In April 2026, the LiDAR industry welcomed a long-awaited technology explosion. Two leading companies released heavyweight new products back to back, and the rivalry was highly charged.
On April 17, Hesai threw out “Picasso” — the world’s first 6D full-color LiDAR chip; on April 21, RoboSense followed closely with its “Genesis” architecture and dual flagship chips. Both companies have placed their bets on the chip track, and the signal behind this is interesting: the LiDAR industry is undergoing a fundamental shift from “mechanics is king” to “chip is king”.
Honestly speaking, this technology race came faster than I expected.
RoboSense “Genesis” Architecture: Chips Define Next-Generation Perception
The “Genesis” architecture released by RoboSense this time is essentially a SPAD-SoC chip-level solution platform. In plain language, it moves the most core performance iteration of LiDAR down from traditional whole-unit tuning to the underlying chip level.
What are the benefits? Just as smartphone chips iterate every year, LiDAR performance improvement and cost optimization can in the future iterate as fast as the semiconductor industry. It is a rather interesting shift in thinking.
Dual-Chip Layout
RoboSense threw out two chips this time, positioned differently.
Phoenix chip is positioned as the core of high-end main LiDAR. It is the world’s first monolithically integrated native 2,160-channel automotive-grade SPAD-SoC. With over 4 million pixels of resolution and a detection distance of up to 600 meters, it has already obtained automotive qualification and plans to enter volume production and vehicle installation within 2026. These specs are indeed impressive.
Peacock chip targets blind-spot filling and multi-scenario applications. It is the industry’s first mass-producible 640×480-resolution fully solid-state large-array SPAD-SoC, with an ultra-wide 180°×135° field of view and millimeter-level precision, planned for volume production in Q3 2026.
Volume-Production Confidence
RoboSense’s volume-production experience indeed provides a solid foundation for landing the new products. The E-platform fully solid-state LiDAR has accumulated over 300,000 units delivered. In the L4 autonomous driving field, the combination of “thousand-channel-class main radar EM4 + fully solid-state blind-spot-filler E1” has become the preferred solution of over 90% of leading customers such as Baidu Apollo Go, DiDi Autonomous Driving and Pony.ai. In the ADAS field, it has also won blind-spot-filling design wins from several heavyweight automakers.
Hesai “Picasso” Chip: Full-Color Perception Defines a New Standard
The “Picasso” SPAD-SoC released by Hesai this time is claimed to be the world’s first 6D full-color LiDAR chip. The chip achieves chip-level, pixel-level fusion of color perception and TOF ranging — outputting X/Y/Z 3D spatial coordinates and R/G/B color information synchronously on the same chip, directly generating native color point clouds.
This “full color” is not a gimmick. It can directly distinguish the red, yellow and green of traffic lights and is immune to environmental interference such as strong light, shadows, rain and fog. It can also clearly identify roadside pedestrians and reflective signs at night — a genuine improvement in the safety performance of autonomous driving.
Technical Specs
Photon detection efficiency exceeds 40%, reaching the international top level. It supports up to 4,320 laser channels, 22.5 times the current mainstream 192-channel product. Point cloud density reaches 4K ultra-HD level. The ETX-series LiDAR offers multiple configurations of 1,080, 2,160 and 4,320 channels. Hesai has also self-developed photon isolation 2.0 technology and a waveform decoding engine capable of filtering 99.9% of noise points.
For long-range identification, the maximum ranging distance reaches 600 meters, with stable ranging of 400 meters at 10% reflectivity. This metric is quite competitive in the industry.
Industrialization Capability
Hesai has already demonstrated strong industrialization capability. Cumulative delivery of self-developed chips has reached 230 million units, expected to exceed 300 million by the end of 2026, ranking first globally in the industry. In February 2026, its automotive LiDAR market share rose to 51%, remaining first in the industry for 13 consecutive months. It has won volume-production design wins from 40 automotive brands covering over 160 models. This scorecard is indeed dazzling.
Industry Landscape: From “Channel-Count Race” to “Full-Dimension Perception”
Diverging Technology Paths
Although the two giants have both bet on the chip track, their technology paths still differ clearly:
Competitive Situation
It can be seen that RoboSense emphasizes “fully solid-state” and “volume-production landing”, while Hesai pushes harder on “channel count” and “full-color perception”. Which path will ultimately fare better is genuinely hard to say now. But one thing is certain — 2026 will be a pivotal year.
Industry Impact and Outlook
Trend of Chip-Based Competition
With the emergence of chip-level platforms such as “Genesis” and “Picasso”, the core competition of the LiDAR industry has formally shifted to the chip level. This is a big deal for the whole industry — by following the iteration laws of the semiconductor industry, LiDAR is becoming stronger and more reliable while its cost is also expected to fall systematically.
Cost-Drop Forecast
Industry forecasts suggest that within the next 1 to 2 years, mainstream vehicles in the CNY 100,000 class may carry entry-level full-color LiDAR. If this forecast comes true, it will truly achieve “technology democratization”, letting more consumers enjoy the safety and convenience brought by advanced intelligent driving.
Expansion of Application Fields
6D full-color perception serves not only autonomous driving but can also extend to robotics, drones, industrial inspection and many other fields. LiDAR is becoming the core perception infrastructure of the physical AI era, and this positioning is getting clearer.
Final Words
Having watched the new product releases of these two manufacturers, my takeaway is: the era of “image-grade perception” for LiDAR has truly arrived.
From the early “mechanically rotating scanning”, to the later “solid-state solutions”, and now to “chip-level full-color perception”, the pace of technology iteration in this industry is far faster than I imagined.
Future competition will be a comprehensive contest of underlying chip technology, architectural innovation and ecosystem integration capability. Whoever can push forward on all these dimensions at once will gain the first-mover advantage in this competition.
As a practitioner, I am quite looking forward to 2026 — by then we will see the real market performance of these “futures” products after they land.
References:
- RoboSense 2026 Technology Open Day release information
- Hesai 2026 Technology Open Day release information
- Official releases of RoboSense
- 36Kr and other industry media reports



Leave a Reply