Enterprises Ditch 'Tokenmaxxing' as AI Cost Efficiency Takes Center Stage at Microsoft Build and Snowflake Summit

Enterprises are abandoning costly 'tokenmaxxing' AI strategies as efficiency dominates Microsoft Build and Snowflake Summit, with major tech firms unveiling cost-cutting tools and experts urging businesses to question whether AI is even needed for every task.

• The era of 'tokenmaxxing' is coming to an end as enterprises grow increasingly cost-conscious about AI spending, with efficiency emerging as the dominant theme at both Microsoft Build and Snowflake Summit this week.
• Major tech companies are releasing efficiency-focused products in response to the shift, including Snowflake's Cortex Training and Adaptive Compute systems, and Microsoft's new 35-billion-parameter reasoning model alongside local-compute hardware designed to reduce cloud token costs.
• Industry experts are urging enterprises to question whether AI is even necessary for a given task, warning that every prompt and agentic tool call adds to the bill, and that some automation can be achieved without AI altogether.