Yann LeCun's New AI Venture Targets Aggressive Funding Goals with World Model Technology

The legendary AI researcher behind deep learning has formally unveiled his latest venture, Advanced Machine Intelligence (AMI Labs), marking a significant new chapter in the world model development race. While LeCun assumes the role of Executive Chairman, the company has enlisted Alex LeBrun — formerly helming medical AI platform Nabla — to lead operations as CEO.

Ambitious Valuation and Fundraising Plans

AMI Labs is already positioning itself aggressively in the capital markets, reportedly pursuing €500 million in funding at an initial €3 billion valuation (approximately $3.5 billion) — this before the company has even launched its first product. The valuation reflects investor enthusiasm for world model research, though it remains notably lower than some competing efforts in the space.

For context, when Mira Murati (the former CTO at OpenAI) established her AI startup last year, backers valued Thinking Machines Lab at $12 billion during its seed round — nearly four times higher. Yet LeCun brings formidable credentials to the table: the Turing Award winner spent years as Meta’s Chief AI Scientist and currently holds a professorship at New York University, lending significant weight to the venture’s credibility.

The World Model Difference

AMI Labs is specifically targeting world model architecture as its core technology — a fundamentally different approach compared to large language models. Rather than generating text based on probabilistic pattern matching, world models aim to build comprehensive understanding of environmental dynamics, enabling AI systems to simulate cause-and-effect relationships and predict outcomes across hypothetical scenarios.

This design philosophy directly addresses a critical weakness in LLMs: their tendency toward hallucination. Because language models are inherently non-deterministic systems trained to be generative and creative, they cannot be engineered to guarantee factual accuracy. World models, by contrast, promise a more grounded, deterministic approach to reasoning.

LeBrun’s Leadership Transition

LeBrun brings substantial experience navigating AI commercialization. His career spans early work at Nuance Communications — which powered Apple’s Siri before Microsoft’s acquisition — multiple natural language processing ventures, and a leadership role managing Facebook’s AI infrastructure prior to founding Nabla in 2018.

Under his tenure, Nabla has demonstrated impressive growth metrics, more than tripling its annual recurring revenue this year with sights set on reaching $1 billion. The medical transcription AI company secured $70 million in its Series C round this past June, bringing total capital raised to $120 million from prominent backers including LeCun himself, alongside HV Capital, Highland Europe, and Tony Fadell’s Build Collective.

Interconnected Ecosystem

The relationship between AMI Labs and Nabla remains intertwined: Nabla has secured a partnership agreement to integrate AMI’s models as they achieve production readiness. LeBrun’s departure leaves the medical AI platform in the hands of co-founder and Chief Operating Officer Delphine Groll on an interim basis, though the board is actively recruiting a permanent replacement.

Competitive Landscape

The world model category is attracting heavyweight attention. Google DeepMind and researcher Fei-Fei Li’s startup World Labs are pursuing similar technical directions. When World Labs initially raised $230 million at a $1 billion valuation in August 2024, the figure was considered exceptional. Today’s fundraising environment for AI has accelerated dramatically, making AMI’s ambitions considerably more feasible within current market dynamics.