INTERNATIONAL BUSINESS MACHINES CORP

INTERNATIONAL BUSINESS MACHINES CORP (IBM) — Investment Memo

🐂 The Bull Case (Warren's voice)

• Why is the moat durable and why does it compound?
  The mainframe lock‑in is real. Banks, airlines, and governments rely on decades of custom COBOL and regulatory‑grade integrations. Ripping them out would take a decade and cost more than the GDP of a small country. This creates steady, predictable cash flow — $12.1B FCF in 2025, consistently above net income. The subscription‑software shift (Red Hat, hybrid cloud) is real, and the $67.5B revenue base is now 60%+ recurring. Switching costs compound because every year of additional integration makes the moat deeper for existing workloads.

• What makes the economics exceptional — specifically?
  Software margins are excellent (gross margins >75%), and the Financing segment earns a spread on customer loans with IBM’s own low‑cost debt. The combination of high‑margin subscriptions and captive financing creates a capital‑light cash machine within a capital‑heavy shell. FCF yield at current prices is ~6.5% — decent for a bond‑like equity with a possible growth kicker once the pivot to cloud automation matures.

• At what price range does this become genuinely attractive to Berkshire?
  If the stock falls below $100 (a 50% discount to DCF), the 3.4% dividend yield becomes a safe 6.8%, and the debt load is covered by a 12% FCF yield — a classic Buffett “cigar butt” with a decent chance of surviving its own decline. At $140 (25% discount), the margin of safety is thin given the structural headwinds, but the yield alone (~5%) compensates for zero growth. Below $100, the case improves.

🐻 The Bear Case (Charlie inverts)

Munger's rule: "Show me where I'll die and I won't go there."

• Scenario 1: Cloud‑native mainframe emulation
  A well‑capitalized competitor (Amazon, Google, or a startup with $5B in VC) builds a drop‑in replacement that runs legacy COBOL on cheap x86 architecture. If a bank can move its core transaction processing to AWS for ⅓ the cost, switching costs collapse overnight. IBM’s $54.8B debt and $12.1B FCF would be exposed — mainframe maintenance revenue (a huge chunk of Infrastructure) would evaporate. Likelihood: Moderate, timeframe 5–10 years.

• Scenario 2: Debt spiral from a recession
  IBM’s $54.8B debt is 4.5x FCF in a normal year. If a recession cuts FCF to $6B (as in 2022’s near‑miss) and interest rates stay high, interest coverage goes below 2x. The dividend (eating $5.8B annually) becomes unsustainable. IBM would be forced to cut the dividend, sell assets, or issue equity at depressed prices — a classic value trap. Likelihood: High in the next downturn, timeframe 1–3 years.

• Scenario 3: The “consulting‑as‑a‑service” commoditization
  Consulting revenue (32% of total) is low‑margin and project‑based. Clients are increasingly using AI tools (e.g., ChatGPT for code, automated workflow) to replace IBM’s strategic advice. If consulting margins compress to single digits, the $19B revenue stream becomes a net drag on capital. IBM’s pivot to “hybrid cloud” is an admission that its highest‑margin businesses are shrinking. Likelihood: High, already underway.

The most likely permanent impairment is the combination of 1 and 3 — a slow bleed where mainframe revenue declines 3‑5% annually, consulting margins erode, and debt stays stubbornly high. No sudden death, but a decade of zero real value creation.

💰 Valuation & Margin of Safety

• DCF estimate (provided): $178.1B total / $189.84 per share (3.0% FCF growth, 10% discount rate, 3% terminal growth).
  Note: This DCF assumes the moat persists — a generous assumption given the bear case.

• 25% margin of safety entry: $142.38 per share
  Conservative — still assumes no impairment, but compensates for the risk of a modest recession.

• 50% margin of safety entry (Buffett’s ideal): $94.92 per share
  At this price, the dividend yield ~7.2%, FCF yield ~13%, and the debt is covered by a real cash flow margin. Only here does the risk/reward become compelling.

• Current status: IBM trades near $200 (as of early 2025). That is ~5% above the DCF intrinsic value — fair, not cheap. With structural risks and a decade of zero revenue growth, paying $190 for a melting ice cube is a mistake.

Verdict: PASS

PASS. At current prices, IBM offers no margin of safety against a deteriorating moat and a leveraged balance sheet. The DCF implies fair value near $190, but structural risks — cloud‑native emulation, debt vulnerability, consulting commoditization — make even that optimistic. We wait for a 50% discount to $95 before considering any entry, and even then, only as a short‑term cash‑generation play, not a compounder.