Back to News
Source: Letsdatascience
Published July 8, 2026Read original source

Ex-Tesla Scientist Launches European Humanoid Robot Startup - Let's Data Science

A Europe-first humanoid vendor shifts the operational and integration priorities AI and robotics teams must weigh, including local regulation, industrial workflows, and demonstration-driven learning pipelines.

Ex-Tesla Scientist Launches European Humanoid Robot Startup - Let's Data Science - Image 1
humanoid
robot
ShareLinkedInX

Key takeaways

  • The humanoid‑robot sector is accelerating worldwide as Chinese leaders Unitree and UBTECH have just opened full‑cycle facilities in Tianjin’s Economic‑Technological Development Area, creating a local manufacturing base that will handle research, production, after‑sales service and secondary development, while the Chinese government pushes a six‑month commercialization challenge for commercial‑grade humanoids.
  • In Europe, former Tesla Autopilot engineer Rémi Cadène’s Paris‑based startup UMA unveiled its first lightweight humanoid, Northstar, at the Machina Summit, showcasing a “real‑time learning” system that lets the robot acquire skills by watching demonstrations and targeting manufacturing, logistics and eventually home use; the company is already in talks with about 50 potential customers and is backed by investors such as Greycroft and Kima Ventures.
  • In the United States, Apptronik has expanded its Austin “Robot Park” to collect massive data sets and train its Apollo humanoid platform together with Google DeepMind, while AGIBOT introduced the A3 humanoid in Europe and launched a UK robot‑as‑a‑service offering, and its sister firm Agibot celebrated the roll‑out of its 15,000th unit, the G2 industrial‑grade robot.
  • Meanwhile, the U.S. startup Humanoid announced its KinetIQ Ascend reinforcement‑learning approach, claiming 99.9 % manipulation reliability at human speed and securing partnerships with Bosch and Schaeffler to scale production of its HMND line.
  • Business Insider’s recent coverage notes that Agility’s Digit has secured more than $300 million in multiyear orders and that 1X plans to ship over 10,000 home‑use humanoids later this year, even as some Silicon‑Valley investors remain skeptical of mass‑market humanoid demand.

A Europe-first humanoid vendor shifts the operational and integration priorities AI and robotics teams must weigh, including local regulation, industrial workflows, and demonstration-driven learning pipelines. Business reporting shows that UMA, a Paris-based startup led by Rémi Cadène, unveiled a lightweight humanoid called Northstar and a learning architecture named Real-Time Learning at the Machina Summit, according to BusinessWire. Electrek and The Next Web report Cadène previously worked on Tesla's Autopilot and Optimus and later led the open-source LeRobot effort at Hugging Face. Electrek and The Next Web say UMA emerged from stealth in December 2025, is talking to about 50 potential customers (Electrek/Bloomberg), and lists investors including Greycroft and angel backers such as BusinessWire reports that UMA, a Paris-based physical AI company, unveiled the design of its first humanoid and introduced Real-Time Learning, "a learning architecture that enables robots to acquire new skills through demonstration rather than manual programming," at the Machina Summit. Electrek and The Next Web report UMA's humanoid, called Northstar, is being developed for manufacturing plants, warehouses, logistics centers, and eventually homes, and that the company emerged from stealth in December 2025. Electrek reports founder Rémi Cadène said UMA is already talking to about 50 potential customers. The Next Web and Electrek list backers and advisers including Greycroft, Red River West, Kima Ventures, and angel supporters such as Yann LeCun and Thomas Wolf; Electrek additionally names View 2 more sources

  1. 04Ex-Tesla scientist Rémi Cadene unveils plans for European humanoid robotmoneycontrol.com
  2. 05Ex-Tesla Optimus scientist unveils European humanoid robot startupelectrek.co

Practice with real Ad Techdata

90SQL & Python problems · 15 industry datasets

Used by DS/ML engineers at top companies

Active Search Campaigns by BudgetEasyHigh CPC Clicks & Poor Landing PagesMediumCampaign ROAS by Attribution ModelHard

250 free problems · No credit card

See all Ad Tech problems

More AI & Data Science News

Jul 7

IBM Adds Compact z17 and LinuxONE AI SystemsJul 7

QBE Details Agentic AI Risk to Commercial EspionageJul 7

Samsung Results Trigger Global Chip Stock Selloff6.8

Jul 7

Micron Faces Pressure From AI Memory Demand and Market Sentiment

Mentioned in this article

Read original sourceMore robotics news