotics, autonomous vehicles, smart city, ITS.

Native flash approach advantages: simplicity (no new ASIC process), maturity (shipping since May 2026), competitive pricing (replaces 2-3 cameras + 1 LiDAR, reduces BOM by 40-60%), and integration with the existing Ouster SDK and ROS2.

3. Detailed comparison table

Key comparison criteria: Hesai ETX focuses on maximum resolution (4,320 channels, 4K color, 600 m range) with a dedicated ASIC approach, while Ouster REV8 focuses on immediate availability (production, ~$2,000), BABA certification, and integration simplicity. Hesai targets premium automotive and robotaxis; Ouster covers robotics, industry, and smart cities with more accessible price points.

Both sensors are solid-state, with no moving parts. The fundamental difference lies in the color acquisition method: pixel-level spectral analysis (Hesai) vs native RGB capture in the SPAD detector (Ouster).

4. Two approaches, one paradigm

What strikes about these simultaneous launches is the convergence toward the same thesis: a single perception sensor replaces the fusion of multiple sensors.

Hesai bets on maximum ASIC-level integration — by etching the entire processing chain on a single chip, they achieve unmatched spectral resolution (4,320 channels, 4K). It's the most technically advanced solution, but also the most complex and recent.

Ouster bets on adapting its existing architecture — the REV7 to REV8 transition adds color without fundamentally changing the sensor design. Less spectacular on paper (256 channels max), but available now at a controlled industrial cost.

For integrators, the choice depends on the use case: mobile robotics/AMR → Ouster REV8 (available, simple, sufficient for 90% of tasks); autonomous vehicle/robotaxi → Hesai ETX (superior range and resolution); smart city/ITS → REV8 with BABA certification (US public markets).

5. The competitive landscape

Hesai and Ouster are not alone in this color race. RoboSense offers the Active Camera AC1/AC2, a module merging dToF + stereo RGB + IMU (120°×90° FoV) — not true pixel-level color LiDAR, but a super-sensor combining multiple modalities.

Innoviz Two ULR focuses on raw performance with 1 km range, without native color (external fusion). Blickfeld Qb2 offers a software fusion approach (simultaneous point cloud and video stream). The real duel is between silicon-level fusion (Hesai, Ouster) and module-level fusion (RoboSense).

6. Practical applications

Color LiDAR enables use cases impossible with monochrome LiDAR: AMR navigation (distinguishing a blue sign from a gray wall), industrial inspection (detecting rust on a metal structure), humanoid robotics (recognizing objects by color + shape), ITS traffic management (classifying vehicles by color), precision agriculture (simultaneous NDVI + crop height), and perimeter security (alarm + visual identification).

7. Availability and outlook

Hesai ETX (Picasso): samples, production S2 2026, price TBD (high-end, likely >$3,000). Ouster OS0/OS1/OS2 REV8: in production, immediate shipping, ~$2,000–$6,000 depending on model. RoboSense AC1/AC2: samples, production Q3 2026, ~$1,000–$2,000 estimated.

LiDAR sensor costs have dropped 40% in two years. The arrival of native color could accelerate this trend by reducing the number of sensors per system.

Conclusion: The post-fusion era begins

Color LiDAR marks the end of the perception system integration nightmare. Whether through Hesai's ASIC approach or Ouster's native flash, the message is clear: the single-sensor 3D color perception is available today.

For robotics, automotive, smart cities, and industry, this is a leap comparable to the transition from monochrome to color photography. Integrators adopting this technology in 2026 will have a head start.

Jovaxis recommendation: immediate need → Ouster REV8 (available, native ROS2); max performance → Hesai ETX (4,320 channels, 600 m, wait for production); general robotics → evaluate RoboSense AC for best feature/price ratio.