AI will shift to “pay-for-results”? OpenAI chairman predicts token payment model will overturn the industry within a year

```

AI token costs are troubling many corporate financial decision-makers, but OpenAI's chairman believes this issue will disappear within a year.

OpenAI Chairman Bret Taylor said on Monday that as the AI market matures, companies will eventually no longer need to manage tokens themselves and will instead pay directly for business outcomes. This assessment is highly consistent with the current industry trend of growing focus on AI investment returns. Previously, the rapid rise in token costs had prompted many companies to cut back on AI spending and reexamine their ROI.

Taylor also expressed caution regarding the Kimi K3 model released last week by Chinese startup Moonshot. Although the model is said to achieve performance comparable to the flagship models of OpenAI and Anthropic at a lower cost and has attracted widespread attention in Silicon Valley, Taylor stated that the key question of "whether it’s really cheaper to use" remains undecided at present.

The Token Dilemma: Cost Pressure Drives Industry Change

Tokens are the basic measurement unit for AI models to process text and are currently the main basis by which companies calculate AI usage costs. Earlier this year, the spiraling token costs forced many companies to cut AI budgets, triggering heated debates around investment returns.

In an interview with CNBC, Taylor pointed out that the root of today's various "token" issues is that the AI market is not yet mature. He likened today’s AI market to the early days of the Internet—when building a website was far more expensive than it is now. "We are still at an early stage of this technology, which is a call to entrepreneurs to develop solutions so companies no longer need to deal with these things."

From Managing Tokens to Paying for Outcomes

Taylor predicts that in the future, other companies will shoulder the burden of token management for enterprises, and the industry will evolve toward a "pay for outcomes" model. This view aligns with the positions of other tech executives who emphasize ROI.

He cited the recent launch of a token spend tracking tool from financial management platform Ramp, as well as vertical AI startups like the legal tech company Harvey— the latter has already begun managing token usage on behalf of customers— as evidence of this shift.

Taylor further predicted that within the next 12 months, IT departments will become much more mature in industrial AI applications. "For the marketing department, you might use one solution; for software engineers, another. This way, you won’t even need to think about the word 'token' anymore."

A Cautious Stance on Low-Cost Models

Although the industry generally expects improved AI model efficiency to drive long-term token cost reduction, Taylor does not wholly agree with this view.

Regarding the Kimi K3 model released by Moonshot last week, Taylor reserved judgment. "'Whether it’s really cheaper to use' is the most critical question for anyone considering it, and I believe that conclusion has yet to be reached." His comments suggest that before the cost advantage is fully validated, companies should continue to rationally assess emerging low-cost models.

Risk Disclosure and DisclaimerThe market has risks, and investments must be cautious. This article does not constitute personal investment advice and does not take into account the individual investment objectives, financial situations, or needs of any particular user. Users should consider whether any opinions, views, or conclusions in this article are suitable for their specific circumstances. Investments made accordingly are at your own risk. ```