30 years on the road and counting...

Bringing Autonomous Driving
to reality

Fully AI-based autonomous driving software stack efficiently running on Ambarella CV3-AD processors. From silicon to software, every detail matters.

Technology at a Glance

Vertically integrated AD stack from sensing to driving — software, silicon, and sensing designed together to improve overall efficiency
Fully owned toolchain, from data acquisition, to labelling, to testing — shortening the loop boosts speed and accuracy

VisLab milestones from 1998 to today

25+ Years on Public Roads

VisLab shaped the field of autonomous driving — from ARGO (1998) to production-ready deployment today.

Neural network-based perception

Fully AI-Based

Pure AI approach without reliance on HD maps. Towards real-time E2E with VLM for slow stack on the same chip.

LiDAR perception and vertical integration for autonomous driving

Vertical Integration

Software, hardware, and sensing technology designed together within Ambarella. Efficiency improves when software and hardware are matched.

Ambarella CV3-AD domain controller

CV3-AD Platform

Full stack running on a single chip: scalable 5nm SoC for L2 to L4: unified SDK, high power efficiency, modern GenAI support, and patented CVflow® IP.

Full technology overview

Pioneering autonomous driving since 1998

It all started when Autonomous Driving was unexplored terrain; then several decades of development, tests, and real challenges — VisLab left its footprint in the history of this field, reaching iconic milestones.

Consider the companies pursuing autonomous driving today. How many of them existed in 1998, during our ARGO project’s 2,000-kilometer autonomous journey?

ARGO vehicle, first autonomous drive on public roads, Mille Miglia 1998
1998

ARGO and MilleMiglia

TerraMax at the DARPA Grand Challenge
2005

TerraMax — DARPA Grand Challenge

VIAC vehicles during the Parma to Shanghai intercontinental autonomous drive
2010

Porter and VIAC

Our Roots →See all Milestones →

Cars or Robotaxis?

Although robotaxis are already operating in several parts of the world, most current solutions achieve autonomy by relying on shortcuts such as HD maps, expensive high-performance sensors, and massive onboard computing power with significant energy consumption. Our goal is different: to deliver scalable autonomy through an efficient, AI-based approach where software, hardware, and sensors are designed together to maximize performance while minimizing cost, power, and operational complexity.

We are working to deliver safety through autonomy to all our future cars (not just robotaxis!)

Technical direction

Our technical direction combines decades of leadership in computer vision and silicon architecture with a shared focus on bringing safe, efficient autonomous driving from research to reality.

Alberto Broggi

Alberto Broggi

General Manager, VisLab

General Manager at VisLab and formerly full professor of computer engineering at the University of Parma. Widely considered a leading pioneer in machine vision for automotive applications, he has been active in autonomous driving R&D for over twenty-five years. IEEE and IAPR Fellow; two European Research Council (ERC) recognitions; 2017 IEEE Medal for Environmental and Safety Technologies for leadership in vehicular perception and worldwide milestones in safe intelligent vehicles. In 2018, Popular Mechanics named his self-driving car patent one of the "15 Patents That Changed the World."

See Alberto accepting the medal →

Les Kohn

Les Kohn

CTA, Ambarella

Well regarded in Silicon Valley as one of the leading chip architects. Combines experience as a silicon architect with a talent for innovative solutions to complex problems. Prior to Ambarella, CTO and co-founder at Afara (acquired by Sun); later Sun Fellow. Was chief architect at C-Cube (three generations of digital video CODECs, including first single-chip MPEG-2 and HD/multi-stream CODECs). Distinguished engineer at Sun (chief architect of UltraSPARC I and II). Chief architect of the Intel i860; co-architected the National 32000 microprocessor family.

See Les' interview in the Computer History Museum →

Career

Join us. We are looking for innovators in machine learning, simulation, autonomy, and control. We offer a dynamic, informal, and international environment with daily connections to our colleagues in California, Taiwan, and China.

View open positions

Get in touch

Contacts

Address: Parco Area delle Scienze 49, 43124 Parma, Italy

Tel.: +39.0521.16023

Email: info@vislab.it

Legal data

VisLab S.r.l.
Registered address: Parco Area delle Scienze 49, 43124 Parma
Tel: +39.0521.16023 · PEC: vislab@mypec.eu
Fiscal Code, VAT: 02516520349
Company Register: PR-244856
A sole shareholder company under the management and coordination of Ambarella Inc.