IBM Launches Power11 Chips to Simplify Enterprise AI Integration

On July 8, IBM unveiled its new Power11 chips and a line of accompanying servers, aiming to deliver more power-efficient, secure, and AI-ready infrastructure for enterprise data centers. Available starting July 25, the Power11 systems are designed to ease the deployment of artificial intelligence into real-world business workflows.

Focus on AI Inference

Unlike Nvidia’s offerings geared toward AI model training, IBM's Power11 systems are built to accelerate AI inference, the process of applying existing models to real-time tasks.

“It’s not going to have all the horsepower for training,” said Tom McPherson, general manager of IBM Power Systems, “but it’s going to have really good inferencing capabilities that are simple to integrate.”

Integrated Systems with Security and Uptime in Mind

IBM emphasized the reliability of its tightly integrated hardware and software stack. The Power11 servers will feature zero planned downtime for software updates and average just 30 seconds of unplanned downtime per year. They can detect and react to ransomware attacks in under a minute, McPherson noted.

AI-Ready, with Future Integration Ahead

IBM plans to pair Power11 with its Spyre AI chip later this year to further boost inferencing capabilities. The systems are expected to serve industries like finance, healthcare, and manufacturing, competing with chips from Intel and AMD in these performance-intensive sectors.