Yann LeCun Says xAI Has Failed and the Current AI Business Model Is Unsustainable

The giants of the tech world keep investing in the future of artificial intelligence (AI) with budgets in the hundreds of billions of dollars. Lately, however, not only independent experts and economists but also pioneers who played a critical role in AI's development have been voicing serious concerns about whether this trajectory is sustainable.

Yann LeCun, regarded as "one of the founding fathers of AI" thanks to his revolutionary work in deep learning, sharply criticized both the economic foundations of today's AI ventures and the technological approaches dominating the industry in an interview with CNBC. According to LeCun, surviving in the long run with today's business models does not look possible, and xAI, one of the most popular players in this race, has already failed.

Why Does LeCun Think Elon Musk's xAI Has Failed?

Meta's Chief AI Scientist Yann LeCun painted a rather bleak picture of xAI, the venture founded by Elon Musk that drew attention with its rapid growth. According to LeCun, xAI has fallen short of its targeted success criteria. He highlighted the following main points to support this claim:

* Departure of Key Staff: LeCun notes that a significant share of the most talented engineers and researchers who were part of xAI's founding phase have left the company, and he argues that this exodus has weakened the organization's technical depth. * Hiring Difficulties: LeCun says Elon Musk's turbulent and high-pressure relationships with employees in the past have created a negative image among elite AI researchers, and that xAI is struggling to attract the most talented people. * Financial and Economic Pressure: xAI's recent massive investments in hardware and data centers were not enough to convince LeCun either. He believes these investments are in fact a sign of the economic stress gripping the entire AI industry. The costs of training and running models are very high, and this spending has not yet been balanced by a sustainable revenue model.

Not Just xAI: The Whole Industry Faces a Sustainability Test

The core problem LeCun points to is not limited to xAI. The renowned researcher argues that the industry's overall economic structure rests on a major imbalance, noting that the operating costs of AI services are not falling fast enough to deliver profitability.

On the other hand, the prices users are willing to pay for these services are quite limited. Even though companies raise billions of dollars, at the end of the day they face very large losses. Indeed, recent figures showing that OpenAI's accumulated losses have reached $39 billion support LeCun's concerns.

LeCun predicts that if current market conditions continue, the industry will enter a sharp correction. Companies will have to either radically raise subscription prices or cut operating costs significantly; otherwise, the bursting of the multibillion-dollar AI bubble will be inevitable.

"Large Language Models (LLMs) Alone Are Not Enough"

Yann LeCun's criticism is not limited to financial matters. He also emphasizes the technological limits of Large Language Models (LLMs), which sit at the center of today's AI, reminding us that despite their success at tasks like generating text or writing code, these models have fundamental shortcomings when it comes to understanding the real world.

LeCun argues that the future of AI will not rest on language models alone, and that the "World Models" approach he has long championed needs to be adopted. Under this approach, AI must be able to grasp the physical world around it, cause-and-effect relationships and how objects interact, just as a human or an animal does. He states that the path to reliable and truly general-purpose AI (AGI) systems is only possible through this technological shift.