The Robot Learning Company
Open-source AI-native robotics dev kits and software for affordable physical automation.
Website: https://www.robot-learning.co/
Cover Block
Public sources
| Name | The Robot Learning Company |
| Tagline | Open-source AI-native robotics dev kits and software for affordable physical automation. [Company site, July 2026] |
| Headquarters | Munich, Germany [Y Combinator, June 2025] |
| Founded | 2025 [Y Combinator, June 2025] |
| Stage | Pre-Seed [PitchBook, Feb 2025] |
| Business Model | Hardware + Software [Altis Research] |
| Industry | Deeptech [Altis Research] |
| Technology | Robotics [Altis Research] |
| Geography | Western Europe [Altis Research] |
| Growth Profile | Venture Scale [Altis Research] |
| Founding Team | Solo Founder [Altis Research] |
| Funding Label | Pre-seed [PitchBook, Feb 2025] |
| Total Disclosed | $500,000 [PitchBook, Feb 2025] |
Links
Public sources
- Website: https://www.robot-learning.co/
- LinkedIn: https://www.linkedin.com/company/the-robot-learning-company
- GitHub: https://github.com/robot-learning-co
Executive Summary
Public sources The Robot Learning Company is building a low-cost, open-source hardware and software platform designed to make AI-native robotics accessible, a bet that the next wave of automation will be driven by developers rather than large integrators [Y Combinator, June 2025]. Founded in 2025 by Jannik Grothusen, the company emerged from his research at UC Berkeley and TU Munich, where he demonstrated the potential of combining low-cost hardware with advanced AI models to perform physical tasks [TechCrunch, Nov 2024]. Its core product is the TRLC-DK1, a developer kit that packages a compact robot arm, teleoperation hardware, cameras, and a cross-platform SDK into a single unit priced in the low thousands of dollars, aiming to serve as a robotics-specific MLOps layer [PlatoSeed] [Altis Research].
Grothusen's background in mechatronics and robot learning research provides technical credibility, though the exact composition of the founding team beyond him is not fully corroborated by primary sources [Fundraising Fox, Sept 2026]. The company raised a $500,000 pre-seed round from Y Combinator in April 2025, which remains its only publicly disclosed financing [PitchBook, Feb 2025]. Its business model combines hardware sales with a software-as-a-service layer, evidenced by a lease option for the dev kit starting at $499 per month [SVRC, 2026]. Over the next 12-18 months, the key milestones to watch are the transition from pilot discussions to named commercial deployments and the company's ability to attract a developer community around its open-source tools, which will validate both its market wedge and its scalability beyond academic and research labs.
Lightly corroborated -- Core product and funding round confirmed by multiple sources; team background and estimated metrics rely on limited or secondary corroboration.
Taxonomy Snapshot
| Axis | Classification |
|---|---|
| Stage | Pre-Seed |
| Business Model | Hardware + Software |
| Industry / Vertical | Deeptech |
| Technology Type | Robotics |
| Geography | Western Europe |
| Growth Profile | Venture Scale |
| Founding Team | Solo Founder |
| Funding | Pre-seed (~$500,000) |
How the Company Got Here
Public sources
The Robot Learning Company was founded in March 2025 by Jannik Grothusen, a researcher with academic ties to UC Berkeley and TU Munich [LinkedIn] [Fundraising Fox, Sept 2026]. The company operates as a dual-headquartered entity, with bases in Munich, Germany, and San Francisco, California, a structure common for early-stage startups with roots in European academia and the Silicon Valley venture ecosystem [LinkedIn] [Perplexity Sonar Pro]. Its formation was catalyzed by a hands-on robotics project Grothusen co-led in late 2024, which demonstrated a low-cost robot arm performing a cleaning task guided by GPT-4o [TechCrunch, Nov 2024].
