Financebotresearch desk研究台

IBM › analyze

IBM · Analyze

Technology

Date: 2026-07-14 · Price: $219 (live, −25% intraday) · Sector: Technology (IT Services / Enterprise Software) · Beta: 0.68 Verdict: 🟡 WATCH — stage-in window opened, but Q2 print is now a binary thesis test. Conviction [6.0]. The valuation objection is gone; a new growth objection replaced it.

🚨 MATERIAL UPDATE — Jul 14, 2026: Q2 profit warning, stock −25% ($293 → ~$219)

IBM pre-announced a Q2 miss: EPS ~$2.93 and revenue ~$17.2B, both below estimates, on weakness in software AND infrastructure. CEO Krishna's framing: "In the last few weeks of June, clients shifted quarterly capex toward servers, storage, and memory to secure supply-constrained infrastructure ahead of expected price increases." Worst single-day drop in decades; it dragged CRM, NOW, ADBE, INTU down with it on fears that enterprises are reallocating budget from software to AI infrastructure.

This is the exact risk flagged as #1 below (AI-disruption to software/services). The fork is now explicit: - Timing blip (bull): a late-June capex deferral toward supply-constrained memory/servers — software demand pushed, not lost. Krishna's story. → today is a gift entry. - Structural (bear): the start of AI-infra cannibalizing software spend — the same repricing already crushing ACN/EPAM, now reaching software. The contagion to CRM/NOW/ADBE says the market is voting bear. - Verdict: value-vs-value-trap, unresolved. Don't catch the full knife. The formal Q2 report (late July) is the referee. See revised valuation & stance below.

Original (pre-warning) read — now stale on price: A genuinely improved IBM (software ~half of revenue, AI book >$6B, best mainframe cycle in years) that the market had already re-rated — stock ~doubled to $293, forward P/E ~23x, yield a decade-low ~2.3%. That "too expensive" objection is now moot; the $293 valuation math below is retained for reference.


1. Fundamentals (Health)

Metric FY2024 FY2025 Read
Revenue $62.75B $67.54B +7.6% YoY (HashiCorp + z17 + Red Hat)
Operating income $10.07B $12.49B +24%; op margin 16.1% → 18.5%
Net income $6.02B $10.59B +76% — but 2024 was depressed by pension/tax one-timers; op income is the cleaner signal
Diluted EPS $6.43 $11.17 Reported jump overstated by 2024 base
Gross margin 56.7% 58.2% Rising — software mix shift working
EBITDA $14.74B $17.51B Margin 23.5% → 25.9%
Operating cash flow $13.45B $13.19B Steady
FCF (OCF − capex, GAAP) $11.76B ~$11.5B IBM-defined FCF ~$13.4B; mgmt guides ~$14.5B for FY2026
Cash & ST investments $14.0B $13.6B

FCF & CAGR. GAAP FCF sits in an ~$11–12B band; IBM's own adjusted definition is ~$13.4B. Capex is tiny (~$1.7B) — this is a low-capex software/services model where the real "investment" is M&A, not plant. Multi-year FCF CAGR is low-single-digit but quality and direction are improving (Q1 2026 FCF $2.2B was the strongest Q1 since 2018). Note: clean 5–8yr CAGRs are muddied by the Nov-2021 Kyndryl spin-off (~$19B revenue removed); pre-2022 comparisons aren't apples-to-apples — flagged.

Capital allocation (FY2025 OCF $13.2B): - Dividends ~$6.26B (the priority — 30 straight years of increases) - M&A ~$8.3B (HashiCorp the big one), funded partly by +$2.9B net new debt - Buybacks ~$0.3B (token — essentially none; used only to blunt dilution) - Capex ~$1.7B → Post-Red-Hat playbook: dividend first, growth-by-acquisition second, buybacks off the table. Reinvestment discipline is decent but the M&A-funded-by-debt pattern is worth watching for empire-building.

Balance sheet. Total debt $64.6B, net debt ~$51B. Debt/EBITDA 3.7x gross / 2.7x net; debt/assets ~42.5%. Interest coverage healthy (EBITDA/interest ~9x). Investment-grade, manageable — but not a fortress, and leverage ticked up in 2025 to fund HashiCorp.

Shares outstanding — the per-share headwind. 937M (2024 diluted) → 949M (2025) → 952M (Q1'26). Rising ~1%/yr. No buybacks + ~$1.7B/yr stock comp + stock used as acquisition currency = a slow dilution drift. Growth is not being amplified per-share; it's being modestly diluted. This is the single biggest knock on the fundamentals.

Scorecard: FCF quality 🟢 · Margins/trend 🟢 · Revenue re-accel 🟢 · Leverage 🟡 · Share count 🔴 (rising) · Dividend safety 🟢 (payout ~55% FCF).


2. Moat & Competitive Advantage

Quantitative base. Gross margin 58% and climbing; estimated ROIC ~12–13% (NOPAT ~$10.6B on ~$84B invested capital) — comfortably above cost of capital, and improving as the mix shifts to software. (roic.ai's reported 4.9% ROE / $234 BVPS are data artifacts — the balance sheet gives equity $32.6B → BVPS ~$34, ROE ~32%, corroborated by Alpha Vantage's 35.8%. High ROE is flattered by a thin, goodwill-heavy equity base.)

