Artificial intelligence entered a new phase after renowned AI scientist Yann LeCun launched a new startup and secured $1.03 billion in funding to build a new generation of AI systems called world models. The startup, named Advanced Machine Intelligence (AMI), aims to design AI that understands the physical world rather than simply generating text.
The funding round placed AMI among the most well-funded AI startups in the world and marked one of the largest seed rounds in Europe. Investors valued the company at about $3.5 billion before new capital entered the business, highlighting strong confidence in LeCun’s vision for the future of artificial intelligence.
This ambitious venture challenges the dominant approach in AI today. While most companies build larger language models, AMI focuses on systems that can reason, plan, and interact with real-world environments.
Who Is Yann LeCun?
Yann LeCun stands among the pioneers of modern artificial intelligence. He helped develop convolutional neural networks, a technology that revolutionized computer vision and machine learning. His work earned him the 2018 Turing Award, one of the highest honors in computer science.
LeCun spent more than a decade at Meta, where he served as chief AI scientist and led the company’s fundamental research efforts through Facebook AI Research. In late 2025, he decided to leave the company to pursue a new direction in artificial intelligence research through his own startup.
The launch of AMI reflects his long-standing belief that current AI systems lack deep understanding. He argues that systems that only predict text cannot reach human-level intelligence.
AMI: A Startup Built to Rethink Artificial Intelligence
AMI, short for Advanced Machine Intelligence, operates as a research-focused AI company with a long-term vision. The startup plans to create systems that can understand how the real world works.
Unlike conventional AI tools that rely heavily on language data, AMI designs models that learn from visual, spatial, and real-world signals. These systems aim to develop internal representations of the environment and use those representations to make decisions.
The company runs from multiple global hubs, including Paris, New York, Montreal, and Singapore, allowing researchers and engineers to collaborate across major AI ecosystems.
AMI also operates under the leadership of CEO Alexandre LeBrun, while LeCun holds the role of executive chair and scientific leader.
$1.03 Billion Funding Signals Massive Investor Confidence
The startup’s funding round attracted strong interest from venture capital firms and technology investors. Major investors include Cathay Innovation, Greycroft, Hiro Capital, HV Capital, and Bezos Expeditions.
The size of the investment highlights growing investor interest in next-generation AI architectures. Many investors believe that current AI models will reach limits unless new research breakthroughs emerge.
The funding will support several major initiatives:
- Expansion of global research teams
- Development of new AI architectures
- Investment in large-scale computing infrastructure
- Collaboration with industry partners
AMI will also use the capital to recruit leading AI scientists and engineers from around the world.
What Are AI World Models?
AMI focuses on developing AI world models, a concept that attempts to mimic how humans learn about the environment.
Traditional AI systems mainly analyze text and predict the next word in a sentence. This method works well for chatbots and language tools but struggles with complex real-world reasoning.
World models operate differently. They attempt to build an internal simulation of the world, allowing AI systems to understand cause-and-effect relationships.
Such systems could learn:
- how objects move
- how environments change
- how actions affect outcomes
By understanding these relationships, AI systems could plan actions, predict consequences, and operate autonomously.
Researchers believe world models could become essential for robotics, self-driving vehicles, and advanced automation.
Why LeCun Challenges the Dominance of Large Language Models
Large language models (LLMs) dominate the current AI landscape. Many companies invest billions of dollars to scale these systems and increase their capabilities.
LeCun strongly criticizes this approach. He argues that predicting words or pixels cannot produce true intelligence.
According to his research philosophy, human intelligence relies heavily on understanding the physical world. Children learn by observing objects, interacting with environments, and forming mental models of reality.
LLMs, in contrast, learn patterns in text data. They often generate convincing answers but may produce inaccurate or fabricated information.
LeCun believes that world models will overcome these limitations and lead to more reliable AI systems.
Industries That Could Benefit From World Model AI
AMI plans to target industries that operate complex systems and require reliable decision-making. These industries include:
- manufacturing
- automotive technology
- aerospace
- biomedical research
- pharmaceuticals
Companies in these sectors often manage complicated processes where mistakes carry serious consequences.
AI systems that understand real-world dynamics could improve automation, safety, and efficiency across these industries.
Over time, the technology could also power consumer applications such as domestic robots and intelligent devices.
A New Direction in the Global AI Race
The global AI race currently centers on companies that build larger generative models. Firms such as OpenAI, Anthropic, and Meta focus heavily on scaling language models to achieve higher intelligence.
AMI introduces a competing strategy.
Instead of scaling language models indefinitely, LeCun’s team aims to build fundamentally different AI architectures. These architectures rely on learning from real-world data rather than purely from text.
This strategy could reshape the AI landscape if the technology proves successful.
The startup’s launch also reflects a broader trend in the AI ecosystem. Researchers increasingly explore alternative approaches that combine reasoning, perception, and planning.
The Road Ahead for AMI
AMI faces significant technical challenges. Building world models requires enormous computing power and complex training methods.
The company will invest heavily in research before releasing commercial products. Early deployments may appear in specialized sectors where companies can test the technology under controlled conditions.
Despite the challenges, the startup’s large funding round demonstrates strong belief in LeCun’s vision.
The AI pioneer has spent decades studying machine intelligence, and his latest venture attempts to solve one of the biggest questions in technology: how to build machines that truly understand the world.
If AMI succeeds, the company could redefine how artificial intelligence evolves in the coming decade.
Also Read – Indian Startups to Create 80,000 Tech Jobs in FY2026