What I Would Say to IBM’s Leadership After the Big Drop
IBM shares fell roughly 25% on July 14, the worst single-day decline in the company’s history, after it pre-announced second-quarter revenue of $17.2 billion, up only 1% and short of expectations.
Is this further proof of the SaaS-pocalypse, or something else, given how much of IBM’s revenue sits in areas exposed to AI disruption, such as COBOL workloads and professional services? That call is not my lane. What I do know is IBM itself. I spent nine years there, five more at Red Hat before IBM acquired it, and IBM remains an MI&S client today; I work with its team every week. Here is the advice I would give its leadership, organized around five moves.
Why IBM Software Needs a Sharper Story, Not Just Better Retention
Software is not where the immediate damage sits, but growth alone will not carry the business through this cycle. IBM Software tends to run like a private equity portfolio: growth through acquisition, profit through operational rigor and customer retention. That model is not broken, but every vendor, SaaS vendors especially, now has to explain not only why it stays relevant in the AI era but where that relevance is heading. IBM’s automation products already tell a solid story about diagnosing and remediating today’s problems autonomously. I made a version of this argument in my own research this spring: the technical pieces are finally starting to cohere, but the go-to-market story connecting them has not caught up. What is missing is the next chapter for these product lines: for instance, predicting customer deployment problems before they happen and before a more advanced AI models catch up.
Where IBM’s Software Go-to-Market Must Get Ruthless
While the software group works on that longer story, its go-to-market motion needs to narrow: put more weight behind what is working, and cut, consolidate or partner around, what is not. Does IBM need its own IDE, or would customers be better served by a partnership with a category leader? Like most software businesses, IBM carries too many products, and it cannot afford to leave that unaddressed.
Why IBM’s Infrastructure Business Needs Less Vision Right Now
This may be an uncomfortable message for the infrastructure team, but customers today are struggling to source memory, accelerators, and servers. They do not need a briefing on quantum computing timelines for 2028 or 2030; that is a long way off given where the market stands. The priority should be solving today’s supply and delivery problems, not showcasing tomorrow’s roadmap. My colleague Matt Kimball made a related point just a day before the pre-announcement: the new rack-mountable z17 and LinuxONE 5 systems can ship and install in days, not the months x86 gear takes in today’s supply crunch. That is the kind of near-term, execution-first move infrastructure needs more of.
How IBM Can Turn AI from a Liability into Its Stickiest Advantage
IBM’s AI positioning has read as defensive more often than it should, even though the underlying work, including Granite, has been genuinely interesting, if arguably ahead of its time. My take is that IBM’s AI story has tracked what the market wants to hear rather than what IBM does best. Consider code migration: those tools are evolving into agents that keep systems continuously current, and a system that stays current is harder to migrate away from. IBM should use AI agents to make its mission-critical systems difficult to leave, turning a migration risk into a retention advantage, provided those agents can be trusted not to introduce risk of their own.
What IBM Consulting’s Agent Experiments Could Become
Consulting is the unit under the most visible pressure right now. IBM Consulting has still done credible work building agents and reusable delivery assets over the past few years. I have no specific recommendation beyond urging IBM to keep experimenting; there may be something durable there. My honest suspicion is that this work may only slow the consolidation already underway in technical services. Then again, for an acquisitive company, that disruption could become an opportunity: IBM strengthened its services position with the PwC Consulting acquisition the last time this industry took a comparable hit.
What IBM Still Needs to Prove
None of this is a quick fix. Each move takes time, patience, and further investment, and I have left out shorter-term fixes on the assumption that IBM’s leadership is already working through those ahead of its full earnings call on July 22. Krishna already owns the immediate problem: in his July 14 letter to investors, he wrote plainly, “We did not adapt and move quickly enough, and numerous large deals failed to close.” That candor is a start, but it is not a strategy. My verdict: IBM has a real window to gain share in a disruption cycle that could prove larger than the shift to distributed computing and the cloud, but only if it treats software vision, infrastructure focus, and AI positioning as three different problems requiring three different postures. Getting that sequencing wrong is what would undercut the case laid out here.
