SafeWorld, a Palo Alto-based artificial intelligence lab focused on robot safety, has emerged from stealth with $12.2 million in seed funding to expand its simulation and testing platform for autonomous machines. The round was co-led by Shine Capital and a16z Speedrun, with participation from BoxGroup, the Carnegie Mellon University Endowment, Innovation Endeavors, SV Angel, and several other venture firms and angel investors. The company is positioning its software as infrastructure for enterprises seeking to evaluate robotic systems before deploying them in environments where machines increasingly operate alongside people.
Building a Safety Layer for Physical AI
SafeWorld is targeting a growing challenge in robotics as artificial intelligence shifts from software-based applications into machines that can move, make decisions, and interact with the physical world. Traditional robot validation often depends on costly and time-consuming field testing, which can make it difficult to examine the large number of rare or hazardous scenarios that may occur during real-world operations. SafeWorld argues that scalable simulation can help manufacturers and operators test edge cases more efficiently while reducing the need to expose people or equipment to unnecessary risk.
How the Platform Works
The company’s platform allows teams to create safety scenarios through a browser using information drawn from previous incidents, robot logs, and established safety requirements. SafeWorld can then run robotic systems through thousands of variations that incorporate realistic human movement and changing environmental conditions, enabling customers to measure how machines respond to potentially dangerous situations. Because new software versions and operating environments can introduce fresh risks, the platform is designed for repeated testing rather than one-time validation before deployment.
Founding Team Combines Research and Startup Experience
SafeWorld was co-founded by Ding Zhao, Kyle Wong, and Simo Rachidi, bringing together backgrounds in autonomous systems research, machine learning, enterprise technology, and startup operations. Zhao directs Carnegie Mellon University’s Safe AI Lab and previously worked at Google DeepMind, while Wong founded the AI and user-generated content platform Pixlee and later served as chief executive of Stanford-affiliated accelerator StartX. Rachidi is a repeat founder and former principal security and machine learning engineer at Salesforce Einstein, where he worked on large-scale data and machine learning infrastructure.
Investors Back the Robot Safety Thesis
Investors supporting the round said the expansion of autonomous machines will require stronger tools for testing and validating how robots behave around people. Shine Capital General Partner Alex Hartz described safety as a continuous layer that must develop alongside increasingly capable machines, while a16z Speedrun General Partner Jon Lai emphasized the shift from manual field trials toward automated simulation. The financing also included Ovo Fund, Valkyrie, Zelda Ventures, Alpha Square Group, Founders Future, Brave Capital, and individual investors from companies including NVIDIA, Google DeepMind, Waymo, Meta, DoorDash, Together AI, and Salesforce.
Early Enterprise Pilots
SafeWorld said it is already running early pilots with large companies across the automotive, warehouse automation, and medical device sectors, although it did not name most of those customers. Anyware Robotics CEO Thomas Tang said simulation tools such as SafeWorld can help robotics companies test difficult situations at scale and strengthen safety processes before real-world deployment. The early commercial work is intended to validate the platform across different robotic systems and operating environments as companies look for practical ways to introduce more capable autonomous machines.
SafeWorld enters the market as robotics companies face increasing pressure to prove that autonomous systems can operate safely in complex human environments. Its $12.2 million seed round gives the company additional resources to develop simulation tools, expand enterprise testing, and build what it sees as a broader safety infrastructure layer for physical AI. If adoption grows alongside the robotics market, the company could play a role in moving safety validation from slow physical testing toward continuous software-based assessment.