IBM Q2 2026 Earnings Analysis

And Alex, this is a juicy one — there was already a pre-announcement before the official call, so expectations were... managed, let's say.

Solid. Okay, let's talk numbers. Revenue growth of just 1% for the quarter — that's a real miss for IBM, which has been on a growth acceleration story for a while now.

Exactly. CFO Jim Kavanaugh explained that in the final weeks of June, clients suddenly redirected spending toward servers, storage, and memory — basically rushing to lock in infrastructure ahead of expected price increases. That caused tens of large enterprise license deals to slip.

Which sounds alarming, but context matters — that's coming off 70% growth in the same quarter last year during the z17 launch. And even with the dip, the z17 cycle is still running at nearly 130% program-to-program versus the z16 cycle, which was IBM's best ever.

That's the bet management is making. And here's a number I liked — about 80% of IBM's software revenue is recurring, subscription or consumption-based, and that grew nicely, accelerating from 7% to 8% year over year. The real damage was isolated to that other 20%, the transactional, CapEx-sensitive chunk.

Which is honestly the more impressive headline given the revenue miss. Kavanaugh pointed to strong underlying EBITDA growth, plus they front-loaded a lot of the cash tax and CapEx headwinds earlier in the year, so those pressures ease up in the second half.

And this is the part analysts kept pushing on during Q&A — is this demand deferred, or demand destroyed? Arvind's answer was pretty specific: of the roughly ten large deals that slipped, about a third had already closed in just the first three weeks of the new quarter. He called that "a good indication, not yet full evidence" that clients are delaying, not canceling.

On the strategic side, there's a lot going on beyond the mainframe story. Distributed infrastructure — think Power and storage — had its best quarter of revenue growth on record, up 37%, with $500 million in backlog, the highest ever.

There's also this new security play called Lightwell, which came out of a vulnerability event called Mythos. IBM's selling access to remediated open-source packages for a million dollars a year, and they've already signed up major banks — Bank of America, Citi, JPMorgan, Goldman, Visa, and others. Arvind pegged the total addressable market at multiple billions of dollars.

One more interesting nugget from Q&A — data segment revenue within software actually underperformed, even with Confluent's strong start. Arvind's explanation was basically that when clients have a fixed budget and something like Confluent looks more immediately valuable, data software gets deprioritized in the short term, but he expects it to recover.

I'd frame it this way — this is a company navigating a timing disruption in its highest-margin, most incumbent business, the mainframe-software bundle, while its newer growth engines — subscription software, infrastructure, consulting — are actually accelerating. Management is leaning on productivity initiatives to protect margins and cash flow even as top-line growth stumbles.

Before we wrap — everything discussed is AI-generated analysis for educational purposes. Past performance doesn't guarantee future results. Please do your own due diligence.

Thanks for listening to Beta Finch — we'll catch you next time.