IBM revenue rises despite concerns over AI

IBM is not being priced like a classic secular-growth software name anymore; it is increasingly a monetization story on installed-base inertia. The key second-order effect is that AI appears to be a net revenue enhancer for the legacy franchise in the near term, because modernization budgets tend to get funded before replacement budgets do. That means AI disruption may actually lengthen the life of the mainframe cash engine rather than shorten it, pushing the pain point from revenue loss to margin/valuation debate over a 12-24 month horizon.

The market is likely underestimating the asymmetry between headline software concern and infrastructure resilience. If mainframe refresh cycles hold while Watsonx-type tools accelerate migration work, IBM can capture spend from both sides of the modernization trade: tooling revenue plus hardware pull-through. The competitive loser is less IBM specifically than pure-play legacy software vendors whose products are easier to unbundle once CIOs buy into AI-driven process automation.

Near term, the real risk is not AI replacement but budget timing: enterprise customers can delay software renewals for a few quarters while they evaluate AI alternatives, which can cap multiple expansion even if revenue stays stable. A sharper downside catalyst would be evidence that modernization tools are displacing IBM’s own software attach rates faster than mainframe hardware can offset, because then the market will re-rate IBM as a melting-ice-cube asset with a temporary AI narrative. On the upside, any proof that AI-assisted modernization increases mainframe utilization should support a multiple expansion over the next 1-2 quarters, since investors will be forced to revise down terminal decline assumptions.

The Middle East noise looks secondary unless it becomes a logistics or customer-credit issue; for IBM, the larger catalyst stack remains product-cycle evidence and AI attach rate commentary over the next 1-2 earnings prints. The contrarian view is that the consensus is treating AI as uniformly disintermediating, when in this case the most likely first-order outcome is productivity lift embedded inside the incumbent vendor’s workflow. That makes the current setup more of a sentiment overhang than a fundamental deterioration story.