IBM CEO: open-weight models deliver 90% performance at 10% cost on premises

IBM's CEO has stated that open-weight AI models now deliver enough performance for enterprise use at a fraction of the cost of closed proprietary systems, requiring only two or three server racks to deploy on premises. The CEO framed this as ninety percent of the performance at ten percent of the price, making the economics of in-house AI infrastructure increasingly attractive compared to cloud-based alternatives.

This reflects a broader industry trend toward smaller, more efficient models that organizations can control and run themselves, moving away from reliance on expensive external AI services.