Neura Robotics, a German-founded company with a €1B+ valuation, is solving the biggest bottleneck in robotics: how to teach robots new skills faster.
The industry standard? A robot needs 2,000 hours of physical practice — repeated motions over 200 days — to learn to fold a towel or tighten a screw.
Neura’s solution? Zero training on the robot itself.
Instead, skills are learned once — in real-world environments — then uploaded to a cloud platform called Neuraverse. Any compatible robot, anywhere in the world, can download and use that skill in under two hours.
This isn’t simulation. It’s real physics — captured in 3D, matched with video and text — then shared like an app.
🔧 The Problem: Every Robot Learns Alone. That’s Too Slow.
Today’s robots are trained in isolation.
A robot in Germany learns to pick up a tool.
A robot in China learns the same thing — from scratch.
Same task. Same motion. Same hardware.
But 2,000 hours of training… twice.
This “data silo” problem is why robotics adoption is slow, expensive, and limited to narrow use cases.
Neura’s insight:
If you can capture a skill once — in a real factory, home, or hospital — why make every robot relearn it?
Their answer: Neuraverse — a global, cloud-based operating system for robot skills.
Think of it like the App Store… for robot movements.
🚀 How It Works: The “Robot Gym” That Teaches 1,000 Robots at Once
Neura built NeuraGym — physical training centers with real-world environments:
- A kitchen (fold towels, load dishwashers)
- A factory floor (tighten screws, handle tools)
- A clinic (assist with mobility, transfer patients)
In each room, a robot performs a task under human supervision.
Sensors capture real 3D motion, force feedback, visual context, and audio cues.
That data is fused with video and textual instructions.
Then, it’s turned into a reusable AI model — and uploaded to Neuraverse.
Result?
A robot in Shanghai can download a “fold laundry” skill trained in Berlin — and use it immediately.
2,000 hours of training → 2 hours of deployment.
No retraining. No manual coding. No engineers needed.
🌍 Why China? Supply Chain + Scale + Data
Neura’s €1.35 billion investment in Hangzhou isn’t just about manufacturing.
It’s about accessing China’s unmatched ecosystem:
- Hardware supply chain: High-precision motors, sensors, and lead screws — cheaper, faster, more mature than in Europe
- Industrial scale: 100+ Chinese robotics firms can now plug into Neuraverse — accelerating their own R&D
- Real-world data diversity: China’s vast logistics, manufacturing, and elder care sectors offer unique, high-volume training scenarios
Neura will open its second NeuraGym in China — larger than the one in Germany — to capture local data, train models, and feed them back into the global platform.
“China isn’t just a market,” says David Reger. “It’s the best place to build the world’s most diverse robot training dataset.”
💰 The Numbers: Orders, Valuation, and Strategic Moves
- €1.2B B-round (2025): Led by Lingotto (Agnelli family), Volvo, Delta Electronics
- Valuation: ~$1B
- Backlog: >€1B in confirmed orders
- Acquisition: ek robotics (German AMR/logistics player, €60M 2024 revenue) — added autonomous mobility to Neura’s stack
- Production: 3 black-factory facilities in Germany — currently 15,000 units/year, targeting seconds-per-unit output
- China hub: 1.35B € investment for R&D, manufacturing, and NeuraGym expansion
Neura does not sell robots.
It sells skills as a service — delivered via Neuraverse.
Manufacturers buy the hardware.
Neura delivers the intelligence.
Neura Robotics – Funding History
| Funding Round | Date | Amount Raised | Investors |
|---|---|---|---|
| Series B | 2025-01-17 | €120 million | Led by Lingotto HV Capital C4 Ventures V-Squared Ventures Volvo Cars Tech Fund L-Bank Delta Electronics InterAlpen Partners |
| Series A+ | 2023-10-10 | $16 million | Led by InterAlpen Partners |
| Series A | 2023-07-18 | €50 million | Led by Lingotto HV Capital V-Squared Ventures Primepulse |
| Seed Round | 2019-05-01 | Undisclosed | Undisclosed |
| Angel Round | 2021-06-01 | Undisclosed | Undisclosed |
Data Source: RuiShou Analytics
🤖 Why This Isn’t Just Another Robot Company
Neura’s business model is two-sided:
| Side | What it does | Why it matters |
|---|---|---|
| Hardware Partners | Use Neura’s chassis, sensors, and control systems | Faster time-to-market, lower R&D cost |
| Platform Users | Download pre-trained skills from Neuraverse | Skip 99% of training time; deploy in days |
This turns Neura into a platform play — not a hardware vendor.
- Volvo uses NeuraGym-trained skills to automate warehouse logistics
- Kawasaki Heavy Industries integrates Neuraverse skills into its factory robots
- A Chinese logistics startup downloads “pack box A, label B” from Neuraverse — no engineers needed
The moat? Network effects.
More robots → more data → better skills → more users → more data.
It’s a flywheel — and Neura built the first one.
⚡ The Real Edge: Platform > Production
Neura is building 15,000 robots per year — but CEO David Reger spends 80% of his time on Neuraverse.
Why?
Because scale is easy. Intelligence is hard.
- Hardware: Can be manufactured in China, outsourced, scaled.
- Skills: Can only be built through real-world data — and shared across users.
Neura’s goal:
Become the Android of robot intelligence — not the iPhone.
The company doesn’t need to make every robot.
It needs to make every robot smarter — by giving them access to a shared brain.
🌐 The Bigger Picture: A New Model for Global Tech Transfer
Neura’s China launch isn’t just about market access.
It’s about cross-border innovation infrastructure.
- European-trained skills → trained on Chinese data → deployed globally
- Chinese startups use Neuraverse to build export-ready products — without reinventing the wheel
- European investors fund Chinese robotics teams via Neura’s platform
This is not a typical “China manufacturing, Europe IP” story.
It’s a two-way, data-driven ecosystem — where skills flow freely, and value is created collectively.
Investors are watching.
Institutional capital from Europe and China co-attended the Hangzhou launch.
This isn’t just a startup. It’s a new model for global robotics collaboration.
🎯 Investment Takeaway: The Next Winner Won’t Build More Robots. It Will Teach Them.
What’s changing in robotics?
From hardware competition → to intelligence sharing.
Neura’s edge:
- Built a real-world skill marketplace (Neuraverse)
- Solved the 2,000-hour training problem
- Turned data into a recurring revenue stream
- Leveraged China’s supply chain to scale hardware — while keeping IP and platform in its control
Investors should ask:
- Is the platform open to third-party developers? → Yes
- Can skills be monetized? → Yes — per download, per license
- Is this defensible? → Yes — network effects grow exponentially
- Is this scalable beyond factories? → Yes — elder care, retail, logistics, homes
The future of robotics isn’t in motors or sensors.
It’s in the cloud.
And Neura owns the operating system.
Major Humanoid Robotics Companies in Europe
| PAL Robotics | Barcelona, Spain | 2004 | One of Europe’s earliest full-body humanoid robot developers. Products: “TALOS”, “TIAGo” — capable of bipedal walking and dual-arm manipulation. |
| Neura Robotics | Germany | 2019 | Focused on cognitive humanoid robots. Next-generation platform: “4NE-1”. |
| 1X Technologies | Norway / USA | 2014 | Building home-assistant humanoid robots (or near-humanoid form). Product: “NEO Gamma”. |
| Furhat Robotics | Sweden | 2014 | Social humanoid robots focused on head-and-face interaction. |
| Humanoid | United Kingdom | 2024 | Goal: A commercially scalable, safe humanoid robot platform for industrial use. |