A key early milestone was acceptance into the Y Combinator accelerator program, which culminated in a $500,000 pre-seed investment in April 2025 [PitchBook, Feb 2025]. This capital injection and the associated programmatic support provided the initial runway to transition from research prototypes to a commercial product. By mid-2026, the company had publicly launched its first product line, the TRLC-DK1 open-source dev kit, and updated its website to reflect this commercial offering [Company site, July 2026] [GitHub, 2026].
The founding team composition presents some ambiguity in public records. While Jannik Grothusen is consistently identified as the Founder/CEO, LinkedIn profiles also list Cristian Pinto and Clément Grivel as co-founders [Y Combinator, June 2025] [LinkedIn, 2026]. The company's public narrative, however, centers on Grothusen's technical vision and research background, with other team members like Rasmus Soeborg and Daniel Friis appearing in associated professional networks [LinkedIn, 2026].
Lightly corroborated -- Core founding date and YC funding confirmed by PitchBook and LinkedIn; team details are partially corroborated but show inconsistencies across sources.
Product and Technology
Sources and analysis
The Robot Learning Company's public proposition centers on a hardware-software bundle designed to lower the barrier to AI-driven robotics experimentation. The company describes its core offering as an "Open Source Dev Kit for AI-native Robotics," a phrase that appears verbatim on its website [Company site, July 2026]. This dev kit is not a single product but a platform, with the TRLC-DK1 for single-arm tasks and the TRLC-DK1-X for bimanual operation [PlatoSeed]. The hardware is characterized by its affordability, with a pair of stationary arms priced "under $10,000" [Y Combinator].
On the software side, the platform functions as a robotics-specific MLOps layer, providing a unified environment for data collection, model training, and deployment [Altis Research]. The workflow is designed to be accessible: users demonstrate a task using a teleoperation "leader" device while cameras record, then train a model locally using the provided SDK, and finally deploy the learned policy to the physical "follower" arm [PlatoSeed]. The company emphasizes techniques like imitation learning and offline reinforcement learning, claiming robots can adapt to new tasks from "just a handful of demonstrations" [Y Combinator, June 2025]. The software stack supports Windows, macOS, and Linux via a USB-C connection [PlatoSeed].
Public demonstrations provide the clearest evidence of the technology's application. In November 2024, founder Jannik Grothusen co-led a project that trained a pair of $120 robot arms to clean a spill using GPT-4o, programming the system in four days [TechCrunch, Nov 2024]. This project, which designed a visual language model for human-robot interaction, serves as a tangible, published proof point for the type of rapid prototyping the company's tools aim to enable. The company's GitHub repository shows active development, including a recent update adding support for bimanual operation to the TRLC-DK1 [GitHub, 2026].
Independently corroborated -- Product specifications and workflow are detailed across multiple independent sources, including the company website, Y Combinator profile, and technical coverage. The GPT-4o demonstration was reported by TechCrunch.
Where the Demand Sits
Public sources The market for affordable, AI-native robotics is emerging from academic labs into industrial pilot programs, driven by a convergence of cheaper hardware, more capable foundation models, and a growing need to automate repetitive tasks. While a precise, third-party TAM for this specific niche is not yet established, the demand is anchored in the broader, multi-billion dollar industrial automation and collaborative robotics (cobots) sectors.
Demand is being pulled by several identifiable tailwinds. The primary driver is a persistent labor shortage for repetitive, low-skill tasks in sectors like manufacturing, logistics, and warehousing, which increases the economic incentive for automation [Y Combinator, June 2025]. Concurrently, the cost of core robotic components has fallen, making hardware more accessible, while advances in AI, particularly in computer vision and large language models, have simplified the programming and adaptability of robots [TechCrunch, Nov 2024]. This has shifted the bottleneck from hardware cost to software complexity, creating a market for tools that lower the barrier to deploying intelligent automation.
