FUNDING DESK · Germany · Physical AI / industrial robotics and embodied intelligence
Agile Robots
Agile Robots is a physical-AI robotics company founded by researchers from the German Aerospace Center (DLR), with a portfolio spanning the Agile ONE industrial humanoid, the Diana 7 force-controlled arm, the Thor series, the Agile Hand, mobile robotics, and the AgileCore software platform.
Company website ↗Funding facts
| This round | Round | Date | Funding data attribution |
|---|---|---|---|
| €451.22M | Series D+ · Equity-stage round | 2026-11-12 · Funding record date | Tech.eu Funding Explorer ↗ |
| Undisclosed | Round unspecified · Type unspecified | 2026-10-20 · Funding record date | Tech.eu Funding Explorer ↗ |
First-party material retrieved
Last retrieval:2026-11-12
Evidence gaps:Pricing · Customer cases · Technical documentation
A newer round is recorded. The product analysis below still refers to the previous round and awaits an update.
Rigid fixed-function industrial automation cells that struggle with high-mix, low-volume and frequent line changeovers, plus manual work in delicate grasping, machine tending and inter-station transport.
The key question is whether it can turn a bundle of humanoid, force-controlled arms, dexterous hands and software into repeatable, standardized product revenue rather than bespoke integration for a few large automotive and electronics customers; public evidence supports product existence, not unit economics or repeat purchase.
The following is editorial analysis based on public material. Inferences and open questions are labeled in the text. Funding is not evidence of revenue or product-market fit.
Product evidence checked 2026-10-21
01
What the product does
Company-reported product scope includes Agile ONE, an industrial humanoid with dexterous hands, camera/LiDAR/speech perception, autonomous movement and human collaboration; Diana 7, a seven-axis arm with torque sensors on every axis; the Thor series with 3, 7, 12 and 20 kg payload options; the five-finger modular Agile Hand; AMR solutions developed with BÄR Automation and idealworks; and AgileCore, a software platform for integrators and operators, with AgileAI described as a natural-language-programmable brain. Precision, cycle time, reliability and safety-certification figures are not given on the retrieved pages.
02
Users, buyers and demand
The site names automotive, electronics, healthcare and service industries as its most demanding customers and claims to serve leading companies among them, without naming accounts, contract sizes or deployment counts. Two buyer groups are implied: plant engineers and system integrators who must fit robots into existing stations, and manufacturing or logistics operations leaders paying for flexible changeover, labor shortages and delicate-manipulation gaps. Pain frequency depends on changeover rates and task complexity, which public material cannot quantify.
03
The actual workflow
A typical flow: a customer defines a station task; Agile Robots or an integration partner configures behavior through AgileCore/AgileAI, selecting Diana 7, Thor or the Agile ONE humanoid, paired with Agile Hand for grasping and assembly; Agile ONE moves between workstations using cameras, LiDAR and speech recognition and signals status via a chest display and lights. The site says its robot foundation models are trained on real production data, simulated datasets, and teleoperation and human data from data farms, but data volume, training cycles and on-site commissioning time are undisclosed.
04
Pricing and unit economics
No pricing, subscription or licensing terms appear on the retrieved pages, so the charging model is undisclosed. A conditional inference is that revenue may combine hardware sales or leasing, AgileCore software licensing and engineering services around deployment, but this is inference, not verified evidence. Unit economics — per-unit gross margin, deployment labor cost, software renewal — are entirely missing and cannot be inferred from the funding record or product pages.
05
Adoption evidence and gaps
Confirmable adoption signals are company-reported only: founded in 2018 by DLR robotics researchers, 2,000+ R&D experts, products aimed at leading automotive, electronics, healthcare and service customers, a physical-AI manufacturing partnership with Hitachi, and an investment in India's XNG Automation. These are unaudited. Customer case pages, deployed unit counts, revenue, repeat purchase and retention data are absent, so partnership announcements should not be read as demand validation.
06
Competition and defensibility
The moat conditions are threefold: DLR-rooted engineering in force control and dexterous manipulation; the ability to integrate hardware, AgileCore software and foundation models into one system; and a training flywheel built from large-customer field data. Versus general-purpose models, the difference is embodied hardware and a real-time control loop; versus legacy automation lines, it is flexible changeover and natural-language programming. Whether these hold depends on standardized delivery and a genuinely closed data loop, which public evidence cannot yet establish.
07
How to read this round
The record shows a funding round dated 2026-10-20 with Nvidia among investors, labeled Other, with amount and original currency undisclosed, so valuation, dilution and lead investor cannot be determined — an investor list is not a lead-investor list. The round signals industrial-capital interest in physical AI but is not evidence of revenue or product success. Use of proceeds is undisclosed; the next milestone should be verifiable standardized delivery and repeat orders, not more partnership announcements.
08
Where it could fail
First, if humanoids remain demos and pilots that miss line requirements on cycle time, yield and safety certification, revenue stays dependent on bespoke integration and does not scale. Second, if AgileCore and foundation models still require heavy on-site engineering, deployment cost erodes gross margin and changeover savings fail to justify a premium. Third, customer concentration and capex cycles in automotive and electronics amplify order risk. Fourth, ecosystem ties to an industrial investor such as Nvidia may constrain neutrality across other compute and software stacks.
09
What you can take from it
A transferable lesson for peers: enter production lines first with separately sellable components such as force-controlled arms and dexterous hands, then connect them through a software platform and data, and only then push a full humanoid. This component-to-system path yields real operating data more readily than selling a humanoid outright. Publish verifiable deployment metrics early rather than only partnership news.
10
What to watch next
Watch for disclosed pricing or leasing terms, named customer cases and deployed unit counts, hard specs for Agile ONE such as cycle time, payload and safety certification, open documentation and a developer ecosystem for AgileCore, and any supplemental disclosure of this round's amount and use of proceeds.