IBM

Executive Summary

Deferral or destruction? One print decides — and it isn't until October

IBM trades at $223.65 after the worst single day in its 115-year history — a 25.2% collapse on July 14, 2026, when management pre-announced a Q2 miss eight days early. The damage was concentrated, not diffuse: IBM Z revenue fell 42% and Transaction Processing fell 8% as enterprise clients rerouted late-June budgets into supply-constrained AI servers, storage and memory. Management calls this deferral, not destruction — roughly a third of the slipped deals closed within three weeks of quarter-end, and ~70% of global transactions by value still run on IBM Z. The software mix is still compounding underneath: Data & AI +19%, Red Hat +11% on $7.8B of software revenue, with non-GAAP operating margin up 0.3pt to 19.2% and the FY26 free-cash-flow guide of roughly +$1B YoY left intact. But the market now pays 17.5× forward earnings for +1% revenue growth — no discount to Oracle, which is growing 21%. Nothing verifies the deferral thesis until the October 21 print, and sell-side targets span $191 to $350. We initiate at HOLD.

QUALITY VS. CYCLE

A software-margin business repriced on a hardware quarter — Q2 FY26 actuals (IBM newsroom, Jul 22, 2026) · market data as of Jul 31, 2026 · revenue and EPS per SEC EDGAR 10-Q

Share Price

$223.65

Jul 31, 2026 · 52-week range $199.19–$332.46

Market Cap

$210.7B

EV ~$267.9B — $62.0B total debt vs $7.2B cash at Jun 30, 2026

Q2 FY26 Revenue

$17.162B

+1% YoY (SEC EDGAR) vs ~$17.86B consensus — a ~$700M miss that cost a quarter of the market cap

Q2 FY26 Diluted EPS

$2.27

GAAP per SEC EDGAR, −2% YoY. Non-GAAP $2.93 vs $3.02 expected

Gross Margin

57.7%

Q2 FY26, −1.0pt YoY as high-margin Transaction Processing fell 8%

Operating Margin (non-GAAP)

19.2%

+0.3pt YoY — cost discipline held even through the revenue shortfall

Free Cash Flow

$2.5B

Q2 FY26, −$0.3B YoY; $4.8B for the half (flat), $13.8B TTM. FY26 guide of ~+$1B YoY reaffirmed

Forward P/E

17.5×

vs SAP 20.4× · ORCL 16.1× · ACN 11.6×. Consensus PT $244 (25 analysts, range $191–$350)

Quarterly Revenue — Trailing 4 Quarters (SEC EDGAR)

$16.3B

Q3 FY25

$19.7B

Q4 FY25

$15.9B

Q1 FY26

$17.2B

Q2 FY26

GROWTH & CATALYSTS

The software mix is compounding — the hardware cycle isn't

+19%

Data & AI Revenue (Q2 FY26)

IBM's fastest-growing software line (+18% cc), inside a Software segment of $7.8B, +5% YoY

+11%

Red Hat / Hybrid Cloud (Q2 FY26)

The open hybrid control plane hyperscalers can't easily displace. Automation +4%, but Transaction Processing −8%

$10B

Quantum Commitment Through 2029

Plus $1B IBM cash alongside $1B of Dept. of Commerce incentives for a dedicated quantum foundry; Starling fault-tolerant system targeted 2029

Feb 23, 2026

First mainframe scare — down 13.2%

Shares fell 13.2% (~$31B) on published claims of AI-automated COBOL-to-Java/Python conversion at high accuracy — a direct challenge to the Z installed base and its attached high-margin transaction-processing software. IBM SVP Rob Thomas rebutted that code translation is not modernization. February supplied the thesis; July supplied the evidence.

Jul 14, 2026

The 25% gap — worst day on record

IBM pre-announced preliminary Q2 eight days early: revenue $17.2B vs ~$17.86B consensus, adj. EPS $2.93 vs $3.02. Shares fell 25.2% ($290.23 → $217.07) on ~64M shares, ~551% of the 3-month average — roughly $67B of market cap, worse than Black Monday 1987. Krishna's letter conceded IBM "did not adapt and move quickly enough."

Jul 22, 2026

Full Q2 print confirms it; FY26 guide cut

IBM Z −42%, Infrastructure $3.8B −7%, Consulting $5.3B flat, Software $7.8B +5%. FY26 constant-currency revenue growth trimmed to 4–5% from "more than 5%" guided in April. Free cash flow guidance of ~+$1B YoY notably was NOT cut.

