CNBC Television · Sep 28, 2026
With Jake Loosararian, Co-founder and CEO of Gecko Robotics
Gecko Robotics has partnered with Nvidia to run AI agents on its robots under limits that humans set and can pull back, Jake Loosararian said, with its Defense Department contracts especially in mind.
The weekend's AI safety news was about software, with reports of security incidents involving OpenAI and Anthropic. Loosararian's concern is robots in the physical world, in harsh industrial settings where models have almost no data to learn from.
"They have advantages of manufacturing where US has advantages on the intelligence and the brain of the robotic stacks."
Jake Loosararian, co-founder and CEO of Gecko Robotics, on CNBC Television, runs a company whose robots collect data on industrial surfaces for customers including the US military, where one contract is aimed at getting ships out of dry dock faster.
The full segment is covered here so you can skip it.
Here are the 4 takeaways that matter.
Key Takeaways
The Nvidia partnership is meant to let robots use AI agents without leaving human control
Self-driving cars already use these models; harsh industrial sites largely do not, because there is almost no data on them
Humans set the guardrails and can reel the robots back in
One use is Defense Department work to get ships out of dry dock faster
China leads on manufacturing and the US leads on the models that run robots, Loosararian said, and the edge goes to whoever deploys safely
1. The Nvidia Partnership
The host opened on the weekend's headlines about security concerns and breaches by AI bots at OpenAI and, to a lesser extent, Anthropic, and asked what Gecko's partnership with Nvidia enables.
Humans stay in control while robots use AI agents
"This whole announcement is all about ensuring that we keep humans in control while deploying agentic agents as fast as you can in the robotic sector, to ensure that we're able to use the benefits of robotics, especially with those DOD contracts that we have at Gecko, but also ensure that safety is paramount." — Jake Loosararian
Agentic agents here means AI systems that plan and take actions on their own, rather than only answering questions.
