Technology

World’s first native color LiDAR gives machines human-like vision

World’s first native color LiDAR gives machines human-like vision
A city block as Rev8 sees it: pedestrians, traffic signals, and building facades rendered in a fully colored 3D point cloud
A city block as Rev8 sees it: pedestrians, traffic signals, and building facades rendered in a fully colored 3D point cloud
View 6 Images
A city block as Rev8 sees it: pedestrians, traffic signals, and building facades rendered in a fully colored 3D point cloud
1/6
A city block as Rev8 sees it: pedestrians, traffic signals, and building facades rendered in a fully colored 3D point cloud
The OS1 Max has a 500 m (1,640 ft) range, and color baked into every data point
2/6
The OS1 Max has a 500 m (1,640 ft) range, and color baked into every data point
From near-darkness to direct sunlight, Rev8 captures color and geometry in a single pass
3/6
From near-darkness to direct sunlight, Rev8 captures color and geometry in a single pass
Ouster's L4 chip, co-developed with Fujifilm, detects up to 20 trillion photons per second – the engine behind Rev8's native color
4/6
Ouster's L4 chip, co-developed with Fujifilm, detects up to 20 trillion photons per second – the engine behind Rev8's native color
Rev8 sensors during testing
5/6
Rev8 sensors during testing
The new LiDAR can make robots see the world like humans
6/6
The new LiDAR can make robots see the world like humans
View gallery - 6 images

For years, machines have navigated the world color-blind. LiDAR sensors – the laser-based eyes of self-driving cars, industrial robots, and inspection drones – build precise 3D maps of their surroundings, but everything is built of monochrome geometric shapes. Ouster's new Rev8 sensor family aims to change that, not by bolting a camera onto a LiDAR unit, but by fusing color directly into every point of data the sensor captures.

Autonomous perception systems generally fall into two camps: camera-only arrays – like the vision system Tesla uses for its underwhelming Full Self Driving tech – or a two-step sensor fusion approach with LiDAR for precise geometry, a camera for color, and a software algorithm to combine them. That stitching process introduces calibration errors, latency, and spatial mismatches – a problem that becomes critical when a robot or vehicle is moving fast through a crowded street.

The new LiDAR can make robots see the world like humans
The new LiDAR can make robots see the world like humans

Rev8 eliminates that architecture entirely. Each point in the 3D map the sensor generates already carries color information at the moment of capture, with no additional software processing required. This, Ouster says, makes it the first LiDAR with native color – though it isn't competing alone for long.

Just a few weeks ago, the Hesai Group, the global leader in LIDAR technology, presented a full-color platform called 6D ETX. The Chinese sensor takes a different approach: rather than simply adding color to the point cloud, it captures six full dimensions of data – X, Y, and Z coordinates alongside reflectivity, velocity, and color – making it less a color LIDAR and more a multi-dimensional perception engine.

Ouster's technology is built around its new L4 chip, which embeds Fujifilm's color science – the same expertise behind the company's imaging technology – to deliver hardware-level color processing. It integrates 42.9 GMACs of processing capacity, detects up to 20 trillion photons per second, and operates at 40 kHz with picosecond-level precision. Those numbers are dense, but they mean a single sensor can now read a traffic sign, detect whether the car ahead is braking by the color of its brake lights or produce topographic maps with real-world color data – all without additional hardware or calibration.

The OS1 Max has a 500 m (1,640 ft) range, and color baked into every data point
The OS1 Max has a 500 m (1,640 ft) range, and color baked into every data point

The flagship model of the Rev8 family is the OS1 Max, a 256-channel sensor with a detection range of up to 200 m (656 ft) at 10% reflectivity – meaning it can spot surfaces that absorb most of the light hitting them – and up to 500 m (1,640 ft) under optimal conditions. Its field of view spans 45 degrees vertically and 360 degrees horizontally, and Ouster claims it doubles both the range and resolution of its previous generation, the Rev7.