Oct 21, 2026

Q3 FY26 — the deferral test

The entire bull case rests on Krishna's claim that this was timing: "tens of large deals failed to close on the timelines we expected," a third closing within three weeks of quarter-end. Q3 software and IBM Z bookings are the first hard verification. Nothing before this date resolves the debate.

2027–2028

Next Z refresh cycle — the structural cure

z17 reached GA in June 2025 and IBM has historically refreshed IBM Z every 2–3 years, putting the next cycle around 2027–28 — the natural remedy for a −42% comp. The quantum optionality behind it has no near-term revenue disclosure and, by Krishna's own framing, no measurable earnings impact until 2028–29.

Earnings & Signals

Forward Estimates, Surprises & Insider Activity

Forward Earnings Estimates

FY+1 EPS Consensus~$12.33 (FY2026E consensus, 25 analysts, stockanalysis.com) — other trackers cluster $12.42–$12.51; the spread reflects how recently each vendor refreshed post-July 14

FY+2 EPS Consensus~$13.43 (FY2027E consensus, ~+8.4% YoY)

PEG Ratio~2.46 on a 3-year forecast EPS growth rate of ~7.5% — expensive on growth-adjusted earnings even after a 33% drawdown

Forward P/E~17.5× forward consensus (~18.1× on the $12.33 FY26E number); trailing ~19.9× on $11.27 TTM EPS

EPS Revisions (90d)↑0 ↓7 (Sharply negative. Seven individually verified target/estimate cuts between Jul 14–24 (HSBC $231→$191, Argus $360→$280, Evercore $310→$250, Citi $375→$255→$245, JPMorgan $291→$250, Jefferies $320→$260, Morgan Stanley $293→$190) and zero raises found. These are verified firm actions, not a full IBES revision tally.)

Guidance AccuracyBroke badly in July after a clean streak. IBM guided to "more than 5%" constant-currency growth in April 2026, pre-announced a Q2 shortfall on Jul 14, and formally cut FY26 cc growth to 4–5% on Jul 22. Free cash flow guidance was not cut. The prior three quarters all beat — this is a discrete credibility break, not a pattern.

Earnings Surprise Track Record

Q3 FY25 (Sep) Est: $2.44 Act: $2.65 +8.6%

Q4 FY25 (Dec) Est: $4.31 Act: $4.52 +4.9%

Q1 FY26 (Mar) Est: $1.81 Act: $1.91 +5.5%

Q2 FY26 (Jun) Est: $3.02 Act: $2.93 −3.0%

Beat Rate3 of 4 (75%) — on operating/non-GAAP diluted EPS, the basis consensus is set on. Q2 FY26 GAAP diluted EPS was $2.27 per SEC EDGAR. Revenue also missed in Q2 FY26: $17.162B vs ~$17.86B expected.

Insider Activity (90 Days)

Net Buying/SellingApproximately zero — no open-market insider purchases or discretionary sales reported in the three months through Jul 31, 2026, including the entire post-crash window

Sell/Buy RatioNet −$6.6M in total insider transaction value over the trailing 24 months (net selling). Over the last 90 days the ratio is undefined: neither meaningful buys nor sells were logged.

Neutral-to-negative — and the negative comes from an absence rather than an action. A −25% single-day dislocation on a $210B blue chip is the textbook moment for insiders to signal conviction with open-market purchases, and none appeared in the two and a half weeks after the gap. With insider ownership at ~0.09%, there is little personal capital at stake to motivate that signal either way. Trading windows were plausibly closed between the Jul 14 pre-announcement and the Jul 22 results, which partly excuses the silence — but not the week that followed. Watch the window that opens after the Q3 print.

Most 2026 insider activity is compensation mechanics, not conviction: Arvind Krishna acquired 83,028 shares via PSU vesting on Feb 1, 2026 and disposed of 41,133 at $303.755 the same day in settlement, plus February RSU releases with shares withheld for taxes. The one genuine open-market buy on record is director Michael Miebach, 434 shares at $233.33 on Feb 25, 2026 — a token position, and now underwater.

