Meet Eno, the anti-humanoid robot

Good morning, robotics enthusiasts. Genesis AI’s Eno robot skips the humanoid cosplay entirely, rolling in on wheels and folding down to suitcase size when the shift’s over.

Backed by Eric Schmidt and a $105M seed round, the French startup is betting that dexterous hands, a shape-shifting body, and a foundation model called GENE can beat a pair of legs and a face.

In today’s robotics rundown:

• Meet Eno, the anti-humanoid on wheels

• Mobileye is building its own robotaxi fleet

• Inside Morph’s pitch for octopus-inspired physical AI

• XDOF nabs $70M to solve robotics’ data problem

• Quick hits on other robotics news

LATEST DEVELOPMENTS

GENESIS AI

🤖 Meet Eno, the anti-humanoid on wheels 

The Rundown: French startup Genesis AI — backed by Eric Schmidt and sitting on a $105M seed round — just unveiled Eno, a wheeled, foldable robot that skips the humanoid template entirely and bets industrial customers want function over a face.

The details:

• Eno runs on a wheeled base topped by an adjustable tower of articulated panels; it can raise its height in real time, then collapse to the size of a suitcase.

• Hands do the humanoid work instead: 20 active, back-drivable degrees of freedom, sized one-to-one to match human hands.

• GENE, Genesis’s foundation model, runs Eno as a “true physical agent” that reasons, adapts, and owns outcomes beyond pre-defined tasks.

• Dozens of units are built already; production scales in later this year, landing first in logistics and manufacturing, then hotels, hospitals, and consumers.

Why it matters: Tesla and Figure are racing to build something that looks like a coworker; Genesis is betting customers just want the job done. With Eric Schmidt’s backing and human-grade dexterous hands behind a non-anthropomorphic frame, Eno could turn the industry’s two-legged consensus into an open question.

MOBILEYE

🚖 Mobileye is building its own robotaxi fleet

The Rundown: Mobileye, the company that built its name selling self-driving software to automakers, is about to become their competitor — launching its own robotaxi fleet in 2027 even as partners like Lyft and Volkswagen’s MOIA build services on its tech.

The details:

• 100 vehicles are set to launch in an unnamed U.S. city in 2027, scaling to roughly 17K robotaxis over five years if it works.

• Mobileye already powers Lyft’s robotaxis (Dallas, as soon as 2026) and VW’s MOIA subsidiary — both now compete with their own supplier.

• The fleet runs through Moovit, Mobileye’s ride-hailing app; press photos hint at a modified Ora iQ from Great Wall Motors, reports Inside EVs.

Why it matters: CEO Amnon Shashua has wanted this since 2020, when he told TechCrunch the driverless “Holy Grail” runs through robotaxis first — a long-stated plan, now shipping. The test is whether “extension, not replacement” holds up once a supplier and its customers are competing for the same riders.

MORPH

🐙 Inside Morph’s pitch for octopus-inspired physical AI

The Rundown: UK-based Morph just came out of stealth with backing from 8VC and, oddly, Pharrell Williams, to build “soft robotic cells” — squishy, sensor-laden modules that change shape and stiffness on the fly, with the octopus as their design muse.

The details:

• Founder Dr. Jean Nehme, a former reconstructive surgeon, previously sold surgical AI startup Digital Surgery to Medtronic in 2020.

• “Soft robotic cells” are modular units that process environmental and user data and change their morphology to achieve specific objectives.

• Launch use cases are athletic performance, injury prevention, and mobility support; healthcare, automotive, and industrial safety categories come next.

• Morph’s strategy is to serve as a tech, design, and manufacturing partner, helping businesses incorporate soft robotics into their own products.

Why it matters: Soft robotics has spent two decades stuck in labs with impressive demos, but no path to scale. Morph says reinforcement learning paired with physics simulation closes that gap, and the investor list (Equinox’s chairman, a health-capital fund, Pharrell) backs up the real target: wearables first, industrial robotics later.

XDOF

💰 XDOF nabs $70M to solve robotics’ data problem

The Rundown: Berkeley-based startup XDOF emerged from stealth today with $70M in funding to solve robotics’ biggest bottleneck — not models or chips, but the physical-world training data that doesn’t exist yet.

The details:

• XDOF has raised $70M led by Thrive Capital, and it’s already serving 20 customers, including several unnamed frontier AI labs.

• The startup builds teleoperation rigs and wearable sensors across a “data pyramid,” ranging from on-robot teleoperation to egocentric data captured by humans performing everyday tasks.

• XDOF is open-sourcing its ABC-130K robot training dataset: 130K manipulation trajectories, 300 hours of simulation, and 100 hours of evaluations.

Why it matters: Robots can’t learn the way LLMs did, scraping the open internet, because there’s no equivalent trove of physical-interaction data sitting around. XDOF is betting that whoever builds the pipeline to collect it, not just the model on top, ends up owning the most defensible layer of the robotics stack.

QUICK HITS

📰 Everything else in robotics today

Uber will launch its premium Lucid-Nuro robotaxi service in Houston by mid-2027, the second city after San Francisco in its bid to take on Waymo.

Alibaba released the open-source Qwen-Robot Suite, featuring three models that hit 91.4% on LIBERO-Plus and work across quadrupeds, arms, and humanoids.

Stellantis, Wayve, and Uber signed a non-binding MOU to co-develop Level 4 robotaxis, with purpose-made taxis built by Stellantis and equipped with Wayve AI.

European Space Agency’s sensor-packed robotic arm, designed for collecting samples on future Moon and Mars missions, is heading into testing at Italy’s Leonardo.

The Italian Institute of Technology built a soft robotic arm modeled on octopus tentacles that can autonomously navigate and explore rocky ocean-floor terrain.

NVIDIA’s ENPIRE system uses AI coding agents to autonomously design, run, and refine real‑world robot training experiments, teaching arms skills like installing GPUs.

Stringman is an open-source ceiling-mounted robot that uses cables and a vision-guided gripper to find, grab, and sort household clutter into designated bins.

Booster Robotics’ T1 humanoid kicks soccer balls hard enough to dent a wall, sparking debate over how to safely manage that much force in robots.

COMMUNITY

🎓 Highlights: News, Guides & Events

• Read our last AI newsletter: Inside the deadlock keeping Mythos offline

• Read our last Tech newsletter: Xbox’s studio crisis gets bigger

• Read our last Robotics newsletter: UBTech’s ultra-realistic robot girlfriend

• Today’s AI tool guide: Build and host a custom CRM with Google Antigravity

See you soon,

Rowan, Joey, Zach, Shubham, and Jennifer — The Rundown’s editorial team