# Flexion Robotics

> Source: https://aiwiki.ai/wiki/flexion_robotics
> Updated: 2026-07-23
> Categories: AI Companies, Embodied AI, Humanoid Robots, Robotics Companies
> License: CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/)
> From AI Wiki (https://aiwiki.ai), the free encyclopedia of artificial intelligence. Reuse freely with attribution to "AI Wiki (aiwiki.ai)".

| Field | Value |
| --- | --- |
| Legal name | Flexion Robotics AG |
| Trading name | Flexion |
| Industry | Robotics software, [physical AI](/wiki/physical_ai) |
| Founded | December 2024 (commercial register entry December 9, 2024) |
| Founders | Nikita Rudin, David Höller, Julian Nubert, Fabian Tischhauser, Marco Hutter |
| Key people | Nikita Rudin (CEO), David Höller (CTO) |
| Headquarters | Affolternstrasse 42, 8050 Zurich, Switzerland |
| US office | San Francisco, California |
| Product | Hardware-agnostic autonomy stack for humanoid robots; Reflect v1.0 platform |
| Total funding | $57.35 million (seed plus Series A, as of November 2025) |
| Investors | DST Global Partners, NVentures, redalpine, Prosus Ventures, Moonfire, Frst |
| Website | flexion.ai |

**Flexion Robotics AG**, usually shortened to **Flexion**, is a Swiss robotics software company that builds a hardware-agnostic autonomy stack, often described as "the brain," for [humanoid robots](/wiki/humanoid_robot).[1][2] The company trains its control and motion systems primarily with [reinforcement learning](/wiki/reinforcement_learning) in large-scale physics simulation rather than with fleets of human teleoperators, and licenses the resulting software to robot manufacturers instead of building robots of its own.[2][3] Founded in Zurich in December 2024 by researchers from [ETH Zurich](/wiki/eth_zurich) and [NVIDIA](/wiki/nvidia), Flexion raised a $7.35 million seed round and a $50 million Series A within its first year and, in July 2026, demonstrated with [Niantic Spatial](/wiki/niantic_spatial) and NVIDIA a zero-shot real2sim2real navigation pipeline on a physical humanoid.[1][4][5]

## Background and founding

Flexion Robotics AG was entered in the commercial register of the canton of Zurich on December 9, 2024, with the stated purpose of developing, producing, distributing, and maintaining software for controlling humanoid robots.[6] The company operated quietly through most of 2025 and came out of stealth on November 20, 2025, when it announced its Series A round.[3][7]

The founding team came out of ETH Zurich's Robotic Systems Lab and NVIDIA's robotics research groups. CEO Nikita Rudin and CTO David Höller worked on GPU-accelerated robot learning at NVIDIA; Prosus Ventures, an investor, credits the pair with pioneering simulation-based robot training at ETH Zurich starting in 2022 and later building [Isaac Lab](/wiki/isaac_lab), NVIDIA's open-source robot learning framework.[8] Humanoids Daily reported that Höller was a research manager at NVIDIA who helped build Isaac Gym, Isaac Lab's predecessor, and that Rudin came from the ETH lab known for the ANYmal quadruped.[9] The other co-founders are Julian Nubert, who leads perception, and Fabian Tischhauser, who leads robotics hardware; Marco Hutter, professor at ETH Zurich and head of the Robotic Systems Lab, is listed by the company as a co-founder and advisor.[10] Beyond the founders, Flexion says its early team drew people from ETH Zurich, NVIDIA, Meta, Google, [Tesla](/wiki/tesla), and Amazon, with backgrounds in reinforcement learning, control systems, perception, and mechatronics.[2][3]

Crunchbase News counted 31 employees at the time of the Series A in November 2025; by mid-2026 the company's own site described a team of more than 60, including former researchers from Meta Reality Labs, NASA JPL, and the [Toyota Research Institute](/wiki/toyota_research_institute).[7][10]

## Technology

Flexion positions itself one layer below the robot makers: it does not build bodies, only the software that runs them. The company summarizes this as "We're not building the body. We're building the brain."[2] The stack is designed to be hardware agnostic, with abstracted interfaces so it can be ported across humanoid platforms from different manufacturers.[2][11]

As described at the Series A announcement, the stack has three layers:[2][12]

| Layer | Function |
| --- | --- |
| Command layer | A [large language model](/wiki/large_language_model) interprets natural-language tasks, decomposes them into steps, and reasons about the environment |
| Motion layer | A [vision-language-action model](/wiki/vision_language_action_model) trained mostly on synthetic data, refined with real-world data, turns perception into motion goals |
| Control layer | Transformer-based whole-body control with a modular skill library executes balance, [locomotion](/wiki/robot_locomotion), and manipulation in real time |

