Advanced Machine Intelligence Labs (AMI Labs)
Category: Foundation Models / LLMs
A next-generation AI research and development lab focused on creating 'World Models' that understand physical reality, causality, and reasoning beyond the capabilities of current Large Language Models. Advanced Machine Intelligence Labs (AMI Labs) was founded in 2025. The company is led by Alex LeBrun. Based in Paris, France. Team size: 10-50. Total funding raised: $1.03B. Latest round: Seed. Key investors include Cathay Innovation, Greycroft, Hiro Capital, HV Capital, Bezos Expeditions, NVIDIA, Temasek, Toyota Ventures, Eric Schmidt, Xavier Niel, Mark Cuban, Samsung, Sea, Publicis Groupe, Bpifrance Digital Venture, Alpha Intelligence Capital, Artémis, SBVA, New Legacy Ventures.
AMW Analysis
Advanced Machine Intelligence Labs (AMI Labs) builds “World Models” based on the Joint Embedding Predictive Architecture (JEPA), which learns from sensory data to understand physical reality, causality, and reasoning—aiming to overcome hallucinations and lack of common sense in current LLMs. The company targets applications in healthcare, robotics, manufacturing, and industry. Since its launch in early 2026, AMI Labs has raised a total of $1.03 B in seed funding, initially reported at $500 M and later closed at €890 M ($1.03 B), giving the startup a $3.5 B valuation. The investor syndicate includes Cathay Innovation, NVIDIA, Bezos Expeditions, Temasek, Toyota Ventures, and others. The funding round has been described as Europe’s largest seed round and one of the largest in AI history for a company only months old. The news flow consistently positions AMI Labs as a contrarian bet against language-only models, signaling a strategic shift toward grounded, physically intuitive AI for safety-critical sectors.
AMW analysis, generated from 8 tracked news signals.
- Founded
- 2025
- Headquarters
- Paris, France
- Team size
- 10-50
- Total funding
- $1.03B
Value proposition
Solves the fundamental limitations of generative AI—such as hallucinations and lack of common sense—by building AI that learns through observation and internal simulation of the physical world.
Products and solutions
JEPA-based World Models (Joint-Embedding Predictive Architecture), Autonomous AI Reasoning Engines, Physical World Simulation Frameworks for Robotics, Advanced Machine Intelligence (AMI) Developer Stack
Unique value
Founded by Turing Award winner Yann LeCun, the company pivots away from the industry-standard 'Auto-regressive LLM' approach toward 'Objective-Driven AI' that mimics human-like learning and planning.
Target customer
Enterprise sectors requiring high-autonomy systems, including robotics manufacturers, autonomous vehicle developers, and industrial automation firms.
Industries served
Robotics and Automation, Autonomous Transportation, Aerospace and Defense, Scientific Research and Simulation
Technology advantage
Utilizes proprietary implementations of Joint-Embedding Predictive Architecture (JEPA), allowing the AI to predict outcomes in abstract representation space rather than generating data point-by-point, leading to superior reasoning and energy efficiency.
How they differentiate
Pivots away from traditional auto-regressive Large Language Models (LLMs) toward 'World Models' using Joint-Embedding Predictive Architecture (JEPA). This allows the AI to understand physical reality, causality, and common sense, avoiding the 'hallucination' and reasoning limitations of current generative AI.
Main competitors
OpenAI, Physical Intelligence (π), Google DeepMind
Key partnerships
Nabla (first disclosed partner for early model deployments), NVIDIA (strategic investor), Toyota Ventures (strategic investor), Temasek (strategic investor), Samsung (strategic investor), Sea (strategic investor), French AI ecosystem (Paris research hub), Academic research community via open publications and open source
Notable customers
Nabla (first disclosed partner for early world model deployments), Stealth/Pre-revenue (Targeting Industrial Process Control, Automation, Robotics, Healthcare)
Major milestones
Yann LeCun exits Meta to found AMI Labs in 2025, Appointment of serial entrepreneur Alex LeBrun as CEO, Secured €890M ($1.03B) seed round — largest seed round for a European company — at $3.5B pre-money valuation (March 2026), Establishment of primary research hub in Paris, France, with offices in New York, Montreal, and Singapore, Launched official website at amilabs.xyz (January 2026), Announced Nabla as first disclosed partner for early world model deployments
Growth metrics
Rapidly scaling research team to 10-50 employees; targeting a €3B ($3.2B) valuation within first year of operations.
Market positioning
Next-generation AI research lab focused on 'Advanced Machine Intelligence' (AMI) for physical-world applications and robotics.
Geographic focus
Europe (Headquartered in Paris, France) and North America.
Patents and IP
Proprietary architectures based on LeCun’s foundational JEPA research; specific commercial patents pending under AMI Labs.
About Alex LeBrun
Serial entrepreneur; Co-founder and CEO of Nabla (medical AI) and Wit.ai (acquired by Facebook); former Director of Engineering at Facebook AI Research (FAIR) and CEO of VirtuOz (acquired by Nuance).
Latest news about Advanced Machine Intelligence Labs (AMI Labs)
- Advanced Machine Intelligence Labs closed a record €890 M ($1.03 B) seed round, valuing the startup at $3.5 B. Backed by investors like Cathay Innovation, NVIDIA and Jeff Bezos, AMI will push Joint Em
- Yann LeCun's AMI Labs raised $1.03B in Europe's largest seed round at a $3.5B valuation, backed by Nvidia, Temasek, and Jeff Bezos. This challenges LLM dominance by pursuing world models that understa
- Yann LeCun's Advanced Machine Intelligence (AMI) has secured $1.03 billion in seed funding at a $3.5 billion valuation, making it one of the largest seed rounds in AI history for a startup just 3 mont
- Turing Prize winner Yann LeCun's AMI Labs just secured $1.03 billion at a $3.5B valuation to build world models based on JEPA architecture that learns from physical reality rather than language alone.
- Yann LeCun launched AMI Labs in Paris, raising $500M at a $3.5B valuation to build world models that learn physics through sensory data, not text. Unlike LLMs, these systems use JEPA architecture to p
More Foundation Models / LLMs companies
Official website: https://amilabs.xyz