Highlights

  • Businesses currently track token usage closely to control their AI spending.
  • OpenAI board chairman Bret Taylor believes companies will soon focus more on results than tokens.
  • Taylor compared the current stage of artificial intelligence with the early days of the internet.

Businesses adopting artificial intelligence are currently keeping a close watch on token usage and the money they spend on AI services. However, OpenAI board chairman Bret Taylor believes this may no longer be a major concern within the next year.

Speaking to CNBC, Taylor said businesses are likely to move away from measuring AI costs through tokens and instead pay based on the outcomes these tools deliver.

Tokens are small pieces of text that AI models process while understanding prompts and generating responses. Most AI companies also use token consumption to calculate how much customers are charged. As businesses increased their use of AI earlier this year, rising token expenses prompted many of them to examine whether their investments were delivering enough value.

“I believe where the world is going is paying for outcomes,” Taylor said, reflecting a wider shift among companies towards measuring the actual return on their AI investments.

Taylor compared the current AI industry with the early days of the internet, when developing and maintaining a website was far more complicated and expensive than it is today.

“We’re just in the early stages of the technology, and I think it’s a call for entrepreneurs to develop solutions so businesses don’t need to deal with this stuff,” he said.

According to Taylor, several companies are already creating tools that manage token usage and related costs in the background. He cited Ramp’s token spending tool and legal AI startup Harvey as examples of platforms that handle these technical details for customers.

He also expects organisations to adopt different AI products for different departments rather than relying on a single tool across the company.

“I think if you fast-forward 12 months from now, IT departments will be really sophisticated about the industrial applications of AI,” Taylor said. “For your marketing department, you might have one thing. For your software engineers, another thing.”

As these systems become easier to manage, he believes employees and businesses may not need to think about tokens at all.

“So you just don’t need to think about the word token at all,” he added.

The cost of using AI is also expected to decline as models become faster and more efficient. Chinese AI startup Moonshot recently introduced its Kimi K3 model, which has drawn attention for reportedly delivering performance comparable to leading AI models at a lower cost.

Taylor said affordability will remain one of the biggest factors influencing AI adoption.

“‘Is it cheaper to use’ is the most important part for anyone considering it,” he said. “And I think the jury’s out on that.”