The robot stops. It’s not stuck on a pallet or confused by a new box shape. It’s waiting for you, the human, to finish your conversation. It has read the room, predicted your path, and decided the safest, most efficient thing to do is nothing at all. This is the moment scaledrive.ai is building for: the pause that proves a machine understands context.
Based in Munich, scaledrive.ai is developing what it calls “certifiable software powered by foundational models” for logistics and industrial robots [scaledrive.ai]. The core proposition is not about making robots faster or stronger, but about making them socially aware. The goal is to move autonomous systems from sterile, controlled settings into the messy, human-dense environments of warehouses and factories, where unpredictability is the main source of downtime.
The bet on Physical AI
The company’s wedge is a proprietary software stack it brands as “Physical AI.” This isn’t just another layer of computer vision. The technology combines generative AI models for human behavior prediction and environmental context reasoning with what the company describes as a “proprietary safety envelope” designed to meet certification standards [scaledrive.ai]. The certification angle is critical. It’s the promised bridge from interesting research to commercially deployable, insurable products. The software is meant to be the reasoning layer that allows any autonomous mobile robot (AMR) or industrial arm to adapt on the fly, slowing, stopping, or rerouting based on a dynamic forecast of human activity, not just a static map.
A team built at the intersection
The founding team brings together academic depth and industrial pragmatism from notable institutions. The team includes alumni from BMW and Stanford University, a combination that speaks to scaledrive.ai’s dual focus on rigorous research and real-world application [scaledrive.ai].
- Denis Azarov (CEO). An entrepreneur with international cross-functional experience in both corporate and startup environments [LinkedIn]. He studied at the Technical University of Munich (TUM).
- Jakob Thumm (Research). A postdoctoral scholar at Stanford University whose work involves “teaching robots to safely work with humans” [jakob-thumm.com]. He completed his PhD in Informatics at TUM in 2025 [ce.cit.tum.de].
- Tim Salzmann (Research). A PhD student affiliated with TUM, Stanford, and Google DeepMind [LinkedIn].
- Ahsan Ahmed (Engineering). Listed as part of the core engineering team on the company’s site [scaledrive.ai].
Traction in early deployments
While specific customer names and deal sizes are not public, the company reports it has “recently started paid physical deployments with autonomous mobile robots” [scaledrive.ai]. This is a significant signal for a pre-seed stage company founded in 2024. The company also participated in the Plug and Play Tech Center accelerator program [Plug and Play Tech Center]. A pre-seed funding round was secured in 2025 [Fundraise Insider, 2026].
The competitive landscape
On paper, scaledrive.ai operates in a space crowded with well-funded giants and ambitious startups. The company lists Figure AI, Waymo, and Anduril Industries as competitors. However, its specific focus on a certifiable software layer for existing industrial and logistics robots suggests a different go-to-market path.
| Company | Primary Focus | scaledrive.ai's Angle |
|---|---|---|
| Figure AI | General-purpose humanoid robots | Software for existing, specialized robots |
| Waymo | Autonomous passenger vehicles | Industrial environments (warehouses, factories) |
| Anduril Industries | Defense and national security systems | Commercial, civilian logistics |
Where the wheels could come off
- The certification marathon. Building a “safety envelope” that regulatory bodies and insurance companies will accept for dynamic human-robot interaction is an untested, multi-year endeavor.
- The integration burden. The value of the software is only realized when it is deeply integrated into a robot’s control systems.
- Proving economic ROI. The company must demonstrate that its software doesn’t just make robots safer, but also more productive and cost-effective.
The next twelve months
The immediate horizon for scaledrive.ai will be defined by scaling its initial deployments into repeatable customer case studies. The next milestone is likely a named partnership with a robotics OEM or a major logistics operator. Given its pre-seed status and active deployments, a seed round to fund further commercial expansion and team growth seems a probable move within the coming year.
Sources
- [scaledrive.ai, retrieved 2024] Company website | https://www.scaledrive.ai/
- [Fundraise Insider, 2026] Funding report | https://www.fundraiseinsider.com/company/scaledrive-ai
- [Plug and Play Tech Center] Accelerator profile | https://www.plugandplaytechcenter.com/startup/scaledrive-ai
- [LinkedIn, retrieved 2026] Denis Azarov profile | https://www.linkedin.com/in/denisazrv/
- [jakob-thumm.com, retrieved 2026] Jakob Thumm research | https://jakob-thumm.com/
- [ce.cit.tum.de, retrieved 2026] Jakob Thumm PhD information | https://ce.cit.tum.de/
- [LinkedIn, retrieved 2026] Tim Salzmann profile | https://www.linkedin.com/in/timsalzmann/