The sensor handles an impressive spread of lighting conditions, from near-total darkness at 1 lux up to 2,000,000 lux, roughly equivalent to intense direct sunlight. Color depth reaches 48 bits with 116 dB of dynamic range – technical shorthand for the sensor's ability to capture fine detail across extremely bright and extremely dark areas simultaneously.

"Rev8 is the most advanced family of LiDAR sensors ever released and sets a new standard in sensing," said Ouster CEO Angus Pacala. "With the L4 Ouster Silicon, we are delivering on the promise of our digital architecture to deliver exponential improvements in performance, doubling our core specs and simultaneously introducing the world’s first native color LiDAR to give machines 3D human-like sight for the next era of Physical AI."

From near-darkness to direct sunlight, Rev8 captures color and geometry in a single pass
From near-darkness to direct sunlight, Rev8 captures color and geometry in a single pass

Early adopters include Google, Volvo Autonomous Solutions, Skydio, PlusAI, and Seegrid, among roughly two dozen companies across robotics, automotive, and smart infrastructure.

The launch also reflects a broader strategic shift. In February 2026, Ouster acquired StereoLabs – a computer vision specialist – for US$38 million in cash plus 1.8 million shares, signaling a move away from selling standalone sensors toward offering a full perception platform. When color and geometry are fused at the source, training AI models becomes significantly easier.

"Rev8 is the foundational technology that will allow customers to move from prototype to commercial production at scale, providing the reliability and affordability required to enable real-world autonomy across industries," added Pacala. Rev8 sensors are available to order (though we've no word on pricing for business customers), with shipments expected to begin this quarter.

REV8 OS1 MAX with Native Color in Chinatown SF

Source: Ouster

View gallery - 6 images
11 comments
11 comments
paul314
reflection and velocity are important, because they give you a clue about what the object is and whether it's moving in a way that will hit you. Nicely done.
Chase
Impressive. I still don't think it will be enough to make L5 a real thing, but it's another step on the long path in that direction.
guzmanchinky
Very cool. In less than a decade I see all new vehicles with a self driving option…
RandD
I'm not surprised that a LiDAR company would use the term "underwhelming Full Self Driving tech".
Davina Cotter
Nobody who has actually DRIVEN a Tesla with FSD(s) would ever call it "underwhelming". I think the adjective you're looking for is "mind blowing".
Edward Vix
RandD, I noticed the snide remark too, but it seems it was from Omar Kardoudi who produced this article. Not sure why he would write that, the YT videos of Tesla FSD, and there are many, demonstrate spectacular and near-miraculous performance.
RandD
Edward, You have that right. Tesla has shown that cameras pick up the information needed and it's really the AI processing that's the secret. My wife and I have been using FSD for 3 years and it's been life changing (for us 80 year olds, especially at night). What's also great is that through OTA SW updates it just keeps getting better.
I was wondering how the LiDAR sees in color. Is it sending out RGB laser beams to see them bouncing back? I reread the article and realized it's using a camera for color.
MrsWeb
In 1999 I "invented" a concept, with 4 sensors for each pixels in a camera. RGB and infrared. When you pressed the "click button" to take a picture, a very short infrared blitz would fire, and the infrared sensors would register the time of the reflection to return. In that way, you would have a crisp colorphoto, with distance of every pixel in the photo, and in thus a fuld colored 3D image, with only one lens. With that system and some computerpower, you would be able to record videos, where you could record movement, distance and direction of every object in the recording - in full color. With much better quality and clearity.
Just my 2 cents...
Cerethus
Hey Musk shills -- if Tesla's self-driving is so "mind-blowing" then why has it ben stuck in development hell for over a decade?
Tommo
As a heavy user of Teslas self driving software I can describe it as the complete opposite of 'underwhelming'. I soon as I read that remark I dismissed any point that article was trying to raise. Do I see China as being able to beat Tesla in the near term? Possibly, I personally think once their flagship cars hit the UK/EU markets then it'll be game over for all other EV manufacturers - However, that article above does nothing at all to reinforce that view with me, all due to that ridiculous comment..
Load More