IBM has announced a new generation of Power11 chips focused on energy efficiency and security, aimed at facilitating the deployment of artificial intelligence for inference tasks in sectors such as financial services and manufacturing.
Introduction: A major update to the Power line
International Business Machines (IBM) on Tuesday announced a new line of data center chips and servers that it says will be more energy efficient than competitors and streamline the rollout of artificial intelligence in business operations. IBM introduced the new Power11 chips on Tuesday, marking the first major update to its “Power” chip line since 2020. These chips have traditionally competed with offerings from Intel and AMD in data centers, particularly in specialized sectors such as financial services, manufacturing, and healthcare.
Focus on reliability and security
Like Nvidia’s AI servers, IBM’s Power systems are an integrated bundle of chips and software. Tom McPherson, general manager of IBM Power systems, said the Armonk, New York-based company used this tight coupling to focus on reliability and security.
Power11 systems, available starting July 25, will not need any planned downtime for software updates, and their average unplanned downtime each year is just over 30 seconds. IBM said they are also designed to detect and respond to a ransomware attack within a minute—where hackers encrypt data and then attempt to extort a ransom for the encryption keys.
Simplifying AI for inference
In the fourth quarter of this year, IBM plans to integrate Power11 with Spyre, the AI chip it introduced last year. McPherson said IBM is not aiming to compete with Nvidia in helping create and train AI systems, but is instead focused on simplifying the deployment of AI for inference, which is the process of putting an AI system into action to accelerate a business task.
McPherson said in an interview last week, referring to work with early customers: “We can seamlessly integrate AI capabilities into this to accelerate inference and help improve their business processes. It won’t have all the horsepower for training or anything like that, but it will have very good inference capabilities that are easy to integrate.”
A different strategy in a crowded market
IBM’s approach represents a different strategy in the AI market. Instead of trying to build the largest and most powerful models for training, the company is focusing on making AI practical and accessible for its enterprise customers. IBM recognizes that many companies do not need to build their own models from scratch, but rather need reliable and secure ways to use pre-existing models to improve their operations.
This focus on inference, security, and reliability could attract customers in heavily regulated industries that prioritize business continuity and data protection over raw performance. By offering an integrated bundle that makes AI deployment easier and safer, IBM hopes to carve out a niche for itself in this competitive space.
Conclusion: A smart step toward practical AI
The launch of IBM’s new Power11 chips and servers represents an important and deliberate step. Rather than entering into direct competition with Nvidia in the costly training race, IBM is focusing on a historical strength of its own: serving the needs of large enterprises with reliable and secure solutions. By facilitating the inference process, IBM may find a receptive audience among companies looking to reap the benefits of AI without the complexities and risks associated with building their own infrastructure from scratch.