Peer Comparison

Relative Valuation vs. Competitors

Company	P/E (fwd)	EV/Rev	Rev Growth	Gross Margin
IBM	17.5×	3.9×	+1%	58.1%
Oracle	16.1×	7.6×	+21%	65.8%
SAP	20.4×	4.8×	+9%	73.7%
Accenture	11.6×	1.4×	+6%	32.0%
Hewlett Packard Ent.	12.4×	2.1×	+40%	33.8%

IBM is caught between two repricings. The IT-services complex has been savaged — Accenture trades at 11.6× forward on +6% growth as the market discounts AI eating billable hours. Meanwhile the beneficiaries of the very capex shift that broke IBM's quarter are being paid for it: HPE grew revenue 40% YoY selling exactly the servers, storage and memory IBM's clients diverted budget toward. IBM's 17.5× multiple and 58.1% gross margin sit on the software side of that divide; its +1% growth sits on the services side. Bull case: a services multiple attached to a software margin structure. Bear case: 3.9× EV/revenue is still full price for a business whose two disclosed growth engines are being swamped by an $8B+ legacy hardware and transaction-processing base in decline. Growth is latest reported quarter YoY; multiples per stockanalysis.com, retrieved Aug 3, 2026.

Valuation

Price Targets & Scenarios

Scenario	Price Target	Assumptions	Probability
Deferral confirmed — Z bookings snap back	$275	October 21 shows the slipped Q2 deals landing in Q3 and IBM Z bookings recovering toward the record z-cycle Kavanaugh described ("high 120s" versus prior programs). FY26 cc growth lands at the top of the 4–5% guide, free cash flow hits the ~+$1B step-up, and FY27 EPS reaches ~$13.80. The multiple re-rates back toward 20× as the "AI loser" label is retired. Requires management credibility to be restored in a single print.	30%
Partial recovery — timing plus some real demand loss	$232	Some deferred deals close but Z and Transaction Processing run structurally softer than pre-July expectations. Software holds +5–6%, Consulting stays roughly flat, Infrastructure remains a drag until the 2027–28 z-refresh. FY27 EPS ~$13.20 against a justified 17.5× — no better than today's multiple, because credibility takes more than one quarter to rebuild. The 3.0% dividend and reaffirmed FCF guide carry the total return.	45%
Structural — the mainframe moat is genuinely eroding	$178	Q3 shows the deals did not come back and a second guidance cut follows. The February COBOL-modernization thesis gains evidence, Transaction Processing decline accelerates, and Consulting turns negative as AI compresses billable hours. FY27 EPS falls to ~$11.80 on a de-rated 15× as IBM is repriced alongside Accenture rather than SAP. $62B of debt limits the buyback offset, and securities-fraud investigations add an overhang.	25%

Probability-Weighted Target: $231 (~3% upside vs $223.65 — probability-weighted across the three scenarios)

$231

Weighted

Bull $27530%

Base $23245%

Bear $17825%

Analyst Consensus

Morgan Stanley

$190

Equal-Weight — target slashed $293 → $190 (Jul 23)

HSBC

$191

Downgraded Hold → Reduce, $231 → $191 (Jul 14)

Citigroup

$245

Buy maintained — cut twice, $375 → $255 → $245 (Jul 24)

Evercore ISI

$250

Outperform maintained — target lowered $310 → $250 (Jul 16)

14 Buy · 9 Hold · 2 Sell — Consensus is still Buy (14 / 9 / 2) with an average 12-month target near $244, about 9% above the close — but that average is the residue of a July repricing in which at least seven firms cut and none verifiably raised. Dispersion is the real signal: an 83% spread from HSBC's $191 Reduce to Oppenheimer's $350 means the Street has no shared view on whether the Z shortfall is timing or structure.

Technical

Key Levels & Options Intelligence

S/RSupport & Resistance

52-Week High

$332.46

200-Day MA

$270.49

50-Day MA

$260.44

Resistance — 50-EMA

$233.20

Current Price

$223.65

Support — post-gap Fib floor

$213.61

52-Week Low

$199.19

OptOptions & Sentiment

RSI (14-day): 41.4 — below the midline but not oversold; the stock has stabilised in a $205–$228 band since the gap rather than bleeding lower
Put/Call Ratio (30-day OI): 0.61 — open interest still skews to calls despite the crash
Implied Volatility (30-day mean): 41.6% — roughly 2× the pre-crash norm but well off the mid-July peak
IV Rank / Percentile: One-month IV hit its ~99.6th percentile on Jul 15; materially decayed by Jul 31 but still elevated
30-Day Historical Volatility: 108.2% — distorted upward by the single −25.2% session. IV at 41.6% prices normalisation, not repeat gap risk
Volatility Skew (30-day): Call IV 42.7% vs put IV 40.5% = −2.2pt reverse skew. Calls bid over puts after a crash reads as dip-buying, not hedging
Short Interest: 33.09M shares — 3.52% of float. Modest; this was never a crowded short and shorts did not cause the move
Days to Cover: 2.75 days — no squeeze fuel. Any recovery has to come from real buyers, not covering

