AI costs

Sam Altman highlighted AI cost reduction as a key topic at the Sun Valley Conference, where tech leaders sought ways to improve AI efficiency and return on investment. OpenAI's new GPT-5.6 models aim to address these concerns with enhanced token efficiency and cost-effective options.

Companies are struggling to reduce rising AI costs, and the trend of 'tokenmaxxing' is ending.

OpenAI CEO Sam Altman and Google CEO Sundar Pichai reported companies are raising concerns about rising AI costs and limited ROI. Uber COO Andrew Macdonald highlighted challenges in linking AI spending to productivity gains, reflecting broader industry anxieties about budget overruns and underwhelming returns.

A report from Bain & Company indicates that 20-30% of business operating expenses will be attributed to AI agent costs in the next three to four years. While 87% of workers use AI at work, only 13% report significant performance improvements, with many using unapproved tools or methods. New AI models like Anthropic’s Fable are twice as expensive as OpenAI’s top model, prompting businesses to adopt cost-effective model mixes.

Datadog has added GPU monitoring to its observability stack, providing AI-focused organizations with deeper insights into GPU efficiency as AI-related costs rise. This enhancement aims to help businesses optimize their expensive hardware resources.