Stefan Seltz-Axmacher knows what it costs to chase autonomy on an open road. As the former CEO of Starsky Robotics, he oversaw the team that put the first driver-out semi-truck on a public highway, a technical feat that ultimately ran into the economic reality of scaling self-driving AI [Business Insider, Jun 2019]. Now, at Polymath Robotics, he and co-founder Ilia Baranov are selling a different kind of autonomy, one that doesn't have to navigate a four-way stop. Their customer is the fleet manager with a hundred 400-ton haul trucks moving ore in a closed-pit mine, or the OEM building the next generation of autonomous tractors. The bet is that the hard part isn't the vehicle, but the reliable, production-ready software stack that makes it safe and useful.
Polymath builds what it calls core autonomy and safety systems for off-highway industrial vehicles. It's a modular planning stack, safety system, and set of infrastructure tools delivered as a SaaS platform, designed to be hardware-agnostic [Polymath Robotics, retrieved 2024]. The company offers a REST API and SDK, bundling path planning, hazard detection, behavior trees, and controls tuning so that robotics teams at large equipment manufacturers or fleet operators can integrate autonomy without rebuilding the foundational layers from scratch [TechCrunch, Jul 2022]. The target is the estimated 50 million industrial vehicles operating in controlled environments today, from agriculture and mining to logistics yards and defense [TechCrunch, Jul 2022].
A wedge of practical abstraction
The company's wedge is abstraction. Where many robotics startups are building a specific autonomous vehicle for a single use case, Polymath is selling the software layer that could enable many of them. Their platform aims to handle the common, complex problems of real-world autonomy,perception, planning, safety monitoring, failover,so that customer engineering teams can focus on the application logic and business rules unique to their industry [Polymath Robotics, retrieved 2024]. This plug-and-play pitch is aimed at reducing development time from years to months, a value proposition that resonates in capital-intensive industries where equipment downtime is measured in thousands of dollars per hour.
It's a classic infrastructure play, positioning Polymath as the Oracle or SAP of industrial robotics, to use the analogy from TechCrunch [TechCrunch, Jul 2022]. The company also offers simulation tools, allowing potential customers to build and test applications on top of their autonomy stack for free in a virtual environment before committing to a hardware integration [Polymath Robotics, retrieved 2024]. This lowers the barrier to evaluation and fits a pragmatic, enterprise sales motion where proof-of-concept is a critical gate.
The team's second act in autonomy
The founding team brings a specific and hard-won credibility to this effort. CEO Stefan Seltz-Axmacher's experience at Starsky Robotics is a double-edged sword: it provides deep, firsthand knowledge of what it takes to deploy real-world autonomy, and it comes with the public lessons of a startup that shuttered, partly due to the escalating costs of AI development [Fortune, May 2020]. That background informs Polymath's constrained focus. They are not solving for every edge case on a public highway; they are solving for the predictable, but still dangerous, environment of a mine or farm.
CTO Ilia Baranov complements this with a strong robotics engineering pedigree. He led teams at Amazon Lab 126 working on the Astro home robot and was an early employee at Clearpath Robotics, a foundational industrial robotics company [RoboBusiness, retrieved 2026]. He also maintained the PR2 ROS repository after Willow Garage shut down, giving him deep roots in the open-source robotics software community that Polymath's platform likely builds upon [Polymath Robotics, retrieved 2026]. The team is reported to be between 17 and 25 people, a size that suggests focused product development rather than a sprawling research organization [Y Combinator, retrieved 2024] [PitchBook, retrieved 2024].
| Role | Name | Key Background |
|---|---|---|
| CEO & Co-Founder | Stefan Seltz-Axmacher | Former Co-Founder & CEO of Starsky Robotics; put first driver-out truck on public highway [Business Insider, Jun 2019]. |
| CTO & Co-Founder | Ilia Baranov | Former Software Development Manager at Amazon Lab 126; early employee at Clearpath Robotics [Boring Business Nerd, Unknown]. |
Traction and the capital runway
Public traction metrics are scarce, which is typical for an early-stage company selling into large industrial customers with long procurement cycles. Polymath states its software "runs across dozens of platforms and is trusted by top OEMs, billion-dollar fleet operators, and government programs" [Polymath Robotics, retrieved 2024]. This language points to design partnerships and early deployments, though specific names and contract values are not disclosed.
