BENGALURU, INDIA | February 12, 2026 – CynLr, the Bengaluru-based deep-tech pioneer, today announced the official launch of its Object Intelligence (OI) Platform, a fundamental leap in robotics that enables machines to learn and adapt on-the-fly without the need for traditional, time-consuming retraining. After five years of intensive R&D and neuroscience-linked collaboration with the Indian Institute of Science (IISc), CynLr is bridging the most significant gap in physical AI: the ability for robots to interact with an unpredictable world.
The End of Pre-Programmed Automation
For over half a century, industrial robotics has been defined by rigidity. Robots were programmed for specific, repetitive movements in controlled environments. If a part shifted by a centimeter or a new object was introduced, the system failed, requiring months of data collection and reprogramming.
CynLr’s OI Platform shatters this paradigm. By mimicking the biological sensorimotor learning of a human infant, the platform allows robots—such as CynLr’s own CyRo, CyNoid, and Mantroid—to perceive, reason, and manipulate unknown objects in real time.
“The last fifty years were about controlled environments—machines that could repeat but not respond,” says Gokul NA, Founder of CynLr. “The next fifty years will be about cognition. We are building the ‘Universal Factory’—a software-defined space where a single robotic line can handle a thousand different SKUs just by updating software, without expensive retooling.”
On-the-Fly Learning: From Months to Seconds
At the heart of this breakthrough is the CLX Vision System, an OI-enabled engine that treats objects not as static images, but as a “recipe” of geometry, texture, reflectance, and grasp possibilities.
While current Physical Intelligence methods rely on massive data centers and months of training, CynLr’s robots can learn to handle entirely unseen objects—including notoriously difficult items like transparent glass, reflective metallic car parts, or irregular electronics—usually within 10 to 15 seconds. This “Vision Force Model” approach allows the robot to learn through interaction. Every failed grasp is a data point for immediate recalibration, creating a closed-loop system where the robot’s “intuition” improves with every attempt.
Bridging the Gap Between AI and Reality
A key differentiator for CynLr is its departure from Vision Language Models (VLMs). According to Nikhil, Founder-GTM at CynLr, current AI struggles because it interprets the world through the lens of human-generated text and static 2D images.
“Human beings act to sense—you pick up a coin and tilt it to see it better. You don’t train on a dataset of blurry coins,” Nikhil explains. “Robots need their own eyes and touch. Object Intelligence enables dynamic sensing, moving away from the ‘blind man talking about rainbows’ approach of current AI.”
From Lab to the Global Supply Chain
Unlike many AI solutions that remain theoretical or confined to research labs, CynLr’s technology is already seeing real-world application. The company’s customer base includes global luxury automotive brands and leading semiconductor automation firms. These robots are being deployed for complex tasks—such as assembly line integration and maintenance in Semiconductor Fab Labs—where traditional automation was previously impossible.
The platform is form-factor agnostic, meaning it can power everything from standard industrial arms to sophisticated multi-arm humanoid systems.
CynLr’s Current Product Suite Includes:
- CyRo & CyNoid: Advanced robotic systems currently being piloted by global manufacturers.
- Mantroid: An upcoming “open hardware” sandbox platform. This allows customers to move away from fixed humanoid shapes and custom-build robot form factors tailored to their specific operational needs.
Global Expansion and the Road to 2028
With a newly established R&D entity in Switzerland and a business development center in the United States, CynLr is positioned for a global scale-up. The company is currently pursuing its next round of funding with an ambitious production goal: manufacturing one robot per day by 2028.
By reducing the reliance on massive datasets and specialized hardware for every new task, CynLr is not just optimizing the factory floor—it is enabling the future of flexible, resilient manufacturing.
The company will soon release technical demonstrations featuring visualizations of neural activations and real-time learning behaviors, providing a window into how “Object Intelligence” is finally giving machines the cognitive spark they have lacked for decades.
