Brief IA

Jensen Huang and Sam Altman: The Future of AI Tokens

🤖 Models & LLM·Tom Levy·

Jensen Huang and Sam Altman: The Future of AI Tokens

Jensen Huang and Sam Altman: The Future of AI Tokens
Key Takeaways
1Jensen Huang from Nvidia suggests integrating AI tokens into engineers' compensation to boost their productivity.
2Sam Altman from OpenAI envisions AI tokens becoming a form of universal basic income, transforming the global economy.
3The rising costs of AI tokens worry companies, as highlighted by Chamath Palihapitiya, amid exponential spending.
💡Why it mattersAI tokens could revolutionize economic and social models, but they pose significant financial challenges for businesses.
Le brief IA que lisent les pros

Le brief IA que les pros lisent chaque soir

Les 7 actus IA du jour, décryptées en 5 min. Gratuit.

Inclus dès l'inscription : notre sélection des meilleurs guides & comparatifs IA.

Choisis ton rythme

Gratuit · Pas de spam · Désabonnement en 1 clic

📄
Full Analysis

Jensen Huang and AI Tokens: A New Compensation Tool

Jensen Huang, the CEO of Nvidia, recently highlighted an analysis that positions his company as the global leader in AI token costs, proclaiming himself the "King of Tokens." According to Huang, it is crucial for companies to allocate substantial budgets for AI tokens for their engineers. This approach aims to integrate these tokens into the compensation package for engineers, a strategy that could become a standard in Silicon Valley. Huang emphasized that access to tokens is a key productivity factor for engineers.

During Nvidia's GTC 2026 in San Jose, Huang explained that tokens have become an important recruitment criterion. "The question is no longer just about salary, but how many tokens are included in my job?" he stated. This approach reflects the growing importance of tokens in the tech industry, where each engineer benefiting from these resources can significantly boost their productivity.

AI Tokens: A Universal Language for AI

Tokens play a central role in the functioning of large language models. They break down words and other inputs into numerical units, simplifying data processing by AI models. On average, each token represents about three-quarters of a word. This method of decomposition has become the norm for billing services from AI model providers, with rates based on the number of tokens used and generated.

Sam Altman and the Vision of Universal Basic Income

Sam Altman, CEO of OpenAI, proposes an even more ambitious vision for AI tokens. He envisions a future where these tokens could constitute a form of universal basic income. Altman imagines a system where every citizen would have a share of AI computing power that they could use, sell, or donate for causes like cancer research. During an episode of the All-In podcast in May 2024, Altman discussed the idea of "Universal Base Computing" rather than traditional basic income, where productivity would be shared in the form of tokens rather than dollars.

Altman also compared the production of raw intelligence by AI to a public utility, such as water or electricity. At the BlackRock Infrastructure Summit in Washington, D.C., he asserted that the sale of tokens would become a central activity for AI model providers.

The Challenges of AI Infrastructures and Satya Nadella's Perspective

Altman's vision relies on the successful construction of vast AI data centers, which Jensen Huang refers to as "token factories." However, Satya Nadella, CEO of Microsoft, warned that AI must prove its utility to justify the use of scarce resources like energy. At the World Economic Forum in Davos, Nadella emphasized that tokens must improve areas such as health, education, and economic competitiveness to maintain their social legitimacy.

Rising Costs of AI Tokens: A Concern for Businesses

Not all industry players share the same enthusiasm for AI tokens. As AI models evolve, users are calling on them for increasingly complex tasks, which can quickly drive up costs for businesses. Chamath Palihapitiya, venture capitalist and founder of 8090, recently expressed his concerns about rising AI costs. His company, which aims to replace legacy software, has seen its expenses triple since November 2025, without a proportional increase in revenue. During an episode of the All-In podcast, Palihapitiya mentioned the high inference costs paid to AWS and expenses with companies like Cursor and Anthropic, amounting to millions.

Brief IA — L'actualité IA en français

L'essentiel de l'actualité de l'intelligence artificielle, décrypté et expliqué chaque jour.