Key adjacent markets provide useful analogies for sizing. The global collaborative robot market was valued at approximately $1.2 billion in 2023 and is projected to grow at a compound annual rate of over 30% [analogous market, Interact Analysis, 2024]. The market for robotic software platforms and simulation tools, a closer proxy for TRLC's MLOps layer, is also expanding rapidly as companies seek to manage fleets and AI training pipelines. The company's stated wedge targets the segment of this market seeking low-cost, stationary automation for tasks like machine tending, inspection, and light assembly, where traditional industrial robots are often over-engineered and cost-prohibitive [PlatoSeed].
Regulatory and macro forces present a mixed picture. In Western Europe and North America, safety standards for collaborative robots are well-established, providing a clear compliance framework for new entrants. However, geopolitical tensions affecting semiconductor and actuator supply chains could impact hardware cost and availability. The most significant macro force is the accelerating enterprise adoption of AI, which is creating organizational readiness and budget for AI-adjacent automation projects, even in early-stage pilots.
| Metric | Value |
|---|---|
| Collaborative Robot Market (2023) | 1.2 $B |
| Projected CAGR | 30 % |
The projected growth rate of the cobot market underscores the underlying demand for more flexible, human-friendly automation, which forms the foundation for AI-native robotics platforms. TRLC's positioning targets a subset of this growth where affordability and ease of AI integration are primary purchase criteria.
Lightly corroborated -- Market sizing is based on analogous sector reports; company-specific TAM/SAM is not publicly quantified by third-party sources.
Competitive Landscape
Sources and analysis
The Robot Learning Company enters a robotics market where the competitive pressure is defined not by a single direct clone, but by a fragmented landscape of high-cost industrial incumbents, a handful of research-focused open-source projects, and a new wave of AI-native startups targeting different layers of the stack.
A direct, named competitor selling an identical open-source AI-native robotics dev kit is not yet present in the public record. The competitive map is instead segmented by approach and customer focus. In the high-performance industrial automation segment, incumbents like ABB and Universal Robots dominate with proprietary, reliable arms priced in the tens of thousands, but their software stacks are not optimized for rapid, demonstration-based AI learning [PUBLIC]. The primary substitute for TRLC's target user,a robotics researcher or startup,has been to build a custom system by cobbling together a low-cost arm from vendors like UFACTORY or Elephant Robotics with separate cameras and open-source software frameworks like ROS and PyBullet. This approach offers flexibility but requires significant integration overhead, creating the very gap TRLC aims to fill.
TRLC's defensible edge today rests on its integrated, open-source hardware/software package and a founder-led focus on the robot learning workflow. The combination of a sub-$10,000 hardware bundle with a pre-configured SDK for imitation and offline reinforcement learning creates a low-friction entry point for experimentation [Y Combinator, June 2025] [PlatoSeed]. This edge is durable if the company can build a community around its open-source libraries and capture user-generated data and models, creating network effects within a niche developer base. However, it is perishable if larger incumbents decide to offer similar low-cost, learning-oriented bundles or if more funded AI robotics startups begin to commoditize the hardware layer to capture higher-margin software and services.
The company's most significant exposure is to well-capitalized startups operating one layer above or below them. A competitor like Covariant, which focuses on AI-powered perception and reasoning for existing industrial arms, could theoretically expand downstream to offer its own integrated hardware, leveraging its substantial funding and enterprise relationships [Crunchbase]. Similarly, a hardware-focused challenger with greater manufacturing scale could undercut TRLC on price for the physical arm component, turning its integrated kit into a less compelling bundle. TRLC also does not own a direct sales channel to large industrial customers, a gap that could be exploited by competitors with established distribution partnerships.
The most plausible 18-month competitive scenario hinges on the adoption of its open-source platform. If TRLC successfully onboards a critical mass of academic labs and early-stage startups, its SDK could become a de facto standard for prototyping AI robotics, creating a moat that attracts partnership opportunities with larger hardware manufacturers. In this scenario, the "winner" would be TRLC, solidified as the infrastructure layer for a new generation of learning-enabled robots. The "loser" would be the DIY approach using generic low-cost arms and disparate software, as the friction reduction of TRLC's kit makes the custom build less justifiable for new projects. Conversely, if adoption stalls and a competitor with similar integration but superior marketing or funding emerges, TRLC's first-mover advantage could evaporate quickly.
