tech
Businesses face up to budget-busting AI bills
Simply sign up to the Artificial intelligence myFT Digest -- delivered directly to your inbox.
TL;DR
- Businesses initially embraced AI with minimal cost constraints, encouraging widespread experimentation.
- A shift to usage-based pricing by AI providers and the surge in AI agent popularity have dramatically increased costs.
- Companies like Uber have already exhausted their AI budgets early in the year.
- Nearly half of global business leaders have scaled back AI agent use due to costs outweighing benefits.
- Strategies to control AI spending include setting token limits, requiring manager approval for increased usage, and exploring open-source models.
- Vendor lock-in is a growing concern, pushing some companies to develop internal tools for selecting the most cost-effective AI models.
- Open-source AI models offer a potential cost reduction of up to 70% and are attractive for businesses prioritizing data security.
- Despite cost controls, overall AI spending and token consumption are projected to increase significantly by 2030.
- A key lesson for businesses is to tie AI experimentation and innovation to measurable work outcomes.