IBM's Stock Just Had Its Worst Day Since 1987. The Real Story Is Where the Money Went.

News Breakdown · FiscEdge Academy

IBM just had its worst trading day since Black Monday. The company's preliminary Q2 numbers, released ahead of its July 22 earnings call, showed revenue of $17.2 billion, short of the $17.86 billion Wall Street expected, with adjusted EPS of $2.93 against a $3.01 consensus. Shares plunged as much as 25% in premarket trading on July 14, the steepest single-day drop for the 115-year-old company since October 19, 1987, when the stock fell 23.7% on Black Monday.

The miss itself is small in percentage terms, a shortfall of roughly 3.7% on revenue. The size of the stock reaction is the tell. Investors weren't just pricing in a soft quarter, they were pricing in a story CEO Arvind Krishna told them directly: "numerous large deals failed to close" because clients redirected their quarterly capital spending toward servers, storage and memory, racing to secure supply-constrained AI infrastructure before expected price increases hit. IBM's own infrastructure segment, including its z-series mainframes, declined roughly 7% as a direct result.

The number that matters isn't the miss, it's the reallocation

Enterprise buyers didn't shrink their budgets. They moved the money. Software and consulting contracts that were already in procurement got shelved so IT departments could lock in hardware ahead of anticipated shortages and price hikes in DRAM, NAND and server components. That is a capex reallocation happening inside real enterprise budgets, in real time, and IBM's earnings warning is the clearest evidence yet that it's large enough to move a Dow-component stock double digits before the opening bell.

For founders selling into enterprise, this is the signal to actually plan around, not the headline number.

What this changes for founders and operators

• Enterprise sales cycles are getting reprioritized around hardware, not paused. If a mid-market or enterprise deal you were counting on this quarter suddenly goes quiet, don't assume it's dead. IT buyers are triaging: hardware now, software later. Build that into your pipeline forecasting rather than writing the deal off.
• Component costs are the real story to watch. If enterprises are stockpiling servers, storage and memory ahead of expected price increases, any founder building AI-heavy products, whether that's inference-heavy SaaS or anything touching GPUs, should expect the underlying infrastructure to get more expensive, not less, over the next few quarters.
• The market is pricing legacy tech companies on their AI story, brutally. A 3.7% revenue miss wiping out roughly a quarter of a company's market cap in hours tells you how little patience public markets have left for "we're transitioning to AI" without hard numbers behind it. If you're modeling a future raise or exit on comparables, assume the bar for proof, not narrative, keeps rising.
• "Picks and shovels" demand is still the safest trade in this cycle. The same dynamic that hurt IBM (customers chasing physical AI infrastructure over software) is the dynamic that's been inflating valuations for chipmakers, memory makers and infrastructure providers all year. If your roadmap touches AI infrastructure at all, that's still where the capital is flowing fastest.

Why this is bigger than one earnings call

IBM is a bellwether for enterprise IT spending broadly, and this quarter it just told the market that AI hardware scarcity is now strong enough to break sales cycles at a company that size. That's not a one-off supply hiccup, it's a preview of how every enterprise software vendor's next few quarters could look if the AI hardware crunch continues: deals slip, not because customers don't want the software, but because they're spending the budget line securing the compute underneath it first.

If you remember one thing

When an enterprise deal goes quiet right now, the first question shouldn't be "did we lose it," it should be "did the budget just get redirected to hardware." The capex reallocation IBM disclosed today is happening across enterprise IT broadly, and founders who model for it in their pipeline and pricing will out-forecast the ones who don't.

---

We break down how to read market-moving earnings and capex shifts like this one in FiscEdge's financial modeling course, and how to build pipeline forecasts that survive enterprise budget reshuffles in startup strategy. If you're building anything that touches AI infrastructure, our AI for entrepreneurs track covers where the capital is actually flowing. Curious how these capex numbers connect to your own unit economics? Start with what are unit economics. Follow @fiscedge for daily Business & AI analysis.