Connect with us

Robotic

KAIST’s Humanoid v0.7 Sprints, Moonwalks, and Kicks Its Way Into the Physical AI Era

South Korea’s KAIST has unveiled a striking field demonstration of its Humanoid v0.7 — a fully homegrown robot that sprints at 7.3 mph, performs a smooth moonwalk, and kicks a soccer ball with precision, powered by Physical AI and deep reinforcement learning.

Published

on

Kaist 0.7 Humanoid Robot

A robot that moonwalks like Michael Jackson, sprints across a soccer field at 7.3 miles per hour, and changes direction mid-stride without losing its footing — South Korea’s KAIST just made the rest of the world’s humanoid robotics labs pay close attention. The university’s latest creation, the Humanoid v0.7, is making headlines this week for a stunning real-world field test that showcases what “Physical AI” can actually look like when it leaves the lab.

Built From the Ground Up — Literally

What makes the KAIST Humanoid v0.7 stand out isn’t just what it can do — it’s how it was built. The entire robot was developed in-house by the Dynamic Robot Control & Design Laboratory (DRCD Lab) under the leadership of Professor Hae-Won Park. That means the motors, gearboxes, and motor drivers were all custom-engineered at KAIST, making the platform almost entirely technically independent from commercial suppliers.

At 165 pounds (75 kg) and standing five-foot-five, the v0.7 is roughly human-sized. But its Quasi-Direct Drive (QDD) architecture — borrowed from the school’s earlier work on legged robots — gives it a key advantage: high torque, low latency, and remarkably smooth force control. That’s the hardware backbone behind every fluid movement you see in the field test.

The Field Test That Went Viral

Footage released this week shows the KAIST v0.7 doing things that would have seemed like science fiction just a few years ago. In the outdoor field test, the robot:

  • Sprints across a grass soccer field at speeds up to 7.3 mph (12 km/h)
  • Kicks a ball toward the goal with accurate follow-through
  • Changes running direction without slowing to a stop
  • Performs a smooth, fluid moonwalk — gliding backward in a way that closely resembles the iconic Michael Jackson move
  • Climbs steps over 12 inches (30 cm) high

The moonwalk, in particular, has attracted enormous attention online. It isn’t a gimmick. According to the DRCD Lab, it’s a demonstration of whole-body balance and fine motor coordination — exactly the kind of capability that separates current-generation Physical AI robots from their predecessors.

The Secret: Physical AI and Motion Capture Priors

Behind the smooth demonstrations is a sophisticated training pipeline built on deep reinforcement learning (DRL). The team trains the robot’s locomotion and manipulation policies entirely in simulation, then transfers them to hardware — a technique known as “sim-to-real” transfer. The magic ingredient that prevents the robot from moving like a stiff, jerky machine is the use of human motion capture data as a behavioral prior.

In practice, this means the robot’s movements are shaped by recordings of actual human motion. Rather than learning locomotion purely from reward signals, the robot learns to mimic the natural dynamics of human walking, running, and kicking. The result is the fluid, almost organic movement quality visible in the field test — a quality that most DRL-trained robots still struggle to achieve.

This approach is central to what the KAIST team calls Physical AI: giving autonomous machines the ability to perceive, interpret, and act on real-world environments without requiring hand-engineered motion primitives. It’s a philosophy that aligns closely with where industry heavyweights like Boston Dynamics, Figure AI, and NVIDIA’s Isaac platform are all heading.

What Comes Next for v0.7

The current v0.7 platform is already impressive, but the DRCD Lab isn’t stopping here. The team has outlined aggressive near-term targets: pushing top running speed to 14 km/h (about 8.7 mph), adding ladder-climbing capability, and achieving step-climbing over 40 cm — more than double the current spec.

Perhaps most interesting is the lab’s work on a system called DynaFlow, which aims to let the robot learn tasks directly from human demonstrations. The concept is straightforward but powerful: a worker performs a task once, and the robot watches and learns to replicate it. If DynaFlow works at scale, it could dramatically reduce the data and programming overhead required to deploy humanoid robots in new environments.

Why This Matters for the Industry

The KAIST Humanoid v0.7 is significant for reasons beyond its impressive party tricks. It represents a national research lab achieving competitive performance with hardware developed entirely in-house — no Boston Dynamics actuators, no off-the-shelf servo ecosystem. South Korea is making a clear statement that it intends to be a top-tier player in the global humanoid race alongside the United States and China.

More broadly, the v0.7 field test is a data point in a rapidly accelerating trend: Physical AI systems that can operate gracefully in unstructured real-world environments are no longer the exclusive domain of well-funded commercial startups. University labs, with the right talent and the right training infrastructure, are closing the gap fast.

