Breaking

Ex-NVIDIA Engineers' Flexion Unveils Humanoid Autonomy Software

June 29, 2026 at 04:03 EDT

Robotics
AI Agents
Funding & M&A

Zurich-based startup Flexion Robotics, built by a team led by former NVIDIA engineers, has released autonomy software for humanoid robots and demonstrated it carrying out office chores from natural-language commands alone. Founded in January 2025, the company was co-founded by CEO Nikita Rudin and CTO David Hoeller, who honed robot reinforcement-learning research at ETH Zurich and NVIDIA.

November 2025 · Flexion Robotics

Teaching humanoids to work from a single sentence

A Zurich startup founded by ex-NVIDIA engineers built a hardware-agnostic autonomy stack. Told only "retrieve the parcel, take the stairs, come up by elevator, then stock the shelf," its robot climbed, walked, grasped and unpacked — fully autonomously, as one continuous workflow.

$57.35M

Total funding raised ($50M Series A + $7.35M seed)

300

Fully autonomous tasks demonstrated at ICRA 2026

2025

Founded in January, Zurich-based with an SF office

From seed to Series A

Series A is roughly 7× the seed round — column height proportional to capital.

$7.35M

Seed

$50M

Series A (Nov 2025)

One command → end-to-end action

A hierarchical AI converts plain English into whole-body control — locomotion and manipulation in a single flow.

Task reasoning

Understands the natural-language goal

→

VLA model

Vision-language-action planning

→

Whole-body control

Transformer drives low-level motion

The bet: solve the data bottleneck in simulation

High-quality training data is the widely cited blocker for humanoid adoption. Flexion accumulates the equivalent of thousands of hours of experience per day via simulation-centered reinforcement learning — building a hardware-agnostic intelligence layer for a market dominated by hardware makers.

Why backers are bullish

Scalable experience in simulation, not the lab
Intelligence layer seen as essential to a viable humanoid market
Hardware-agnostic — works across robot makers

Open questions

Real-world robustness of full autonomy needs more validation
Depends on close collaboration with hardware partners
Intense competition with U.S. and Chinese players

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.