Antioch co-founders Harry Mellsop, Alex Langshur, Colton Swingle, and Collin Schlager. Source: Antioch
Safely validating artificial intelligence to run physical systems in the real world is one of the hardest challenges in autonomy, according to Antioch Inc. The startup has raised pre-seed funding from investors including Shyam Sankar, chief technology officer of Palantir Technologies Inc.
Stanford University graduates Harry Mellsop, Alex Langshur, Colton Swingle, and Collin Schlager co-founded Antioch early in 2025. The New York-based company said it is building a simulation platform to enable robotics teams to build, test, and deploy autonomous systems entirely in software.
Antioch builds infrastructure to democratize development
Antioch’s stated mission is to eliminate lengthy testing cycles that slow robotics progress by giving developers access to “Tesla-level infrastructure.”
“At Tesla, I saw first-hand how the right simulation tools let you develop and iterate exceptionally fast,” said Mellsop, who previously worked on the Tesla Autopilot vision team. “We built Antioch to give every robotics company that same advantage — the ability to test and scale their robotics system at the speed of software.”
Antioch’s other co-founders have complementary experiences. Swingle previously led several large-scale model validation infrastructure projects at Google’s DeepMind unit, while Schlager played a leading role in foundational research programs for Meta Reality Labs’ recently released Neural Band wearable device.
Both oversaw the development of simulation and testing infrastructure for the validation of these products, and they said they saw firsthand the need to bring these capabilities to the wider market.
“At Meta, evaluating our hardware and models – particularly in a manner representative of real-world use – was critical to moving fast and to providing an informative feedback loop,” recalled Schlager. “However, developing the platforms and tooling to carry out these evaluations required enormous investment. With Antioch, we’re making this publicly accessible for the first time, so teams can focus on perfecting their products, and not building the evaluation infrastructure around it.”
Instead of laboriously resetting a physical machine that takes up energy, space and money every time it’s tested, Antioch is intended to allow developers to refine their robots with minimal overhead. Engineers can spin up thousands of digital twins in parallel using Antioch Cloud, integrate with continuous integration and deployment (CI/CD) workflows, and track performance metrics before any hardware ever leaves the lab, claimed the startup.
“Right now, the way companies train AI for robots is absurdly manual, ” said Mellsop. “We’ve talked to teams literally renting Airbnbs just so they can test their household robots overnight or spending millions building fake warehouses and neighborhoods to simulate the real world. It’s slow, expensive, and wildly inefficient. With Antioch, you can do all of that virtually.”
Founders apply experience to national needs
Market leaders in the autonomy space such as Tesla, Anduril, and Waymo spend hundreds of millions of dollars a year on end-to-end system evaluation, said Antioch. This creates a significant market opportunity, but it also puts these tools out of reach for smaller players and new entrants, it noted.
Mellsop and Langshur asserted that Antioch’s mission is not only about responding to the acute commercial need, but it also reflects their commitment to building infrastructure that is critical for national security but neglected by existing players.
The pair previously built Transpose, which they sold to the U.S. contractor Chainalysis in 2023. Transpose served intelligence and national security agencies, law enforcement, regulators, and financial institutions in the U.S. and abroad. It provided a security and intelligence layer that helped government and the private sector defend against nation-state cyber threats.
“Over the last 40 years, the manufacturing capability that set the United States apart from the rest of the world has been systematically eroded by offshoring,” Langshur said. “We see rapid re-industrialization as imperative from a national security perspective, and the only economically viable way to do this is through augmenting our workforce with robotics and automation.”
“Efficient, scalable testing to ensure that robots deployed are safe and effective is now the rate-limiting step in making this happen, which is why we decided that Antioch needed to be built now,” he added.
Antioch integrates with existing AI tools
Antioch said it has tightly integrated with the technology of industry leaders, including tools like Omniverse and Cosmos from NVIDIA and observability from AI robotics company Foxglove.
The company said it is already working with many leading robotics and autonomy companies, including Fortune 500s, in verticals like construction robotics, smart security systems, and foundation model AI developers. It expects a full commercial launch in late 2025.
“We’ve taught AI to think,” Mellsop said. “Now it’s time to teach it to act in the real world safely, intelligently, and at scale.”
Investors hope to accelerate robotics innovation
A* led Antioch’s pre-seed round, with participation from Abstract Ventures, MaC Venture Capital, and Box Group. Other investors included angels Adrian Macneil, the CEO of Foxglove; Palantir’s Sankar; and New Zealand firm Icehouse Ventures.
Antioch’s core technology is an equalizer for companies to explore the next generation of robotics,” said Bennett Siegel, co-founder and general partner of A*.
“Antioch sits at the intersection of AI and robotics and will unlock the sci-fi future we’ve often dreamed of,” he said. “This platform removes the friction from physical-world testing and enables a new generation of embodied AI startups to scale their inventions globally. Companies are currently spending hundreds of millions a year to have a horse in the AI robotics race, but technology like Antioch’s can reduce the hurdles considerably, giving more businesses a chance to innovate.”
Antioch has raised $4.24 million, according to Grishin Robotics. The startup is seeking a founding software engineer. Other companies that have recently raised millions in funding for physical and embodied AI development include Foxglove, Forterra, Physical Intelligence, Flexion, and ROBOTERRA.
Tell Us What You Think!