IBM reported first-quarter revenue of $14.54 billion, up 2% year over year at constant currency, with software revenue up 9% to $6.34 billion and free cash flow of $2.0 billion for the quarter. [1] The headline print beat consensus. The stock traded down 1.4% in the post-market anyway, because the one number Wall Street actually wants from IBM — the cumulative generative AI book-of-business — came in at $13.5 billion, a $1.0 billion sequential addition versus the $2.0 billion Q4-to-Q4 average the Street had modeled. [2]

That is the entire earnings story. Software is doing what software is supposed to do. Consulting grew slowly; infrastructure contracted as expected ahead of the z17 refresh. The AI question is not resolving on this print.

The deceleration is arithmetic, not narrative. IBM's GenAI book runs two-thirds consulting-signings, one-third software-and-infrastructure bookings. The consulting leg is the capacity-constrained piece: each engagement requires consultants, and IBM's utilization is already elevated. The software leg was supposed to compensate, but watsonx subscription momentum flattened quarter over quarter. [1] When Arvind Krishna said on the call that the GenAI pipeline remains "robust and growing," he was telling the truth about the pipeline and saying nothing about the conversion rate.

Bloomberg's lead framed it precisely: "in-line software sales that fail to shake AI concerns." [2] The fail-to-shake framing is the right one. Nothing in the print confirms the bear case that IBM's AI strategy is structurally limited. Nothing in the print disconfirms it either. Q2 is the reporting period when the watsonx-x cloud-provider deals announced in late 2025 should begin booking at scale. If they do not, the bear case stops being a question and starts being a thesis.

Free cash flow guidance for 2026 was reaffirmed at $13.5 billion, and the company kept its full-year revenue outlook at 5% constant-currency growth. [3] The buyback was modest — $1.4 billion in the quarter — and the dividend coverage is uncontroversial. For income investors, the print is fine. For anyone buying the stock for AI exposure, the print is a shrug.

The peer comparison does not help. Accenture reported last month with consulting growth that underperformed IBM's services segment; Oracle's cloud-applications AI book is smaller in absolute dollars but expanding faster in percentage terms. [2] IBM's $13.5 billion cumulative book is still the largest claimed GenAI franchise in enterprise IT. It is also the slowest-growing one relative to its own expectations. That is the compression the headline buried, and it is why the stock did not rally on a beat.

What the paper will watch in Q2: sequential software deceleration or reacceleration, consulting-segment bookings split between AI and non-AI, and any change to the cumulative GenAI book's reporting methodology. Methodology changes would be a tell.

-- THEO KAPLAN, San Francisco