IBM says AI delayed software deals, not killed them—a third already returning

IBM's Q2 earnings call became an extended defense of the company's software business after preliminary results triggered one of its steepest stock declines in years. The trigger: a notable slowdown in large capex software deals at enterprise customers, which rattled investors who feared spending had permanently shifted away from IBM's core software franchise.

CEO Arvind Krishna's central claim was that customers had not abandoned software purchases but rather emptied their budgets on AI infrastructure first. "The majority of what didn't happen in the second quarter was large capex deals at large clients," Krishna said. He pointed to an early recovery signal: "We have been very pleased to see that about a third of those have already closed." He cautioned, however, that this represented "indication, not yet full evidence" that the slowdown was truly temporary. The distinction mattered because Wall Street remained skeptical; during the Q&A, analyst Amit Daryanani from Evercore framed the core debate as "is this demand deferred or destroyed for IBM?"—a question that Krishna continued to answer by insisting enterprise customers had merely temporarily diverted spending toward AI infrastructure rather than fundamentally changing their priorities.

Rather than rest on a defensive argument, IBM spent the second half of the call reframing AI as an opportunity. The company announced Project Lightwell, a new service aimed at enterprises struggling with security vulnerabilities in aging open source software. Krishna noted that Anthropic's release of Mythos in early April had "accelerated the discovery of security vulnerabilities for clients," creating what he called a "multibillion-dollar" total addressable market. Lightwell uses AI to remediate and validate open source packages that enterprises continue to rely on long after community maintainers have abandoned them. The service costs $1 million（約1.6億円） per year per customer. Early adopters already signed up include eight major financial services firms: Bank of America, Citi, Goldman Sachs, JPMorgan Chase, Mastercard, Morgan Stanley, Visa, and Wells Fargo. Krishna framed the initiative as a bet on both AI and open source expertise, positioning it as "a multiple billion-dollar opportunity" that IBM intends to pursue "really fast and hard."

IBM faced a sharp market reaction after Q2 preliminary results revealed a significant software sales slowdown, prompting the company to use its earnings call largely as damage control. CEO Krishna's core argument—that customers had not fundamentally abandoned software but had temporarily redirected capital to AI infrastructure—was designed to combat investor fears that the software dollar pipeline had permanently shrunk. The fact that roughly a third of delayed deals had already closed by early Q3 was presented as early validation of this "deferral, not destruction" thesis, though Krishna himself acknowledged the sample was still small.

Wall Street remained unconvinced enough to press hard on the distinction between deferred and destroyed demand, suggesting the market saw real risk in Krishna's narrative. To shore up confidence, IBM pivoted to positioning AI not as a threat to its software business but as the foundation for a new, adjacent revenue stream. Project Lightwell exemplifies this strategy: it sells enterprises AI-powered remediation for the security vulnerabilities that AI tools like Anthropic's Mythos (released in early April) are helping to discover in legacy open source codebases. By framing the opportunity as "multibillion-dollar" and securing commitments from major financial services firms including Bank of America, JPMorgan Chase, and Goldman Sachs, IBM is attempting to redefine the conversation from lost software revenue to AI-enabled growth.

No comments yet. Be the first to share your thoughts!