Physicl announced on Monday that it has closed a $10 million financing round to accelerate its work on closing the simulation-to-real (sim-to-real) gap for robots. The funding will be used to expand a pipeline originally built to market goods online, adapting it so that virtual models can be transferred to physical robots with predictable performance.
The announcement matters because the sim-to-real gap remains one of the biggest obstacles to scaling robotic solutions in manufacturing, logistics and service sectors. By leveraging a system designed for high-fidelity visualisation of products, Physicl hopes to give developers a reliable way to test and deploy robot behaviours without costly physical prototypes.
What Is the Sim-to-Real Gap and Why Does It Matter?
In robotics, engineers often train algorithms in simulated environments because computers can generate endless scenarios quickly and safely. However, when a robot moves from a virtual world to a physical one, differences in lighting, friction, sensor noise and hardware tolerances can cause failures. This discrepancy is known as the sim-to-real gap.
Closing the gap means that a robot that performs well in simulation will behave similarly when deployed on a factory floor or in a warehouse. Success reduces development time, lowers capital expenditure on physical prototypes, and speeds up adoption of autonomous systems across industries.
How Physicl's Pipeline Was Originally Used
Physicl's technology was first created to help retailers showcase products online. The pipeline generates photorealistic 3D models, simulates lighting and material properties, and allows customers to view items from any angle. By ensuring that the virtual representation matches the real product, retailers can reduce returns and improve buyer confidence.
Key components of the original system include:
- High-resolution 3D scanning of physical objects.
- Physically based rendering (PBR) engines that replicate real-world light interaction.
- Automated calibration tools that align virtual dimensions with measured data.
These capabilities are now being repurposed to create digital twins of robots and their workspaces, providing a foundation for accurate transfer of control policies.
Adapting the Pipeline for Robotics
Physicl's approach involves three steps:
- Digital Twin Creation: Using the same scanning technology, the company builds a precise virtual replica of a robot, including joint limits, sensor placements and end-effector geometry.
- Simulation Environment Alignment: The pipeline integrates physics engines that model contact forces, friction coefficients and material compliance, matching the conditions measured in the physical lab.
- Policy Transfer and Validation: Control policies trained in the aligned simulation are exported to the robot's controller. Real-world trials then verify performance, with the system automatically adjusting parameters to reduce error.
This workflow aims to make the transition from code to hardware as seamless as uploading a product image to an e-commerce site.
Funding Details and Strategic Use of Capital
The $10 million round was led by investors familiar with both robotics and visual technology. While the press release does not list individual investors, the capital will be allocated to three primary areas:
- Hiring additional robotics engineers and simulation specialists.
- Expanding cloud-based compute resources for large-scale simulation runs.
- Developing industry-specific modules, such as for pick-and-place in e-commerce fulfilment centers and for assembly line inspection.
Physicl expects the first commercial version of its adapted pipeline to be ready for beta testing with partner manufacturers by the second quarter of next year.
Industry Context: Why Sim-to-Real Solutions Are in Demand
Global demand for robotics automation is rising. According to the International Federation of Robotics, worldwide robot installations grew by 12 percent in 2023, reaching over 3 million units. Yet many manufacturers cite the cost of moving from simulation to production as a barrier.
Companies such as NVIDIA, Unity and OpenAI have released tools to improve realism in virtual environments, but few provide an end-to-end pipeline that includes calibrated physical deployment. Physicl's unique angle is its heritage in e-commerce visualisation, which already demands sub-millimetre accuracy and realistic material rendering.
Potential Impact on Different Sectors
By reducing the time and expense of physical testing, the technology could benefit:
- Manufacturing: Faster rollout of collaborative robots (cobots) for assembly tasks.
- Logistics: More reliable autonomous mobile robots for warehouse order fulfillment.
- Healthcare: Safer deployment of assistive robots that must operate alongside patients.
Small- and medium-size enterprises that lack large R&D budgets stand to gain the most, as the platform could lower entry barriers to advanced automation.
What Comes Next for Physicl?
In the coming months, Physicl will finalize integration with major physics engines such as MuJoCo and Isaac Gym, and will begin pilot programs with at least two manufacturing partners. The company has indicated that it will publish benchmark results comparing its pipeline's transfer accuracy against existing methods.
Regulatory considerations are minimal for the software itself, but any deployment of robots in safety-critical environments will still require compliance with standards such as ISO 10218 and ISO/TS 15066. Physicl plans to provide documentation that helps customers meet these standards when using its digital-twin workflow.
Stakeholders should watch for announcements of beta partners, performance metrics released by the company, and any subsequent funding rounds that may signal broader market validation.
