Token-maxing is an AI cost sink - how to use agents without busting your budget (www.zdnet.com)

In a recent discussion at Snowflake's Summit 2026, industry leaders highlighted the exploding costs of AI token consumption, emphasizing the need for effective "tokenomics"—a sophisticated approach to managing token usage in AI applications. Lucas, a tech CEO, noted a staggering tenfold increase in spending on AI services over the past year, pointing out the unsustainable nature of such growth and the urgent necessity for business leaders to develop enterprise-ready strategies. The overarching question now is whether organizations can harness AI effectively while ensuring a reasonable return on investment. Snowflake's CEO Sridhar Ramaswamy and Synopsys CIO Sriram Sitaraman echoed this sentiment, advocating for a balance between exploring AI’s potential and establishing guidelines to mitigate costs. Rather than imposing strict limits on token usage, they suggest providing context for AI projects, which can lead to more efficient consumption and innovative solutions. Caution was a recurring theme, with TF1's Francois-Xavier Pierrel comparing unchecked AI usage to a sugar addiction, warning that without careful management, unexpected costs could escalate. As businesses continue to integrate AI, adopting a thoughtful tokenomics strategy will be crucial for sustainable growth and maximizing value.