Where the moat is real (switching costs + entrenchment): - IBM Z mainframe — the z17 cycle shipped >100% new-MIPS growth for four straight quarters; runs the transaction backbone of banks, insurers, airlines. Effectively un-rip-and-replaceable. AI-on-Z (watsonx Code Assistant) is extending the lock-in, not eroding it. - Red Hat (~13% growth) — the open hybrid-cloud control plane (RHEL, OpenShift, Ansible); HashiCorp/Terraform bolt-on aims to own the infra-as-code layer end-to-end. - Transaction Processing software — sticky, high-margin, tied to Z. - Software ARR $24.6B, +10% YoY; software heading to ~50% of total revenue by year-end.

Adversarial stress-test (well-funded rival): Cannot touch the mainframe/Z franchise — no economic path to displace 50 years of embedded COBOL workloads. Red Hat's open-source position is defensible but is contested by the hyperscalers (AWS, Azure, Google) on managed Kubernetes. The soft flank is IBM Consulting (~28% of revenue): classic body-shop labor arbitrage, directly in the blast radius of the same GenAI-disruption-to-IT-services repricing hammering ACN / CTSH / EPAM. A rival (or the client's own AI copilots) can attack that stream.

Evergreen assessment: The Z + Red Hat + software core is genuinely evergreen and even AI-advantaged (mainframe AI inference, watsonx governance for regulated industries). Consulting is not structurally evergreen at current headcount economics. Net: a narrowing-but-durable moat anchored on switching costs and mission-critical infrastructure, with one clearly exposed segment.


3. Valuation

Current ~$293 · Forward P/E ~23x (2026 consensus EPS ~$12.42) · Trailing P/E ~25x · EV/EBITDA ~18.6x · P/S ~3.9x · Div yield 2.28%.

Model Output Weight Note
Graham √(22.5×EPS×BVPS) ~$93 Low Near-meaningless here: negative tangible book (goodwill $68B + intangibles $11B vs. equity $33B). Graham punishes goodwill-heavy, asset-light software cos. Ignore the point estimate.
Bogle expected return ~5–8%/yr Medium 2.3% yield + ~6% earnings growth, minus likely P/E compression (23x → mid/high-teens over 5yr) = mid-single-digit total return. Modest.
Dividend Yield Theory Overvalued signal High (dividend name) Yield 2.28% sits at the low end of IBM's decade band (~4–5% in 2018–2023). By its own dividend history, IBM is expensive, not cheap. The re-rating did this.
DDM (g≈3%, ke≈8%) ~$138 Medium Dividend alone justifies ~$140. The other ~$150 of price is the growth-re-acceleration story — which must actually deliver.
Forward P/E fair band $224–273 Medium 18–22x on 2026 EPS $12.42. Current $293 sits at/above the top.

Fair-value range: ~$240–290. Current $293 is fully valued to slightly rich. Wall Street agrees it's close to done: avg target ~$282–295 (BofA bull $330); mix is 15 buy / 7 hold / 1 strong-sell — mildly positive, not a screaming buy.


4. Type Overlay — Dividend Grower

  • 30 consecutive years of dividend increases (Aristocrat), $6.72/yr, 2.28% yield.
  • But: it's now a low-yield, slow-grower — recent raises are ~1–2%/yr penny bumps, not the 8–10% CAGR the framework prizes.
  • FCF payout ~55% (of GAAP FCF) / ~47% of adjusted — safe and covered, improving as FCF grows.
  • DYT verdict: the yield-compression to a decade low is the clearest single "you missed the cheap entry" flag. This was a 5%-yield value stock at ~$130; at $293/2.3% the dividend thesis is spent — you're now buying the growth story at a full multiple.

5. Synthesis — Weighted Verdict

No material agent conflict. Fundamentals, Moat, and Sentiment all agree the business inflected genuinely positive (software ≈50% of revenue, AI book >$6B and monetizing on Z, FCF re-accelerating, FY26 guide likely raised to ">6% cc" at the Q2 print). Valuation is the lone dissenter — and on a dividend name, in a stock that has already doubled, Valuation gets the heavier weight.

Weighted call: Good, defensive, improving business — but the market has already paid for the improvement. This is the textbook "great company, bad entry" box. For a long-term fundamental investor, chasing $293 offers ~5–8%/yr expected return with real multiple-compression risk and no margin of safety.

A portfolio-specific passage was removed from the public build.

Key risks: (1) P/E de-rating — 23x forward is priced for continued acceleration; any software or z-cycle stumble compresses hard. (2) IBM Consulting (~28% of revenue) shares the GenAI-services-disruption risk repricing ACN/CTSH/EPAM. (3) Rising share count quietly taxes per-share growth. (4) M&A-funded-by-debt cadence (leverage rose for HashiCorp). (5) z17 hardware cyclicality — mainframe cycles roll off.

Golden points: AI book >$6B and monetizing · z17 the best mainframe cycle in years · software ARR $24.6B +10% · FCF strongest Q1 since 2018 · 30-yr dividend streak, ~55% payout · beta 0.68 defensive ballast.


A portfolio-specific passage was removed from the public build.