$IBM fell 25% because the market finally priced the middle. This was not just an earnings miss; it was a repricing of a strategy built on being the bridge between old enterprise IT and whatever comes next. IBM pre-announced Q2 revenue of $17.2 billion, up just 1%, with software up 5%, consulting flat and infrastructure down 7%, then admitted it failed to adapt as clients redirected capex toward servers, storage and memory. The strategy made sense—until the clock sped up. Arvind Krishna’s hybrid-cloud bet assumed large enterprises would keep critical workloads on IBM systems while adding public-cloud capabilities around them. That gave IBM a logical role: monetize mainframes and transaction software, use consultants to manage complexity, and sell Red Hat as the connective layer across private and public infrastructure. The problem is that AI is now changing both the destination and the migration economics. AI hits IBM twice. Corporate tech budgets are being pulled toward compute, storage, networking and cybersecurity, leaving less room for slower-moving software projects and legacy upgrades. At the same time, tools such as Claude Code can automate parts of the exploration and analysis behind COBOL modernization—the exact friction that helped make IBM’s installed base so sticky. AI therefore competes for the customer’s next dollar while lowering the cost of escaping the customer’s previous IBM decision. Then there is quantum. IBM has committed more than $10 billion to quantum and targets a large-scale, fault-tolerant system in 2029, which may be strategically valuable but does nothing to repair today’s revenue mix. This creates a brutal duration mismatch: yesterday’s businesses fund the company, today’s AI cycle reallocates spending away from parts of them, and tomorrow’s quantum payoff remains years out. We have seen the same incumbent trap in media, music and autos: management protects the old cash engine while joining the new cycle without fully cannibalizing itself. The middle feels prudent internally, but markets eventually punish it because it produces too many priorities and no obvious source of leadership. Bottom line: IBM needs to choose a smaller number of markets where it can genuinely lead—and accept that protecting every legacy revenue stream is no longer a strategy.