Large companies plan to expand open-weight AI use as costs rise

Large companies plan to expand open-weight AI use as costs rise
Big firms shift to open AI

Large enterprises are shifting more of their artificial intelligence workloads toward lower-cost open-weight models as spending runs above expectations. A Capgemini study shows most surveyed executives expect that shift to accelerate over the next year, while many still keep closed models for more complex tasks.

Highlights

  • Capgemini reports 59% of large-company executives plan to increase open-weight AI model usage over the next year as AI costs exceed budgets by 24% on average.
  • Open-weight tokens are 87% cheaper than closed tokens, with open models handling 72% of OpenRouter traffic but accounting for just 27% of AI spending.
  • Companies like SAP are substituting expensive AI models with cheaper open-weight alternatives for lower-complexity tasks, while half of surveyed firms reserve closed models for complex applications.

Survey points to broader cost-driven AI shift

As reported by Reuters, citing Capgemini, nearly 6 in 10 senior executives at large companies plan to increase their use of cheaper open-weight AI models over the next year as AI bills overshoot budgets.

The consulting firm says 59% of surveyed executives report that AI spending over the past year is significantly above expectations or budgets, by 24% on average. It also says 53% are worried about AI costs, up from 45% a year earlier.

Capgemini surveyed 500 executives at enterprises with annual revenue of more than $1 billion across 10 countries. The study says open-weight tokens are on average 87% cheaper than closed ones.

Open models gain traction across enterprise AI

Open-weight models, including those from DeepSeek and Mistral, can be downloaded and run in different environments, while closed models such as Claude, ChatGPT and Gemini are available only through the cloud.

Capgemini says open-weight models handled 72% of traffic on OpenRouter in September, although that accounted for just 27% of AI spending. The findings suggest companies are increasingly using lower-cost models for a large share of volume while reserving more expensive systems for selected applications.

That trend mirrors comments from SAP's finance chief last month that the company is replacing some of its costliest AI models with cheaper alternatives where performance remains sufficient. Closed models still retain a role, with half of respondents saying they will keep them for tasks whose complexity justifies the added cost.

In earlier coverage, we examined a KPMG survey on AI agents showing early efficiency gains but limited transformational change at large U.S. firms.

This material may contain third-party opinions, none of the data and information on this webpage constitutes investment advice according to our Disclaimer. While we adhere to strict Editorial Integrity, this post may contain references to products from our partners.