Conviction Analysis

Systematic Conviction Score: 43/100 (Low)

Analyst Alignment

30%

FCF Visibility

25%

Catalyst Clarity

20%

Valuation Safety

15%

Mgmt Quality

10%

Weighted 0.30 analyst / 0.25 FCF / 0.20 catalyst / 0.15 valuation / 0.10 management = 43, firmly in the Low band. Analyst alignment (28) is the biggest drag: an 83% spread from $191 to $350 means the Street itself cannot agree on whether this is a timing problem or a franchise problem. Management quality (36) reflects a direct hit to forecasting credibility — guiding 5%+ cc growth in April and pre-announcing a miss in July that erased $67B. FCF visibility (52) is the one genuine support: the ~+$1B guide was reaffirmed, though it now depends on an unproven H2 skew. Catalyst clarity (45) is middling by design — the date is precise (Oct 21) and highly informative, but its outcome is binary and nothing verifies before then. Valuation safety (61) is the highest input, and even it is capped by $62B of debt. A Low score is the honest read: this is a coin-flip on one print, and position sizing should reflect that rather than the quality of the underlying franchise.

Risk & Ownership

Risk Assessment & Insider Signals

OOwnership & Insider Signals

Institutional Ownership ~64% — high, and mostly passive: Roughly 63–65% institutional depending on 13F cut date, with insiders near 0.5% and the balance retail. Composition matters more than level: the bulk sits in index and quasi-index vehicles that do not sell on a guidance cut. That is part of why the stock gapped −25% and then stabilised in a $205–$228 band rather than bleeding lower — there was a repricing, not a forced unwind.
Big Three control roughly a fifth of the company: BlackRock is the largest single holder at ~77.6M shares (~8.25%); Vanguard filed a 13G reporting 70,227,012 shares, 7.48% of the class as of Apr 30, 2026; State Street holds ~4.5%. Combined, the three passive managers own about 20% of shares outstanding. None are discretionary sellers — but none are marginal buyers either. They provide a floor of indifference, not support.
Insider ownership is functionally nil — ~0.09%: Management and directors hold a negligible personal stake relative to the $210.7B cap. Executive compensation is delivered through RSUs and PSUs that vest on schedule regardless of the July drawdown, so alignment between the people who set April's guidance and the shareholders who absorbed the −25% is weak.
Float is ~99.8% of shares outstanding — no scarcity mechanics: 940.14M float against 942.13M shares outstanding (SEC EDGAR cover page, Jun 30, 2026). Essentially every share is tradeable. Combined with a 2.75-day short ratio, this is a deep, liquid name with efficient price discovery and no structural constraint that would amplify a recovery. Any re-rating must come from fundamentals.
Dow Jones weight ~2% — an index footprint that just shrank: IBM carries roughly a 2% weight in the price-weighted DJIA. Because the index is price-weighted, the fall from $290.23 to $217.07 mechanically cut IBM's index influence by about a quarter — reducing the passive Dow-tracking bid at exactly the moment the stock needed it. In the cap-weighted S&P 500, a $210.7B cap sits well outside the tier absorbing index flows.
Sell-side positioning is unusually fractured: Post-warning targets span $191 (HSBC, Reduce) to $350 (Oppenheimer), with Citi cutting $375 → $245, JPMorgan $291 → $250, Evercore $310 → $250, Argus $360 → $280 and Morgan Stanley $293 → $190. An 83% spread between low and high target on a mega-cap blue chip signals genuine institutional disagreement on whether the Z shortfall is timing or structure — and is the main reason conviction here is Low.