The Robot Report noted that Flexion deliberately avoids "end-to-end monoliths," arguing that modularity keeps interfaces clean and testable and improves generalization.[12] The training approach is simulation first: instead of collecting demonstrations through teleoperation, Flexion generates [synthetic data](/wiki/synthetic_data) in massively parallel physics simulation and relies on [sim-to-real transfer](/wiki/sim_to_real_transfer) to move policies onto hardware, an approach the founders helped establish through Isaac Gym and Isaac Lab.[8][12] The company's launch material put it bluntly: "No scripts, no tele-op farms, no brittle task-specific logic."[13]

Flexion's public demonstrations have run on hardware from [Unitree](/wiki/unitree). Its debut video, released with the Series A in November 2025, showed a [Unitree G1](/wiki/unitree_g1) walking through uneven forest terrain in the Swiss Alps, bending down to pick up trash, and dropping it in a bin; Humanoids Daily reported the robot carried a custom perception setup with a ZED stereo camera and [NVIDIA Jetson](/wiki/nvidia_jetson) Orin compute.[9] The Robot Report likewise pictured a Unitree humanoid running Flexion's software.[12]

## Funding

Flexion raised $57.35 million across two rounds in roughly one year, an unusually fast [ramp](/wiki/ramp) for a European [robotics](/wiki/robotics) software startup.[7]

| Round | Announced | Amount | Investors |
| --- | --- | --- | --- |
| Seed | Closed in 2025, disclosed November 2025 | $7.35 million | Frst, Moonfire, redalpine |
| Series A | November 20, 2025 | $50 million | DST Global Partners (lead), NVentures, redalpine, Prosus Ventures, Moonfire |

The Series A was led by DST Global Partners with participation from [NVentures](/wiki/nvidia_nventures), NVIDIA's [venture capital](/wiki/venture_capital) arm, along with redalpine, Prosus Ventures, and Moonfire.[3][7][12][18] European outlets reported the round as roughly EUR 43 million to EUR 50 million depending on conversion.[4][14] No valuation was disclosed.[7] The company said the money would go toward growing the Zurich R&D team, scaling compute and robot fleets, opening a US presence, and commercializing with what it called major OEM partners; it has since listed a San Francisco office at 1004 Treat Avenue alongside its Zurich headquarters.[2][3] Crunchbase News reported that Flexion plans to charge manufacturers an annual per-robot software license.[7]

Investors framed the bet around training data. Prosus Ventures argued that the binding constraint on humanoids is scalable, high-quality training data, and that Flexion's simulation-first reinforcement learning lets robots accumulate thousands of hours of virtual experience per day, more than teleoperation pipelines can produce.[8] Startupticker summarized the target markets as industrial settings, logistics, manufacturing, disaster response, and planetary exploration.[3]

## Reflect v1.0

On June 29, 2026, Flexion released Reflect v1.0, which it calls a robotics intelligence platform for "long-horizon" autonomous humanoid work.[11][15][17] The release video shows a modified Unitree humanoid receiving a single natural-language instruction,[19] then autonomously retrieving a delivered snack parcel from a ground-floor delivery area, opening doors, taking stairs and an elevator, unpacking the box, and placing the items in a designated drawer, with no human operator.[11][16]

Architecturally, Reflect v1.0 puts a custom [vision-language model](/wiki/vision_language_model) at the top as a mission controller that watches the robot's camera feed and continuously replans. Below it, a vision-language-action model trained on real-world data works together with reinforcement-learning skills and a whole-body controller, on top of a runtime that handles communication, process isolation, low-latency inference, logging, and safety checks.[11][15] On an internal 16-step mission evaluation, Flexion reported that supervised fine-tuning alone completed 38 percent of missions end to end, and that reinforcement-learning fine-tuning raised this to 90 percent; eWeek noted these are company benchmarks without independent validation.[15][16]

Flexion was unusually direct about limits: the platform operates within bounded task distributions rather than being a general-purpose worker, some objects remain hard to grasp, the mission controller can make wrong visual assumptions, and recovery behaviors cover some but not all failure modes.[11]

