Open Models Now Handle 80-90% of Enterprise AI Tokens, Says Ollama CEO

Jeffrey Morgan's statement represents validation of a long-predicted industry shift: open-source LLMs are now the dominant force in enterprise AI token consumption. This metric—measuring actual production usage rather than theoretical capability—signals that organizations have moved beyond experimentation with local models and are deploying them at scale. The 80-90% figure encompasses both self-hosted and on-device deployments, reflecting mature tooling, cost advantages, and acceptable performance characteristics for real-world workloads.

This trend accelerates the investment and focus on projects like Ollama itself, which has made local LLM deployment accessible to non-specialists. When enterprises consume 80-90% of tokens from open models, infrastructure improvements and optimization become commercially justified, attracting both commercial and open-source development resources.

For practitioners, this market validation means the open-source and local inference ecosystem will continue receiving substantial investment and innovation. It also suggests that choosing open models for local deployment is increasingly a mainstream decision rather than an alternative approach, with corresponding improvements in tooling maturity and model quality.

Read the full article on finance.biggo.com.

Source: finance.biggo.com · Relevance: 9/10