Risk Heatmap

Quantified Risk Assessment

Severity Risk Factor Prob. PT Impact

High IBM Z / Infrastructure cycle broke down — the trigger for the −25% gap

On Jul 14, 2026 IBM pre-announced Q2 revenue of $17.2B (+1% YoY) vs ~$17.86B consensus and adj. EPS $2.93 vs $3.02, and the stock fell 25.2% — its worst day since at least 1968, roughly $67B of market cap. Confirmed Jul 22: Infrastructure −7% ($3.8B), Consulting flat ($5.3B), Software +5% ($7.8B), gross margin 57.7% (−1.0pt), with IBM Z itself down 42%. Krishna attributed it to clients rerouting quarterly capex into AI servers, storage and memory ahead of price increases. A single quarter's Z shortfall re-rated the entire multiple, which makes the Z cycle the dominant swing factor for the next two prints.

55% -18%

High Guidance credibility impaired — risk of a second cut in October

As recently as April 2026 IBM guided to 5%+ constant-currency growth; on Jul 22 that reset to 4–5% cc with software narrowed to 6–8%. Free cash flow was held at roughly +$1B YoY, but H1 FCF of $4.8B was flat YoY and Q2's $2.5B fell $0.3B — the FCF guide now leans on a large H2 skew that the same Z/software mix just failed to deliver. A second cut on Oct 21 would confirm the market's read that this is structural rather than timing, and is the single largest threat to the base case.

40% -20%

High AI-driven COBOL modernization attacks the mainframe moat

This was IBM's second mainframe scare of 2026. On Feb 23 the stock fell 13.2% (~$31B) on published claims of automated COBOL-to-Java/Python conversion at high accuracy — a direct threat to the Z installed base and the high-margin transaction-processing software attached to it. IBM publicly rebutted that code translation is not modernization. The July Z shortfall makes the February thesis materially harder to dismiss; the two events compound, because the market now has both a narrative and a data point.

30% -25%

Medium Consulting has stopped growing — no offset when hardware cycles down

Consulting was $5.3B in Q2 FY26, flat YoY (+1% cc), and IBM's own warning language put consulting alongside software as deferred by the AI capex reallocation. The segment was built as counter-cyclical ballast to lumpy Infrastructure; with it flat, nothing structurally damps the Z cycle. It is also the segment most exposed if AI coding agents compress the billable-hours model — the same force behind Accenture's de-rating to 11.6× forward.

50% -10%

Medium Securities-fraud investigations opened by multiple firms

Following the July 14 drop, several plaintiffs' firms opened investigations into whether IBM made false or misleading statements about the pace of securing large deals and the strength of the IBM Z outlook — specifically the gap between April guidance and the July pre-announcement. These are investigations, not certified class actions, so direct near-term financial impact is limited. The real cost is the overhang on management credibility at exactly the moment the Z narrative most needs the benefit of the doubt.

45% -6%

Medium $62B debt against a 31-year dividend streak and softening free cash flow

Total debt was $62.0B at Jun 30, 2026 against $7.2B cash (EV ~$267.9B vs $210.7B market cap). The dividend yields ~3.0% on 31 consecutive years of increases — a streak management will defend at nearly any cost. If FCF misses the +$1B guide, the squeeze falls on buybacks, R&D or the Z roadmap rather than the payout. That is precisely the value-trap mechanic, and P/FCF of ~15.3× leaves little cushion if the denominator slips.

30% -12%

Recommendation

Summary

Rating

HOLD

Conviction

Low

Price Target

$232

Timeframe

12 mo

Upside

+4%

Position Size

1.5%-2.5%

Entry Strategy

Tranche 1 — 30%

~$213

Start only at the post-gap Fibonacci floor near $213.61, roughly 5% below spot — not here. At $223.65 the stock offers ~4% to our base-case target, which is not compensation for binary event risk. Waiting costs little when the catalyst is 11 weeks out.

Tranche 2 — 35%

~$200

Add on a retest of the 52-week low at $199.19. This is the level at which the bear case is roughly half-priced — $200 sits between our $178 bear target and $232 base, so the risk/reward turns favourable even without knowing the October outcome.

Tranche 3 — 35%

~$186

Reserve the final tranche for a capitulation below the 52-week low — the zone implied by a second guidance cut, near HSBC's $191 and Morgan Stanley's $190 targets. Deploy only if free cash flow guidance survives; if the ~+$1B FCF guide is cut alongside revenue, the dividend-supported floor is gone and this tranche should not be used at all.

1-Year Price & Key Levels