Antioch, a startup building cloud-scale simulation infrastructure for robotics and autonomous-system developers, has raised a $32 million Series A led by Greylock, with participation from A*, Category Ventures, BoxGroup, and Icehouse Ventures. It follows an $8.5 million seed round that closed in April 2026 at a $60 million valuation — a fast jump in both round size and backer profile for a company barely a year and a half old.
The pitch behind Antioch is straightforward once you see the cost problem it’s aimed at: training and validating a physical robot against real-world conditions is slow and expensive, because every test requires an actual machine, an actual environment, and actual time. Antioch’s platform lets robotics and autonomy teams build digital twins of their systems and environments, generate synthetic training data, and run thousands of parallel test scenarios in simulation before ever deploying hardware into the real world. The company describes itself as bringing “software speed” to a development process that has historically moved at the pace of physical prototyping.
Already Working With Amazon’s Ring
Antioch says it is already working with leading physical AI teams, including Amazon’s Ring, to accelerate development cycles while dramatically increasing how much testing can be run before a system ships into the real world. That’s a meaningful validation point in a category where enterprise customers are typically the hardest sell — a robotics or autonomy team has to trust that a simulated test result actually predicts real-world behavior closely enough to bet a product launch on it, and landing a name like Ring suggests at least one serious operator is willing to make that bet.
The round’s angel roster is also telling about who’s paying attention to this space. Backers include Shyam Sankar, Palantir’s chief technology officer; Adrian Macneil, chief executive of robotics tooling company Foxglove; and Ian Andrews, an Nvidia executive and former Groq executive — a lineup that spans defense-adjacent software, robotics infrastructure, and AI chip hardware, suggesting Antioch is being read as infrastructure relevant across all three of those worlds rather than a niche tool for one narrow vertical.
Simulation as the New Bottleneck Everyone’s Racing to Fix
Antioch’s raise lands in the middle of a broader surge of capital into what investors are now calling “physical AI” — the application of large-model-style AI techniques to robots, autonomous vehicles, and other systems that have to operate in the physical world rather than purely in software. That category has attracted some of the largest rounds of the year, and the shared thesis across most of them is that physical AI’s real bottleneck isn’t model capability anymore, it’s the cost and speed of generating enough real-world-equivalent training and testing data to make those models reliable. Simulation infrastructure sits directly on top of that bottleneck, which is why investors have been willing to write large early checks into companies like Antioch well before they’ve built out a long enterprise customer list — the market opportunity scales with every other physical AI company that needs the same underlying tooling.
What This Means for Philippine Founders
Antioch’s category — venture-scale simulation infrastructure for humanoid robots and autonomous vehicles — isn’t one Philippine founders are likely to compete in directly any time soon; it demands the kind of deep technical bench and capital intensity that’s still rare locally. But the underlying idea behind the raise, that testing in simulation is dramatically cheaper and faster than testing in the physical world, applies directly to problems the Philippines already has real commercial stakes in. The country’s e-commerce fulfillment and logistics sector is scaling warehouse automation and delivery-robotics pilots without anything close to the simulation tooling a company like Antioch provides, which means every pilot currently burns real time and real hardware to learn lessons that could, in principle, be learned in software first.
There’s also a more immediate opportunity in the Philippines’ own manufacturing and electronics-assembly base, which already serves as a testing and light-assembly hub for hardware built elsewhere. As physical AI and robotics companies abroad scale up the volume of real-world validation they need to run, some of that testing and calibration work — quality assurance, sensor calibration, edge-case data collection — is exactly the kind of structured, procedure-driven technical labor the Philippines’ BPO and engineering-services sector has built decades of credibility doing for other industries. A founder who can package that expertise specifically for physical AI and robotics clients, rather than trying to build the simulation platform itself, has a genuinely reachable wedge into a fast-growing category most local players currently have no foothold in at all.
The broader pattern in Antioch’s raise is also worth studying on its own terms, separate from robotics specifically: investors are rewarding startups that remove a real, quantifiable bottleneck from someone else’s expensive R&D cycle, rather than startups building yet another end-user product. That’s a lesson that travels well beyond physical AI. Whatever category a Filipino founder is building in, the equivalent question is worth asking directly — what is the most expensive, slowest, most repeated step your target customer currently has to do in the real world, and can a piece of software plausibly let them do a version of it faster, cheaper, and mostly in simulation or in software first. Antioch didn’t need to build a robot to raise $32 million; it needed to identify the one step in every robotics company’s roadmap that was genuinely too slow and too expensive, and build the tool that fixes just that.
Share this article