AI Tokens: All You Need To Know About The New Currency Of Computing

Read why Nvidia CEO Jensen Huang believes computers will soon be treated like manufacturing equipment that generate AI tokens instead of standard software.

Advertisement
Read Time: 2 mins

As AI becomes more and more central to global workforces, Nvidia CEO Jensen Huang, whose GPU units power the majority of the world's AI, has coined a new concept that could disrupt the way legacy software or even AI is perceived.

Speaking at Nvidia's GTC Conference, Huang labelled future computers as 'manufacturing equipment' that produce tokens, thus bringing about a potential revamp as to how businesses think about technology cost.

Advertisement

What are tokens?

Tokens can be thought of as just another currency. In simple terms, they are small chunks of text that AI systems process. A token could be a full word or part of a word. It could even be a few characters.

When an individual types a prompt into AI systems such as ChatGPT or Gemini, the system counts how many tokens are used in both the input and the response. This could determine how much the service costs.

Advertisement

The more text you generate or analyse, the more tokens are used and therefore, the higher the cost.

Why it matters

It is perhaps fair to acknowledge Jensen Huang's idea, as it could become mainstream in the world of AI in the coming years or even months.

Unlike traditional software, which is usually sold through subscription or fixed fees, AI could indeed become price-based on usage. Some entities have already adopted this token-based system, such as Claude.

Advertisement

That means tokens could become a regular expense for companies, much like cloud storage or software licenses. Huang suggested that in the future, engineers could even be given “token budgets” to do their work more efficiently. In some cases, these budgets could be significant, especially if AI tools boost productivity.

A shift in how work gets done

The idea is tied to the rise of AI “agents” — systems that can perform tasks on their own. These agents could run continuously, generating tokens around the clock, even when users are not actively working.

This could lead to a world where computers are always active, producing output and consuming resources in the background.

For businesses, the takeaway is simple: as AI becomes more embedded in daily operations, understanding tokens — and managing their costs — will become increasingly important.

Essential Business Intelligence, Continuous LIVE TV, Sharp Market Insights, Practical Personal Finance Advice and Latest Stories — On NDTV Profit.

Loading...