Before Tuesday's open, IBM did something companies almost never do: it pre-announced preliminary Q2 2026 results eight days ahead of its scheduled July 22 report, paired with an unusually confessional letter to investors from CEO Arvind Krishna. The stock cratered ~25% to ~$216 from a $290.23 prior close — pacing its worst single session since October 19, 1987 (Black Monday), when it fell 23.7%.
| Metric | Q2 Preliminary | Consensus / Comparison | Read |
|---|---|---|---|
| Revenue | $17.2B (+1% YoY) | $17.86B (LSEG) | Miss by ~$660M (3.7%) |
| Adj. EPS | $2.93 | ~$3.01–3.02 | Miss ~3% |
| GAAP EPS | $2.27 | — | — |
| Software | +5% | +11% in Q1 '26 | Growth halved |
| Red Hat | +11% | — | Bright spot |
| Consulting | ~flat (+1% cc) | +1% cc in Q1 | No acceleration |
| Infrastructure | −7% | +12% cc in Q1 | Mainframe air pocket |
| Op. gross margin | 59.4% | −70 bps | Slightly softer |
| H1 free cash flow | $4.76B | $4.80B a year ago | Essentially flat |
"In the last few weeks of June, we saw clients shift their quarterly capex spend toward servers, storage, and memory purchases to secure supply-constrained infrastructure ahead of expected price increases… we did not anticipate the magnitude of the capex reprioritization. This quarter we faltered. We did not adapt and move quickly enough, and numerous large deals failed to close on the timelines we expected." — Arvind Krishna, letter to IBM investors, July 14, 2026
Krishna also told CNBC's Sara Eisen that Anthropic's new "Mythos" AI cybersecurity model is making enterprise buyers "pause on new deals until they know" how much to spend on cyber — while insisting "we don't see our software being disrupted by AI at all."
IBM operates five segments: Software (~45% of revenue, ~two-thirds of profit; 82.8% gross margin in Q1), Consulting (~27.5% margin), Infrastructure (IBM Z mainframes, Power, Storage; ~57% margin), Financing, and Other. The multi-year strategy under Krishna has been to shed the "legacy mainframe" label and become the hybrid-cloud + enterprise-AI plumbing vendor — the governance, data, security and orchestration layer that regulated enterprises use to run AI across clouds. That pivot has been fueled by acquisitions: Red Hat, HashiCorp, and — closed March 2026 for ~$11.6B — Confluent.
This is the crux of my analysis and the reason the reaction is rational, not hysterical. Total software grew 5%. But BofA estimated Confluent alone would add ~$340M of Q2 revenue. Work the bridge:
| Software growth bridge (est.) | Value |
|---|---|
| Q2 2025 software revenue | $7.387B |
| Q2 2026 implied by +5% | $7.756B |
| Total YoY increase | ~$369M |
| BofA est. Confluent contribution | ~$340M |
| Confluent share of the increase | ~92% |
| Implied organic software growth | ~0.4% |
IBM entered 2026 guiding to double-digit full-year software growth. Its June-2 all-time high of $332.46 was investors paying up for a high-margin software engine to accelerate. If ~92 cents of every incremental software dollar is bought (Confluent), not organic, the quality of the growth — and the multiple it deserves — collapses. Software is the profit center; when it stalls, a $660M revenue miss legitimately erases ~$68B of equity value because it re-rates years of forward cash flow, not one quarter.
Not everything broke. Red Hat +11%; Distributed Infrastructure +37% (best on record) with ~$500M backlog; z17 mainframe program running at ~130% of the comparable z16 cycle; generative-AI "book of business" >$12.5B (note: this is cumulative signings, not ARR). Management framed slipped deals as deferred, not lost. These argue against a demand collapse — but they don't resolve the organic-software question.
The tape today was risk-on everywhere except software: a shockingly cool June CPI (headline +3.5% YoY vs 3.8% exp; −0.4% MoM, biggest drop since April 2020; core ~2.6%) sent the Nasdaq +1%, chips ripping (MU +4.6%, SNDK +5%, NVDA +2.5%, AMD +4.8%, SK Hynix ADR +15–17%). Against that backdrop, IBM's collapse dragged software specifically:
| Ticker | Move (intraday) | My read |
|---|---|---|
| ServiceNow (NOW) | ~−5% ($105.79) | Overdone — SaaS demand is not IBM's problem |
| Microsoft (MSFT) | ~−1.6% ($384.79) | Noise; buyable weakness |
| Oracle (ORCL) | ~−2.1% ($128.74) | Already on my SHORT/AVOID list |
| Accenture (ACN) | ~−1.7% ($136.16) | Consulting crowding-out is the real tail risk here |
| IGV (software ETF) | ~−4% | Single-name shock traveling through a concentrated fund |
My distinction: IBM's miss is 60% idiosyncratic (mainframe/Z air pocket + acquisition-dependent software + botched deal execution) and 40% a real macro signal — the AI-hardware capex supercycle is crowding out other IT budgets. Krishna's own words: clients bought servers/storage/memory ahead of price hikes and postponed software. That is not "AI automating away software"; it's "AI hardware eating the wallet." The memory names IBM blamed (MU, SK Hynix, SNDK) are the ones soaring today. The budget is real — it just went to silicon.
No red-flag informed selling into the print. The late-June Form 4s were routine annual director equity grants (code A at $281.21, Jun 30) and an officer's option-exercise/tax-withholding (code M/F ~$286.72, Jul 1) — mechanical, not discretionary open-market dumping. Prediction markets (Polymarket) price only a ~25.5% probability IBM beats on July 22, and Wall Street targets span an extraordinary $191 (HSBC Reduce) to $375 (Citi Buy) — a 96% spread that tells you the Street itself has no idea whether this is timing or structural.
| Firm | Rating | Target |
|---|---|---|
| HSBC (downgrade) | Reduce | $191 |
| UBS | Neutral | $236 |
| Morgan Stanley | Equalweight | $293 |
| RBC | Outperform | $300 |
| Susquehanna | Neutral | $303 |
| Barclays / Oppenheimer | Overweight / Outperform | $350 |
| Citi | Buy | $375 |
Note: most targets predate the warning and will be revised down. HSBC's $191 is a sum-of-parts that values quantum separately (~$35B) — meaning the core business is being marked toward the low $150s.
| Scenario | Prob. | Trigger on July 22 | Stock |
|---|---|---|---|
| Timing problem | 35% | Deferred deals close post-quarter, FCF guide held, organic software improves | Relief rally to $250–265 (gap-fill) |
| Valuation reset (base) | 45% | Some deals close, FCF guide trimmed modestly, organic software stays soft | Range $205–235; grind |
| Structural warning (bear) | 20% | Deals lost, organic software flat/negative, FCF cut materially | $191 → low-$150s core value |
Fair value: ~$235–250 in a base case (13–14x a slightly-trimmed ~$15B FCF guide, plus a modest quantum option). That's above today's ~$216 — but the path there runs directly through a binary July 22 event, and I am not willing to underwrite "the deals were just delayed" with my capital before management proves it.
WATCH SPECULATIVE IBM — do not chase the knife.
Sources (primary & key): Arvind Krishna's letter to IBM investors; CNBC; INDmoney breakdown (segment/FCF/analyst tables); Investing.com (HSBC downgrade); 24/7 Wall St. (historical/contagion); IBM 2Q26 earnings event (July 22, 5:00pm ET). Prices intraday July 14, 2026 via Finnhub.
This is analysis and opinion for informational purposes only — not personalized investment advice. Do your own due diligence.