Lightly corroborated -- Competitive analysis is inferred from public positioning of incumbents and adjacent players; no direct, named competitors are cited in primary sources for TRLC.
Opportunity
Public sources The Robot Learning Company’s opportunity rests on a simple premise: if it can become the default, open-source infrastructure layer for AI-native robotics, the value accrual could mirror the early days of software-defined automation.
The headline opportunity is to become the foundational software and hardware platform for a new generation of affordable, general-purpose robotic workers. This is not a niche tool for research labs. The company’s stated goal is to build tools for the next era of physical automation, targeting businesses seeking low-cost automation for repetitive, stationary tasks [Y Combinator, June 2025]. The evidence that makes this outcome reachable, rather than purely aspirational, is the company’s early demonstration of a complete, working stack. They have already shown a functional dev kit that combines hardware, teleoperation, and a cross-platform software SDK [PlatoSeed]. More importantly, founder Jannik Grothusen previously demonstrated the core technical workflow by training a pair of inexpensive robot arms to clean a spill using GPT-4o in just four days [TechCrunch, Nov 2024]. This proves the underlying concept,rapid, low-cost robot training via AI,is technically feasible. The company’s current product, priced under $10,000 for a pair of arms, directly attacks the primary barrier to adoption: cost [Y Combinator]. By providing an open-source, all-in-one platform, TRLC aims to be the rails on which thousands of small automation projects are built, moving from a dev kit vendor to the de facto standard for robot learning infrastructure.
Growth could follow several distinct, concrete paths. The following scenarios outline plausible routes to scale, each grounded in the company’s current positioning and public statements.
| Scenario | What happens | Catalyst | Why it's plausible |
|---|---|---|---|
| The Research & Developer Standard | TRLC-DK1 becomes the default hardware/software stack for academic labs and AI robotics startups, creating a pipeline of trained engineers and derivative commercial products. | A major research institution (e.g., UC Berkeley, TU Munich) adopts the kit for a flagship course or lab, publishing results using the platform. | The product is explicitly marketed to robotics researchers and developers as an open-source, all-in-one kit [PlatoSeed]. Founder Grothusen has research ties to both UC Berkeley and TU Munich, providing a natural initial beachhead [Fundraising Fox, Sept 2026]. |
| The SME Automation Wedge | The company successfully converts pilot discussions into paid deployments with small-to-medium enterprises (SMEs), using a low-cost, task-specific automation offer to land and then expand within facilities. | Securing a first named, referenceable customer in a vertical like light manufacturing or lab automation, demonstrating a clear ROI on a repetitive task. | The company’s Y Combinator profile states it is seeking pilot partners for its affordable platform, indicating an active commercial push beyond pure research [Y Combinator]. The hardware is priced for SME budgets, and the focus on stationary, repetitive tasks aligns with common SME pain points. |
| The Embedded Infrastructure Play | TRLC’s software stack (the robotics-specific MLOps layer) is adopted by larger robotics OEMs or system integrators as the preferred middleware for AI-driven automation, decoupling from their own hardware sales. | A partnership announcement with a established robotics hardware manufacturer to bundle or recommend TRLC’s software platform. | Third-party analysis already frames the product as a robotics-specific MLOps layer, a description that implies a standalone software value proposition [Altis Research]. An open-source core lowers adoption barriers for partners. |
Compounding for TRLC would manifest as a classic adoption flywheel, powered by open-source distribution and data network effects. Early adoption by developers and researchers generates a growing corpus of task demonstrations, training code, and model checkpoints shared within the TRLC ecosystem. This public repository of “robot skills” makes the platform more valuable for the next user, who can build upon existing work rather than start from scratch. The company’s GitHub, which already hosts forks of key perception libraries like SAM2 and LeRobot, is the early nucleus of this flywheel [PlatoSeed]. As the library of shared skills grows, TRLC’s software stack becomes the path of least resistance for implementing new automation, creating a subtle but powerful form of distribution lock-in. Furthermore, every deployment generates unique physical data that can be used to improve the robustness of the underlying imitation and reinforcement learning models, potentially creating a data moat for common industrial tasks over time. The flywheel is in its earliest stage, but the open-source foundation is explicitly designed to trigger it.