At InteliDroid, we’ll be watching closely as KAIST pushes toward its next performance milestones — and as the rest of the field races to answer back.

Robotic

Germany’s NEURA Robotics Lands $1.4 Billion From Amazon, Nvidia, and Tether — The Largest Humanoid Funding Round Ever

German humanoid robotics company NEURA Robotics has closed a record-breaking $1.4 billion Series C backed by Amazon, Nvidia, Tether, and Qualcomm — cementing Europe’s place at the center of the global race to build cognitive robots.

Published

on

By

When the world’s biggest names in tech and logistics write a single company a check for $1.4 billion, you pay attention. On June 10, 2026, Germany’s NEURA Robotics announced the largest funding round in the history of humanoid robotics — a Series C led by Tether and backed by Amazon, Nvidia, Qualcomm, Bosch, Schaeffler, and the European Investment Bank. The round values the Metzingen-based startup at approximately $7 billion and sends an unmistakable signal: the race to build cognitive, general-purpose robots is no longer theoretical.

A Round Built for Scale, Not Just Survival

NEURA describes the $1.4 billion as a ceiling tied to performance milestones rather than a single upfront transfer — a structure that aligns investor capital with real-world deployment progress. The lead investor, Tether, is best known as the issuer of the USDT stablecoin, but the company has been quietly building a portfolio of physical AI bets. Joining them are names that speak directly to NEURA’s roadmap: Amazon brings logistics deployment expertise and a massive warehouse network hungry for automation; Nvidia brings the compute backbone powering physical AI training; Qualcomm brings edge inference silicon; and Bosch and Schaeffler bring deep manufacturing integration across European industry.

The European Investment Bank’s participation is particularly notable. It signals that policymakers see humanoid robotics not just as a commercial opportunity but as a strategic infrastructure investment for Europe’s industrial competitiveness.

Meet the 4NE-1: A Robot Designed for Anyone

At the center of NEURA’s product lineup is the 4NE-1 (pronounced “for anyone”) — a full-size humanoid standing approximately 180 cm tall, capable of lifting 100 kg, and walking at 5 km/h. The Gen 3.5 version, unveiled at CES 2026 with aesthetics co-designed by Studio F.A. Porsche (the team behind the Porsche 911), features patented artificial skin engineered for safe physical collaboration with humans.

What makes NEURA’s pricing unusual in an industry where most companies guard their cost structures is radical transparency: the 4NE-1 Gen 3.5 starts at €98,000 per unit, dropping to €60,000 for fleet purchases of 20 or more. In a market where Tesla’s Optimus and Figure’s 03 remain effectively priceless to outside buyers, that kind of public pricing is a statement of commercial confidence.

Beyond the humanoid, NEURA offers a broader portfolio: the MAV and MiPA mobile robots, the LARA and MAiRA robot arms, and a SenseKit sensor suite — making the company a full-stack physical AI platform rather than a one-product bet.

NEURA Gyms and the “Machine Economy” Vision

Capital from this round will fund something NEURA calls NEURA Gyms — dedicated real-world training environments where cognitive robots learn physical tasks through repeated interaction with the environment, rather than simulation alone. The concept mirrors the insight that drove breakthroughs in language AI: you can simulate a lot, but real-world data is irreplaceable.

CEO David Reger has framed NEURA’s mission around what he calls the “Machine Economy” — a future where robots don’t just automate predefined tasks but participate as productive agents in manufacturing, logistics, healthcare, and eventually consumer life. The $1.4 billion is meant to accelerate serial production to multi-million units by 2030 and expand NEURA Gym infrastructure globally.

Why This Round Matters for the Entire Industry

NEURA’s raise doesn’t happen in isolation. Earlier in 2026, robotics companies collectively raised $55.8 billion globally — nearly double the previous annual record. Figure AI scaled its BotQ factory to one robot per hour. Boston Dynamics shipped the first Atlas units to Hyundai and Google DeepMind. The industry has entered a phase where capital is abundant, but execution is everything.

What NEURA’s round adds to that picture is a European anchor — proof that the humanoid race isn’t solely being run out of Silicon Valley or Chinese state-backed labs. With Bosch and Schaeffler as investors, NEURA has direct pathways into European automotive and industrial manufacturing at a scale few startups can access.

What Comes Next

NEURA has not announced a specific deployment timeline tied to this funding, but the combination of Amazon’s logistics network and Qualcomm’s edge chips suggests warehouse and industrial automation are the near-term targets. NEURA Gyms will likely begin appearing in Europe first, where partnerships with Bosch and Schaeffler facilities offer ready training grounds.

