IBM Expands Collaboration and Leverages On-Premise Data Centers for AI

IBM IBM is expanding its push to deliver AI inside customers' own data centers, leaning into the concept of 'data gravity,' the tendency of enterprise data to stay on-premise due to egress fees, latency, security, and regulatory constraints . Rather than pulling workloads to the public cloud, IBM is bringing AI to where the data already lives.

The company is using its Power servers and Z mainframes to deploy AI capabilities directly within clients' private infrastructure . The approach contrasts with a cloud-first model and competes with hyperscalers, even as rivals such as Oracle, CrowdStrike and Seagate also court demand for on-premise and hybrid solutions.

IBM's bet is that regulated, data-heavy enterprises will increasingly want AI without moving sensitive data off-site, playing to its installed base and services depth. Analysts will watch whether this hybrid strategy translates into durable software and infrastructure revenue, or whether workloads ultimately gravitate back to public cloud.