The funding picture is similarly lean and early-stage. The company has raised a total of $500,000 in a round led by Y Combinator in September 2022 [PitchBook, Sep 2022]. Tracxn also lists a $1 million seed round in February 2025, though the lead investor is not specified [Tracxn, Feb 2025]. This level of capital, while modest, aligns with a capital-light, software-focused model and Y Combinator's typical seed-stage checks. The investor list includes Samsara Ventures, TRAC, Bruckner Ventures, Blackbird Ventures, and Global From Day One, a mix that suggests interest from both generalist tech investors and those with a transportation or industrial focus.
2022 Early Stage VC | 0.5 | M USD
2025 Seed | 1 | M USD
Where the integration risk lives
For all its promise of abstraction, Polymath's success hinges on a deeply concrete challenge: integration. The company's platform must interface seamlessly with a wild variety of vehicle chassis, sensor suites, and legacy control systems from manufacturers like John Deere, Caterpillar, or CNH Industrial. Each integration represents a custom engineering effort, which can erode the margins of a pure software play and slow deployment velocity.
- Sales cycle length. Selling to top-tier OEMs and billion-dollar fleets involves long, multi-stakeholder evaluations. Polymath's simulation tools help, but the path from a successful pilot to a fleet-wide rollout can take years.
- The in-house build option. Large industrial players have the resources to develop their own autonomy stacks. Polymath must convince them that buying is faster, cheaper, and more reliable than building, a argument that often comes down to the quality and robustness of the core safety systems.
- Competitive depth. The field includes well-funded specialists like Outrider (yard trucks) and SafeAI (mining), as well as the internal projects of the equipment giants themselves. Polymath's agnosticism is its differentiator, but it also means competing with focused solutions that may be more optimized for a specific vertical.
The company's most plausible answer is that the complexity of a production-grade autonomy stack is consistently underestimated. By productizing the hardest parts,certifiable safety, reliable planning, and developer tooling,they offer a shortcut that even large engineering organizations may find compelling as time-to-market pressures increase.
The next twelve months
The immediate watch points for Polymath are commercial. The next twelve months should reveal whether their early partnerships mature into announced, referenceable customers with public case studies. Another funding round, likely a Series A, would be expected to scale the sales and solutions engineering teams needed to support those deployments. Technically, the focus will be on expanding the library of pre-validated integrations for common vehicle platforms, reducing the friction for new customers.
The ideal customer profile here is not a startup building a single robot. It's the autonomy program lead at a major industrial OEM or a large private fleet operator, someone with a budget to automate a fleet of existing assets and a mandate to do it without assuming untenable operational risk. They are buying a safety-certifiable software foundation, not just a navigation module.
The realistic competitive set splits into three lanes. First, the vertical specialists like Outrider and SafeAI. Second, the autonomy divisions of the equipment manufacturers themselves. Third, a handful of other generalist robotics software providers, though few are as sharply focused on the off-highway, industrial vehicle niche. Polymath's path is to out-execute the specialists on breadth and out-innovate the giants on speed and focus. It's a classic infrastructure bet, and in the dusty, demanding world of industrial vehicles, a reliable API might just be the most valuable tool in the shed.
Sources
- [Polymath Robotics, retrieved 2024] Polymath Robotics | Autonomy & Safety Systems for Off-Highway Vehicles | https://www.polymathrobotics.com/
- [TechCrunch, Jul 2022] Polymath Robotics launches to bring plug-and-play autonomy software to any industrial vehicle | https://techcrunch.com/2022/07/29/polymath-robotics-launches-to-bring-plug-and-play-autonomy-software-to-any-industrial-vehicle/
- [Business Insider, Jun 2019] Starsky Robotics achieved the industry feat of running driverless semi-trucks on the open road | https://www.businessinsider.com/starsky-robotics-driverless-trucks-highway-self-driving-2019-6
- [Fortune, May 2020] Starsky Robotics shuttered, with founder attributing failure partly to AI costs | https://fortune.com/2020/05/19/starsky-robotics-self-driving-truck-startup-shuts-down/
- [RoboBusiness, retrieved 2026] Ilia Baranov profile | https://robobusiness.com/speaker/ilia-baranov/
- [Y Combinator, retrieved 2024] Polymath Robotics company profile | https://www.ycombinator.com/companies/polymath-robotics
- [PitchBook, Sep 2022] Polymath Robotics funding details | https://pitchbook.com/profiles/company/523452-04#funding
- [Tracxn, Feb 2025] Polymath Robotics funding rounds | https://tracxn.com/d/companies/polymath-robotics
- [Boring Business Nerd, Unknown] Founders background reference
- [Forbes, Jun 2020] Stefan Seltz-Axmacher profile as autonomy pioneer