For the humanoid robotics industry as a whole, $1.4 billion flowing to a full-stack European player is a sign that the field is maturing fast — from research curiosity to capital-intensive infrastructure buildout. At InteliDroid, we’ll be watching NEURA’s production ramp closely: the distance between a record funding round and a robot actually working beside humans is exactly the gap the next 18 months will have to close.

Continue Reading

Robotic

NVIDIA Picks Unitree H2 Plus as Its First Research Humanoid Robot Platform

NVIDIA has selected Unitree’s H2 Plus as the hardware backbone for its Isaac GR00T Reference Humanoid Robot, a new open platform shipping to Stanford, ETH Zurich, and other top research institutions in late 2026.

Published

on

By

NVIDIA just made its most ambitious move yet in the humanoid robotics space — and it’s betting on a Chinese startup to deliver it. On June 1, 2026, NVIDIA announced the Isaac GR00T Reference Humanoid Robot, an open research platform built around Unitree’s H2 Plus chassis, and it’s already headed to some of the world’s most prestigious research institutions.

What Is the Isaac GR00T Reference Humanoid?

The Isaac GR00T Reference Humanoid is NVIDIA’s first complete robotics system sold directly to researchers. It combines four major components into a single, ready-to-research package:

  • Unitree H2 Plus — a humanoid chassis standing nearly 6 feet tall and weighing 150 pounds, with 31 degrees of freedom across the body
  • Sharpa Wave tactile five-finger hands — dexterous end effectors with 22 degrees of freedom each, bringing the total to 75 DOF across the full system
  • NVIDIA Jetson Thor — onboard compute module designed for advanced reasoning and real-time robot control
  • NVIDIA Isaac GR00T open software and models — a full software stack with foundation models for perception, planning, and manipulation

Together, these components give researchers a human-scale platform capable of the kind of dexterous manipulation and full-body coordination that most academic labs have never had access to before.

Who’s Getting One — and Why It Matters

NVIDIA isn’t just selling hardware. It’s seeding a new generation of humanoid robotics research. Institutions confirmed to receive the H2 Plus include Stanford Robotics Center, ETH Zurich, UC San Diego’s Advanced Robotics and Controls Laboratory, and Seattle-based Ai2 (Allen Institute for AI).

The significance here is hard to overstate. Until now, most academic research on humanoid robots has been constrained by cost, availability, or the need to build custom platforms from scratch. By offering a standardized, fully integrated system backed by NVIDIA’s software ecosystem, the GR00T Reference Humanoid dramatically lowers the barrier to entry for serious full-body robot research.

For Unitree, the partnership is equally transformative. The Chinese startup — which has built a reputation for affordable, high-performance robots — gains instant global credibility by becoming NVIDIA’s chosen hardware partner. The announcement coincided with Unitree’s move to raise 4.2 billion yuan ($620 million) through a listing on Shanghai’s STAR board.

The Isaac GR00T Software Edge

Hardware alone doesn’t make a research platform — the software stack is where NVIDIA adds its deepest value. The Isaac GR00T framework includes foundation models for robot perception and manipulation, simulation environments for training in synthetic data, and tools for transferring learned behaviors from simulation to physical hardware (sim-to-real transfer).

Researchers at partner institutions will be able to build on top of NVIDIA’s pre-trained models rather than starting from scratch, potentially accelerating timelines for new capabilities by months or years. The open nature of the platform also means that breakthroughs from one institution can be shared across the research community.

Availability and What Comes Next

The NVIDIA Isaac GR00T Reference Humanoid Robot will be available from Unitree in late 2026. NVIDIA has also indicated it plans to expand the program to additional US and European humanoid robot manufacturers, suggesting this is the first step in a broader research ecosystem strategy rather than an exclusive Unitree deal.

For the humanoid robotics field, the timing couldn’t be better. With commercial deployments at BMW, Japan Airlines, and Toyota already proving the concept at scale, academic research is the next frontier — developing the algorithms and capabilities that will define the next generation of industrial and consumer robots.

The Bigger Picture

NVIDIA’s move into humanoid research hardware is a natural extension of its dominance in AI compute. By owning the platform — chips, software, and now the robot itself — NVIDIA is positioning itself as the essential infrastructure layer for the entire humanoid robotics industry, from research lab to factory floor.

For InteliDroid readers, this signals something important: the gap between research and deployment is narrowing fast. When Stanford and ETH Zurich start running experiments on the same hardware that could ship to a warehouse next year, the path from academic paper to real-world robot gets a lot shorter.

Continue Reading

Humanoid Robots

Gatsby Sends a Humanoid Robot Into an American Home — History Made at $150

San Francisco startup Gatsby made U.S. history on May 14, 2026, dispatching a humanoid robot to complete the first-ever paid residential cleaning for an American consumer — at a flat rate of $150 per clean.

