arvind krishna

Mainframe remains architecturally superior for workloads needing protection and burst capacity at 5-15x lower unit cost vs alternatives, with 85% of Z-17 buyers increasing capacity; software portfolio shifting to consumption model to accelerate growth; quantum computing enters engineering realm with 100M-computation machine targeted by 2029 and trillion-dollar TAM by 2030s, reinforced by HRL acquisition.

Mainframe demand repricing is a capex reprioritization, not structural decline—Z-17 at 130% capacity growth, 85% of buyers increasing capacity, and mainframe 5-15x cheaper per unit for resilience workloads. Quantum computing is now in engineering realm: 100M computations by 2029, $1T total market by 2030s, accelerated by HRL acquisition.

IBM has successfully pivoted to hybrid cloud and AI software (nearly half of revenue), spun off declining managed infrastructure services, acquired Red Hat and Confluent to strengthen portfolio, and is integrating AI into mainframes and building enterprise-focused AI solutions.

Confluent's Kafka backbone is the best infrastructure for moving and exposing real-time data to AI systems, solving the critical challenge of making data available for AI workloads across enterprises.