Recent data from Vercel’s AI Gateway reveals that AI models are losing relevance faster than ever, with most token expenditure concentrated on models released within the last three months. This accelerated cycle favors open-weight models, which now command a majority share of token volume but a fraction of overall spend, signaling a strategic pivot in AI deployment economics.

  • AI models older than three months are increasingly retired and replaced with newer versions.
  • Open-weight models now account for 56% of token volume but just 14% of token spend.
  • Average token price has declined 23.2% in September, marking a third consecutive monthly drop.

Market signal

The AI market is experiencing unprecedented churn as models become outdated at record speeds, with the most recent iterations attracting the majority of token purchases. This rapid obsolescence challenges traditional procurement models that rely on longer-term token usage and raises questions about capital efficiency in AI deployments.

Another significant market development is the increasing dominance of open-weight AI models. These models, which allow self-hosting and customization, are driving a restructuring of token economics by offering a lower-cost alternative to closed, proprietary systems. Token volume share for open-weight models jumped from 13% in April 2026 to 56% in September, reflecting growing operator preference for adaptable AI solutions.

Operator impact

Operators must adapt to accelerated AI model refresh cycles that are compressing the effective lifespan of token investments. Committing budget to tokens tied to older models risks rapid write-offs as newer, more advanced models replace them within months. This environment increases the need for agile token purchasing strategies aligned with evolving AI capabilities.

The growing popularity of open-weight models also offers operators more procurement flexibility and potential cost savings. Enterprises can leverage tunable, self-hosted models to reduce reliance on expensive closed-box AI systems, optimizing their AI infrastructure spend. However, this also raises operational considerations around deployment complexity and support.

What to watch next

Monitor continued token pricing trends, especially for open-weight versus closed models, as this will indicate how market competition and AI innovation influence operator economics. Sustained declines in token costs could further accelerate the shift to open-weight alternatives.

Evaluate developments from emerging AI providers such as TypeSafe AI with their Jev model, which has seen rapid adoption due to low token pricing and unique data input methods. The success of these new entrants may reshape available AI options and impact legacy model viability.

Source assisted: This briefing began from a discovered source item from TechRadar. Open the original source.
How SignalDesk reports: feeds and outside sources are used for discovery. Public briefings are edited to add context, buyer relevance and attribution before they are published. Read the standards

Related briefings