IBM’s third quarter answered the question its second quarter raised. Revenue rose 4.7 percent to 17.10 billion dollars, beating consensus of 16.87 billion dollars, and adjusted earnings rose 11 percent to 2.95 dollars per share, beating the 2.88 dollar consensus. In July the company had warned that large software transactions slipped as clients shifted spending to supply constrained servers, storage and memory. This quarter, chief financial officer James Kavanaugh said about two thirds of those delayed deals closed, contributing roughly a quarter of the sequential acceleration in software growth.
Software, the largest segment, grew 8.2 percent to 7.80 billion dollars, with Hybrid Cloud up 11 percent on Red Hat, Data up 20 percent at constant currency helped by Confluent, and Transaction Processing rebounding 10 percent after a soft quarter. Software annual recurring revenue reached 24.9 billion dollars, up 8 percent, OpenShift recurring revenue passed 2.4 billion dollars, up more than 30 percent, and software segment profit rose 15.2 percent to 2.73 billion dollars. Consulting was flat at 5.3 billion dollars. Infrastructure rose 4 percent to 3.7 billion dollars, with distributed infrastructure up 31 percent and IBM Z mainframe revenue down 20 percent, a decline the company attributes to the launch cycle after the strongest start to a Z programme in nearly two decades.
Free cash flow for the quarter was 3.0 billion dollars, and management raised the full year outlook: revenue growth of 4 to 5 percent at constant currency, unchanged, and free cash flow growth of more than 1 billion dollars year on year. The board approved a quarterly dividend of 1.69 dollars per share. The shares rose 3.94 percent in the regular session on Wednesday and fell in after hours trading despite the beat, a reminder that after July’s collapse the stock is being held to a prove it standard.
On the call, chief executive Arvind Krishna pointed to sovereign AI as a growing tailwind, arguing that governments and regulated companies increasingly want AI deployed privately, on premise, where no traces of their data sit in a public model’s logs. That is the demand pool IBM’s mainframe and hybrid cloud business is built to serve, and it is the part of the AI market that does not depend on hyperscaler capital spending.