## Real2sim2real demonstration with Niantic Spatial and NVIDIA

On July 20, 2026, Niantic Spatial and Flexion published a joint technical post, also carried on Flexion's site, showing what they described as an end-to-end real2sim2real pipeline: an RGB-only navigation policy trained entirely in simulation that transferred zero-shot to a real humanoid robot navigating a real office.[1][5] NVIDIA's [Isaac Sim](/wiki/nvidia_isaac_sim) and Isaac Lab frameworks provided the simulation and training infrastructure.[5]

The pipeline starts with a single walkthrough of the deployment site using an off-the-shelf 360-degree RGB camera. Niantic Spatial reconstructs the scene as a [digital twin](/wiki/digital_twin): a 3D Gaussian splat (see [Gaussian splatting](/wiki/gaussian_splatting)) supplies photorealistic RGB rendering, while MVSAnywhere, a zero-shot multi-view stereo model, generates an aligned collision mesh that holds up even on low-texture surfaces. Both are packaged as a USDZ file in NVIDIA's NuRec volume format and loaded directly into Isaac Sim and Isaac Lab, where Flexion trains navigation policies with massively parallel reinforcement learning, domain randomization, and image encoders trained offline that run identically in training and on the robot.[1][5] Niantic Spatial says a five-minute 360-degree capture in its Scaniverse app can become a simulation-ready environment this way.[1] The division of labor: Niantic Spatial handles "faithfully bringing reality into the simulator," Flexion delivers "policies and deployment software tuned to the specific hardware."[1]

The teams benchmarked four policy variants, each evaluated over 1,024 rollouts with identical spawn and target poses, in two reconstructed offices:[5]

| Policy variant | Sensor | Training environment | Flexion office | Niantic Spatial office |
| --- | --- | --- | --- | --- |
| Baseline | Depth (ZED X, neural mode) | Untextured navigation mesh | 93.8% | 70.9% |
| RGB, generated | RGB | Untextured navigation mesh | below depth baseline | below depth baseline |
| RGB, synthetic | RGB | Synthetic textured office | below depth baseline | below depth baseline |
| RGB, reconstruction | RGB | Gaussian splat of the actual site | 97.8% | 75.0% |

Only the RGB policy trained inside the Gaussian-splat reconstruction beat the depth baseline; RGB policies trained on generic or synthetic environments did worse than depth in both offices.[5] The post highlights three failure modes where RGB beat depth sensing: semantic hazards such as a blue mat that is obvious in color but nearly invisible in a depth stream, thin structures like tripods, railings, and cables that stereo depth blurs or misses, and transparent surfaces such as glass doors and windows.[5] The authors argue the result matters commercially because deploying a policy to a new site has traditionally taken months of on-site adaptation, while this pipeline compresses site capture to a single walkthrough and training to simulation time.[5] The blog does not name the humanoid platform used in the demonstration.[1][5]

Stated limitations include that a reconstruction captures a single moment in time with lighting and reflections baked in, and that scenes are static, with dynamic agents left to future work. The teams listed follow-ups including capture from iPhones and fisheye cameras, open-vocabulary semantic labels, scene transformations such as lighting variants and moved furniture, and extending from local navigation to full-task autonomy with language-conditioned behavior.[5]

## Position in the humanoid market

Flexion is a software-only bet in a market where most prominent humanoid companies, and several [robot foundation model](/wiki/robot_foundation_model) startups, pursue vertical integration or general-purpose models trained heavily on teleoperated demonstrations. Humanoids Daily described Flexion's approach as a horizontal software layer, "the Android of humanoids," that hardware manufacturers can license instead of funding their own autonomy R&D.[9] The company's pitch leans on the founders' [robot learning](/wiki/robot_learning) pedigree: the same simulation-based reinforcement learning line of work, from ETH Zurich's legged robots through Isaac Gym and Isaac Lab, that much of the [embodied AI](/wiki/embodied_ai) field now builds on.[8][9] NVIDIA sits on both sides of the relationship, as an investor through NVentures and as the supplier of the Isaac simulation stack Flexion trains in.[5][7]

Whether the model works commercially is still open. Flexion's published results, the 16-step mission evaluation and the office navigation benchmarks, are company-run evaluations, and its OEM partners have not been named publicly.[12][16] The demonstrations so far run on Unitree hardware, and the hardware-agnostic claim, portability of the same stack across many manufacturers' robots, has not yet been shown in public across multiple platforms.[9][12]

## See also

- [Humanoid robot](/wiki/humanoid_robot)
- [Niantic Spatial](/wiki/niantic_spatial)
- [Isaac Lab](/wiki/isaac_lab)
- [Sim-to-real transfer](/wiki/sim_to_real_transfer)
- [Reinforcement learning](/wiki/reinforcement_learning)
- [Unitree G1](/wiki/unitree_g1)
- [Gaussian splatting](/wiki/gaussian_splatting)
- [Physical AI](/wiki/physical_ai)