The size of the win, should the company capture a leading position in this nascent infrastructure layer, can be framed by looking at comparable companies that defined software layers in other hardware-centric domains. A relevant, though not perfect, analog is the trajectory of companies like NVIDIA with its CUDA platform for GPU computing. While TRLC operates at a vastly smaller scale and in a different market, the principle of building the essential software layer for a new computing paradigm is similar. More directly, the market for industrial robot software and services is projected to reach tens of billions of dollars globally within the decade. If TRLC executes on the “Embedded Infrastructure” scenario and captures even a single-digit percentage of that software value pool, the outcome would be a company valued in the hundreds of millions to low billions of dollars. This is a scenario-based outcome, not a forecast, but it illustrates the magnitude of the prize for the company that successfully standardizes the robot learning stack.
Lightly corroborated -- The opportunity analysis is based on the company's stated goals and product capabilities from its website and YC profile, as well as the founder's prior demonstrated work. The growth scenarios are plausible extrapolations but lack confirmation from commercial customer announcements or partnership deals.
Sources
Public sources
[Company site, July 2026] The Robot Learning Company | https://www.robot-learning.co/
[Y Combinator, June 2025] The Robot Learning Company: Tools for the next era of physical automation. | https://www.ycombinator.com/companies/the-robot-learning-company
[PitchBook, Feb 2025] The Robot Learning Company 2025 Profile: Valuation, Funding & Investors | https://pitchbook.com/profiles/company/840193-21
[Altis Research] The Robot Learning Company Funding | Complete Analysis | https://www.extruct.ai/hub/robot-learning-co/index.html
[LinkedIn] The Robot Learning Company | https://www.linkedin.com/company/the-robot-learning-company
[PlatoSeed] TRLC-DK1: The Robot Learning Company | https://www.robot-learning.co/
[TechCrunch, Nov 2024] Built in four days, this $120 robot arm cleans a spill with help from GPT-4o | https://techcrunch.com/2024/11/04/built-in-four-days-this-120-robot-arm-cleans-a-spill-with-help-from-gpt-4o/
[Fundraising Fox, Sept 2026] Jannik Grothusen , Founder at The Robot Learning Company | https://www.fundraisingfox.com/jannik-grothusen
[GitHub, 2026] robot-learning-co/trlc-dk1 | https://github.com/robot-learning-co/trlc-dk1
[SVRC, 2026] TRLC-DK1: The Robot Learning Company | https://svrc.ai/trlc-dk1
[LinkedIn, 2026] Jannik Grothusen - The Robot Learning Company | https://de.linkedin.com/in/jannik-grothusen/en
[LinkedIn, 2026] Cristian Pinto - pluto house | https://www.linkedin.com/in/cristian-pinto-marinho/
[LinkedIn, 2026] Clément Grivel - Co-founder & CTO @ Condor Energy (YC ... | https://fr.linkedin.com/in/clement-grivel
[LinkedIn, 2026] Rasmus Soeborg - Thor Commerce | https://www.linkedin.com/in/rasmus-soeborg-4ba65a1b6/
[LinkedIn, 2026] Daniel Friis - Stealth Physical AI | https://www.linkedin.com/in/dfriis/
[Crunchbase] The Robot Learning Company - Crunchbase Company Profile & Funding | https://www.crunchbase.com/organization/the-robot-learning-company
Articles about The Robot Learning Company
- The Robot Learning Company's $10,000 Kit Aims to Wire AI into the Roboticist's Workbench — A Y Combinator-backed team is betting that open-source hardware and a unified software stack can accelerate the move from robot research to affordable automation.