Published

on

By

A San Francisco startup just quietly rewrote history. On May 14, 2026, a humanoid robot entered a customer’s apartment, cleaned it from top to bottom, and walked back out — the first time a humanoid machine has ever performed a paid residential cleaning for an end consumer in the United States.

The company behind the milestone is Gatsby, founded in January 2026. Most people hadn’t heard of it. That changes now.

The Moment Happened Quietly — But It Changes Everything

Gatsby selected its first customer entirely at random from a waitlist of eager San Francisco residents. The customer booked through the Gatsby iOS app like any ride-share or food delivery order. The humanoid robot arrived, navigated the apartment autonomously, cleaned it, and left. No human supervisor on-site. No controlled media environment. Just a machine doing housework in a stranger’s home.

This wasn’t a demo for investors. It wasn’t a proof-of-concept with a pre-vetted partner. It was a real commercial transaction — the first of its kind in American history.

Gatsby founder and CEO Aron Frishberg, who left the University of Chicago to build the company under parent firm West Egg Labs, was direct about what’s at stake: “Housework is the largest unpaid job in human history, and it falls hardest on the people with the least time to give. We’ve mapped every neuron and synapse in a fruit fly’s brain, yet we still clean our homes the same way our ancestors did hundreds of years ago.”

$150 to Have a Robot Clean Your Apartment

Gatsby charges a flat rate of $150 per cleaning, regardless of apartment size. Professional human cleaning services in San Francisco typically run between $150 and $300 per visit. On price alone, the robot is immediately competitive.

The service is currently live only in the San Francisco Bay Area, but the waitlist has expanded well beyond the city. Gatsby has signaled plans to scale nationally as operations mature.

For context: consumers have been willing to pay $30 for a 20-minute Uber ride and $15 for grocery delivery. A $150 apartment cleaning — with no scheduling headaches, no background check anxiety, and guaranteed consistency — sits in a price range that millions of households already spend on cleaning services. The robot just removes the human friction entirely.

Gatsby Isn’t Building a Robot — It’s Building the Platform

Here’s what makes Gatsby’s approach strategically distinctive: the company is hardware-agnostic. It does not manufacture its own humanoid robot. Instead, it is building the consumer distribution layer — the software stack, home navigation systems, booking interface, and operational infrastructure required to deploy any humanoid robot into a real residential environment.

Think Uber, not General Motors. Think Airbnb, not Marriott.

While Tesla Optimus, Figure AI, 1X Technologies, and others are spending billions racing to build the ideal mechanical body, Gatsby is betting that the distribution layer — the interface between robots and real consumers — is where the lasting value accumulates. If a cheaper, more capable robot ships next quarter, Gatsby can integrate it and immediately upgrade its service fleet without rebuilding its business model from scratch.

The company is backed by NVIDIA Inception and Entrepreneurs First, two organizations with strong track records of identifying foundational infrastructure plays in emerging tech categories.

Why Cleaning First — and Why It Matters

Cleaning was selected as Gatsby’s launch market with deliberate logic. It is a service that is universally disliked, already commands substantial consumer spending, involves highly repetitive and learnable tasks, and — crucially — has seen almost zero technological disruption since the Roomba introduced robotic vacuuming in 2002.

The humanoid form factor changes the equation. Unlike wheeled robots confined to flat floors, a humanoid can climb stairs, open doors, move objects between rooms, and operate standard household appliances without requiring any modification to the home environment. For the first time, whole-home autonomous cleaning is technically feasible at scale.

Gatsby is explicit that cleaning is a starting point, not a destination. The underlying platform is designed to extend across any domestic service category where a human worker currently enters the home — from laundry and errands to elderly care assistance and package handling.

The Bigger Picture for Humanoid Robotics

For years, the humanoid robotics industry has been defined by warehouse deployments, factory floor integrations, and carefully staged demos. Gatsby’s May 14 milestone represents something qualitatively different: a humanoid robot operating inside the messy, unstructured environment of a real consumer home, completing a task that a paying customer booked through a smartphone app.

This is the consumer era of humanoid robotics beginning in earnest. As hardware costs fall and robot capabilities improve, Gatsby’s platform model positions the company to benefit from every advance made by the underlying hardware manufacturers — regardless of which platform ultimately wins the robot wars.

Mark the date. The robots aren’t just sorting packages in warehouses anymore. They’re cleaning our homes. And if Gatsby’s early trajectory holds, the $150 cleaning will look like a historical footnote in a few years — the moment the robotic home services economy quietly switched on.

Continue Reading
Advertisement

Trending