Imagry CEO Eran Ofir believes the autonomous vehicle industry has misread where demand actually lies.
In an interview with Zag Daily, Ofir argued that robotaxis, which have attracted the bulk of AV investment and attention, do little to address the transport challenges governments and public authorities are trying to tackle.
“Robotaxis serve one or two passengers at a time and are empty 60% of the time,” he said. “They add to congestion. They don’t address the driver shortage. They don’t solve any of the problems that governments are actually trying to solve.”
Imagry, founded in 2018, develops autonomous driving software for buses, roboshuttles and passenger vehicles, with buses accounting for around 80% of the business.
The company has active deployments in Israel and Latvia, and test fleets running in Germany, Japan and the United States.
The shortage driving demand
Europe currently has around 105,000 unfilled bus driver positions, a figure the International Road Transport Union projects will reach 275,000 by 2028.
In Japan, the shortage has prompted a direct government response. The Japan Bus Association projects a shortfall of 36,000 drivers by 2030 as the existing workforce ages out. It is the market where Imagry says it has found regulators most willing to move quickly and where it is working with Toyota Tsusho on a pilot deployment.
“In Europe and Japan, the conversation is about a critical shortage of bus drivers,” Ofir said. “In Japan specifically, 40% of bus drivers are expected to retire within the next ten years.”
European authorities have typically required 12-month pilots with a safety driver before approving driverless operations. Ofir said he expects full Level 4 driverless bus legislation across most of Europe within five years.
In the US, where many states allow operators to self-certify as Level 4 compliant, the use case for autonomy is different. Ofir cited rural and suburban access, corporate shuttle networks and airport services as the primary use cases.
Even there, he said, labour shortages remain a major factor. Fixed-route shuttle services can be difficult to staff consistently, making automation attractive for operators facing the same workforce pressures seen elsewhere.
The camera-only argument
Imagry’s system uses eight cameras and no LiDAR and operates without HD maps or a cloud connection, a position that remains contentious in the industry.
Tesla’s camera-only approach has drawn sustained criticism and regulatory attention and sceptics argue that a full sensor suite is necessary for safe autonomous operation.
Leading autonomous vehicle expert and former NHTSA advisor Missy Cummings previously told Zag Daily that “no safety-critical system should ever be reliant on a single sensor type,” drawing a comparison with the Boeing 737 Max crashes, where faulty reliance on a single sensor led to catastrophic outcomes.
Ofir distinguishes his company’s approach from Tesla’s on the basis of training methodology. Imagry has used supervised learning since 2018 with human reviewers assessing and correcting the system’s decisions throughout training. Tesla, he said, currently uses unsupervised learning without that human oversight.
“Over many years and hundreds of millions of images, you reach a level of driving safety that regulators are increasingly accepting,” he said.
Addressing the counter-argument that imitating human drivers encodes human error, he said human reviewers assess driving decisions during training and remove behaviours that do not meet the company’s safety standards.
He also argued that cameras carry capabilities that LiDAR does not. “Cameras understand what they see. They can read traffic light colours, road signs, lane markings. LiDAR cannot do that.”
Because the system operates without any cloud connection, it functions identically in areas with poor connectivity, which Ofir said matters for the rural and suburban markets Imagry is targeting.
That focus on practical deployment, he believes, is what separates autonomous buses from much of the wider AV industry.
