Vayu Robotics is taking a direct shot at one of the most expensive pieces of the autonomous vehicle and robotics stack: LiDAR. The startup, co-founded in 2022 by Anand Gopalan, is building delivery robots that rely on foundation models and passive sensing instead of the laser-based technology that has become common across autonomous systems.
The move stands out because Gopalan is not an outsider to LiDAR. He was the former CTO and CEO of Velodyne, described in the source as a LiDAR leader, and took Velodyne public via SPAC two years before co-founding Vayu Robotics.
A LiDAR veteran makes the case against LiDAR
LiDAR has long been treated as a key technology for autonomous vehicles and robotics. Its value is clear: it helps machines understand the world around them. But the source article highlights a major drawback that has followed the technology into commercial robotics: high cost.
Vayu Robotics is leaning into that weakness. Rather than treating LiDAR as a required component, the company is positioning LiDAR-free navigation as one of its main advantages. Its larger goal is to make delivery robotics cheaper and more scalable.
That is why the company’s technical shift matters. The traditional mobile robotics path has often meant adding multiple sensors to a robot and building software modules that each handle narrow tasks. In Gopalan’s account, that approach can create expensive hardware and software that struggles when the world becomes uncertain or unfamiliar.
Vayu Robotics is trying a different architecture. According to Gopalan, the company uses a transformer based mobility foundation model along with a new type of powerful passive sensor. The stated aim is to remove the need for lidar, especially in low speed applications.
Foundation models move into physical delivery
Foundation models are the machine learning systems behind the recent generative AI boom. Vayu Robotics is applying that category of technology to mobility, not just content generation or software tasks.
For delivery robots, that change is significant because the operating environment can be messy. Streets, curbs, traffic patterns, obstacles and customer drop-off scenarios do not always fit neatly into prewritten software modules. The company’s argument is that a foundation model can help address that uncertainty more flexibly than a stack built around many separate, task-specific modules.
The source does not provide the full technical details of Vayu’s model or passive sensor. What it does make clear is the company’s intended tradeoff: reduce reliance on costly sensor hardware and use a machine learning approach that can support broader deployment.
In practical terms, that puts Vayu Robotics in the middle of two trends. One is the push to automate last-mile goods delivery. The other is the expansion of foundation models beyond chatbots and image tools into systems that move through the physical world.
The first target is goods delivery
Delivery robots are Vayu Robotics’ first step. The source describes the market as large and growing, while also noting that the category has faced plenty of pitfalls. That combination explains why cost, deployment scale and reliability matter so much.
Vayu Robotics says it has attracted investors, including Khosla Ventures, and has raised $12.7 million to date. Funding alone does not prove a robotics company can scale, but the source points to a more concrete commercial milestone: a deployment agreement.
The company says it has signed a “substantial commercial agreement with a large e-commerce player to deploy 2500 robots to enable ultra-fast goods delivery, with similar commercial customers in the pipeline.” The customer has not been disclosed, and the source says Vayu has not shared the specifics of that deal.
Still, the planned deployment size is central to the company’s story. A 2500-robot agreement suggests the company is aiming beyond a small pilot. For a delivery robotics startup, the difference between testing and broader commercial deployment is one of the most important gaps to cross.
Why the on-road design matters
Vayu Robotics is also separating itself from the sidewalk delivery robots that have become familiar in the category. Instead of focusing on slow-moving sidewalk machines, the company is pursuing an on-road approach.
According to the source, Vayu says its system can carry a 100-pound payload at speeds of up to 20 miles an hour. Those two details help explain the company’s commercial focus. A robot that can move heavier goods at road speeds is being positioned for faster, more scalable goods delivery than sidewalk-only machines.
That choice also raises the importance of the navigation system. If the robot is operating on-road, the company’s LiDAR-free stack has to support the kind of mobility Vayu is promising. The source frames the company’s technology package as the answer to problems that have limited delivery robots over the past decade.
Gopalan’s claim is direct: the technologies developed at Vayu have allowed the company to solve longstanding delivery robot problems and create a system that can be deployed at scale for cheap goods transport. The source does not independently verify that claim, but it shows how Vayu is presenting its bet to customers, investors and the robotics market.
A test of cost and scale
Vayu Robotics is not simply swapping one sensor for another. It is making a broader argument about how mobile robots should be built. The company believes delivery robots can be cheaper and more scalable if they avoid LiDAR, use passive sensing, and rely on a transformer based mobility foundation model.
That strategy puts pressure on execution. The delivery robotics industry has already shown that demand for faster goods movement does not automatically translate into easy deployment. Hardware costs, software reliability and real-world uncertainty all remain central challenges.
For Vayu, the next stage is about proving that its architecture can support the commercial scale it is promising. A disclosed raise of $12.7 million and a stated 2500-robot agreement give the company a bigger platform than an early experiment. The harder question is whether LiDAR-free delivery robots can turn that platform into repeatable deployment.