## References

1. Closing the Sim2Real Gap for Humanoids. Niantic Spatial, July 20, 2026. https://www.nianticspatial.com/blog/flexion-humanoid-real2sim-sim2real
2. Flexion Raises $50M to Build the Brain of Humanoid Robots at Scale. Flexion, November 20, 2025. https://flexion.ai/news/flexion-raises-50m-to-build-the-brain-of-humanoid-robots-at-scale
3. Flexion raises $50M to build the brain of humanoid robots. Startupticker.ch, November 2025. https://www.startupticker.ch/en/news/flexion-raises-50m-to-build-the-brain-of-humanoid-robots
4. Zurich's Flexion raises 43 million euros to build the brains behind humanoids. EU-Startups, November 2025. https://www.eu-startups.com/2025/11/zurichs-flexion-raises-e50-million-to-build-the-brains-behind-humanoids/
5. Niantic Spatial, Flexion, and NVIDIA: Closing the Sim2Real Gap for Humanoids. Flexion, July 21, 2026. https://flexion.ai/news/niantic-spatial-flexion-and-nvidia-closing-the-sim2real-gap-for-humanoids
6. Flexion Robotics AG in Zurich. Moneyhouse (Swiss commercial register data). https://www.moneyhouse.ch/en/company/flexion-robotics-ag-11605579991
7. Exclusive: Founded By Ex-Nvidia Researchers, Flexion Lands $50M To Build The 'Brain' for Humanoid Robots. Crunchbase News, November 20, 2025. https://news.crunchbase.com/venture/robotic-brain-building-startup-flexion-raise/
8. Prosus Ventures Invests in Flexion Robotics to Build the Intelligence Stack for Humanoid Robots. Prosus, November 26, 2025. https://www.prosus.com/news-insights/2025/prosus-ventures-invests-in-flexion-robotics-to-build-the-intelligence-stack-for-humanoid-robots
9. A Brain in the Alps: Flexion Raises $50M to Build the "Android" of Humanoids. Humanoids Daily, November 2025. https://www.humanoidsdaily.com/news/a-brain-in-the-alps-flexion-raises-50m-to-build-the-android-of-humanoids
10. About. Flexion Robotics. https://flexion.ai/about
11. Flexion Reflect v1.0 - The Path Towards Long-Horizon Autonomous Humanoid Work. Flexion, June 29, 2026. https://flexion.ai/news/flexion-reflect-v1.0
12. Flexion to use Series A to build sim-to-real, AI systems powering humanoids. The Robot Report, November 26, 2025. https://www.therobotreport.com/flexion-raises-50m-build-ai-systems-power-humanoids/
13. Video: New brain helps humanoid robot handle uneven surfaces with ease. Interesting Engineering, November 21, 2025. https://interestingengineering.com/ai-robotics/brain-helps-humanoid-robot-conduct-tasks
14. Zurich's Flexion raises 50 million euros to build the brains behind humanoids. BeBeez International, November 21, 2025. https://bebeez.eu/2025/11/21/zurichs-flexion-raises-e50-million-to-build-the-brains-behind-humanoids/
15. Flexion new AI model gives humanoid robots long-horizon autonomy. Interesting Engineering, June 2026. https://interestingengineering.com/ai-robotics/video-new-ai-model-gives-humanoid-robots-90-percent-success-in-complex-missions
16. Flexion Reflect v1.0 Shows Why Humanoid Robot Software Matters. eWeek, June 29, 2026. https://www.eweek.com/news/flexion-reflect-robot-software/
17. Swiss Startup Flexion Robotics Introduces 'Long-Horizon' Autonomous Humanoid Robotics Platform. The AI Insider, June 29, 2026. https://theaiinsider.tech/2026/06/29/swiss-startup-flexion-robotics-introduces-long-horizon-autonomous-humanoid-robotics-platform/
18. Robotics Software Startup Flexion Raises $50m Series A To Power Humanoid Autonomy. Crowdfund Insider, November 2025. https://www.crowdfundinsider.com/2025/11/255869-robotics-software-startup-flexion-raises-50m-series-a-to-power-humanoid-autonomy/
19. Reflect v1.0 - The Path Towards Long-Horizon Autonomous Humanoid Work (video). Flexion on YouTube, June 2026. https://www.youtube.com/watch?v=6dme3JYj3Hg