Bonsai Robotics harnesses AI to create a flexible system that converts 2D images into 3D farm models, enhancing efficiency across various crops.
Ag robotics have long grappled with a significant trade-off: companies must either design machines tailored for specific tasks or crops, which limits versatility, or develop more adaptable systems at a higher cost and complexity. California-based Bonsai Robotics aims to address this issue with a sophisticated vision-based autonomy stack that purportedly transforms 2D camera images into comprehensive 3D models of agricultural environments.
According to cofounder and CEO Tyler Niday, this innovative technology allows Bonsai’s underlying system to operate across various types of specialty crops, such as orchards, vineyards, and berries, without requiring developers to rewrite extensive lines of code for each new application. While Bonsai continues to enhance existing agricultural equipment, the company is also focused on developing its own machinery. Following its acquisition of Farm-ng, Bonsai plans to launch larger, more robust versions of its Amiga robotic platform, designed for tasks like spraying, hauling, and lifting.
Niday emphasizes that the objective extends beyond merely replacing human labor; it's about reducing overall farm operation costs by integrating autonomous systems with cost-efficient machinery.
AI-Powered Vision and Adaptability
In a conversation with AgFunderNews (AFN) during the inaugural Ruggedize ag robotics conference at Reservoir Farms in Salinas, Niday elaborated on how AI is transforming the economics of agricultural robotics. He described Bonsai Robotics as a full-stack robotics company that develops an operating system to manage and control farm operations, alongside an autonomy stack applicable to both existing tractors and new devices.
Bonsai's approach involves significant cost savings not only in the initial purchase of agricultural machinery but also in ongoing operational expenses. For example, one of their machines operates on approximately three gallons of diesel per day compared to a traditional machine that might use around 30 gallons.
Innovative Machine Learning Approaches
Niday discussed the company's transition from processing 2D images to constructing 3D representations and highlighted its reliance on a learned model. Historically, developing autonomous solutions in robotics required extensive coding through “if-then” statements tailored to individual crops or tasks. However, Bonsai Robotics employs a foundation model equipped with varied 3D representations—including elevation maps and voxel maps—that can be adapted for multiple crops with minimal adjustments.
To enhance this model, Bonsai has collected an impressive dataset, having deployed 400 units over roughly one million acres of specialty crops. They also utilize Nvidia's Cosmos family of world foundation models for physical AI which allows for advanced simulation capabilities. The model's ability to generate 3D coordinates and contextual depth is groundbreaking, according to Niday, enabling superior performance even in challenging conditions.
Expanding Crop Adaptability and Market Applications
Niday stated that Bonsai Robotics initially focused on almond orchards but has since broadened its applications. This expansion has been substantially aided by the acquisition of Farm-ng, allowing them to operate in diverse environments, including older orchards that lack GPS connectivity. By overcoming challenges such as dust and terrain variations, Bonsai has successfully adapted its model for a wide range of specialty crops, including strawberries, vineyards, apples, and citrus fruits.
The company’s approach currently balances two revenue streams: retrofitting existing equipment and selling proprietary Amiga machines. The operational flexibility of these machines is crucial. According to Niday, while contractor horsepower will always be a requirement, specific scenarios allow for autonomy-first platform designs that can significantly reduce operational costs.
The Importance of Flexibility and Investment in Ag Robotics
At the Reservoir Farms event, Niday highlighted the strategic value of the location for testing and demonstrating Bonsai's technology. He recalled experiencing initial difficulties in establishing connections within the agriculture community, but the venue provided immediate access to decision-making growers eager to see product demonstrations.
Currently, Bonsai Robotics has sold over 400 units, with approximately 75 of these being OEM retrofits, while the majority are proprietary Amiga machines. The company’s latest hybrid electric platform, the Amiga Max, has garnered significant interest, leading them to sell out of production for the current year.
Niday emphasized the critical role of flexibility in making the economics of these autonomous machines viable. Traditional agricultural tasks often suffer from inefficiencies, particularly with machinery that remains idle for most of the year. Redesigning vehicles to perform multiple roles, rather than serving a single purpose, presents substantial cost-reduction opportunities and appeals to buyers looking for more versatile options.
Looking Ahead: Commercial Viability and Automation in Agriculture
Addressing the growing interest in ag robotics, Niday remarked that past investment cycles have taught valuable lessons, positioning technology in agriculture as increasingly viable. The rapid advancements in the field have opened doors to addressing previously daunting tasks, such as robotic harvesting. According to Niday, the evolution of physical AI could significantly impact agricultural efficiency and labor market dynamics. The growing demand for automation across various sectors positions agriculture as an essential area for innovation.
In summary, Bonsai Robotics is leading the charge in transforming agricultural operations through advanced AI technology that bridges the gap between 2D imaging and 3D productivity. By enhancing flexibility and robustness in machinery, they are well-positioned to meet the demands of diverse crops while offering a tangible return on investment for growers. As the landscape of agricultural robotics continues to evolve, Bonsai aims to capitalize on these advancements, driving efficiency and reducing costs in an industry ripe for change.
*B-roll courtesy of Bonsai Robotics
