Broadcom (AVGO)

Total return at 5 years: 12.5%/year = 11.7% appreciation + 0.8% dividend. The target price ($616) is ex-dividend; the $18 in dividends collected over 5 years are added separately.

By both methods, the value today (DCF $598 · Multiples $510) exceeds the market price ($355).

Today — fairly valued: at $355 trades ~30.4% below its value discounted to today (~$510); the discount is positive but does not reach the margin of safety we require (≥38%).

At 5 years — Infravalorado: the target price ($616) plus dividends yield above the required average return (10%) — the business compounds.

The bridge: the return at 5 years exceeds the risk-free rate (4.5%) — but the discount does not reach the required margin of safety (≥38%). To require a 15% annual return, it would need to be bought at ~$318.

Educational / informational. Does not constitute investment advice.

The inverse approach: instead of projecting growth to obtain the value, the market price ($355) is taken as given and it solves for what annual owner-earnings growth would need to hold for 5 years for the present value —at the method's rate (4.5%, no-growth terminal)— to equal that price. It is the disconfirmation test: the expectations the price already pays for, contrasted against the method's projection.

The market discounts less growth (14.3%/year) than we project (27.4%/year) → if the base case holds, there is margin: perception is more pessimistic than the estimated reality.

That growth implies ~$83.4 bn of owner earnings in year 5 (vs ~$143.2 bn of our base case). It recalculates if the DCF assumptions are edited.

Year-by-year projection of the selected scenario. From each year, two versions of the flow are derived: growth FCF (operating flow − total capex, the cash surplus) and maintenance FCF (the owner earnings: what the business yields if it only sustains its capacity). Cash accumulates the retained surplus —what is not returned as dividend or buyback—, so that EV falls and multiples compress going forward. The valuation is done on EV/NOPAT (resultado operativo económico, neto de SBC) · derivado de las partes (semis + software). In edit mode, revenue, margins, capex, and exit multiples can be adjusted.

Rolling grid: the base year is the real trailing twelve months (revenue $89.10 billion and GAAP operating income $42.81 billion at August 2, 2026, built by levels from the September 2, 2026 release: fiscal 2025 plus the first three quarters of fiscal 2026 minus the first three of the prior fiscal year, because the quarter is not yet in the Securities and Exchange Commission's structured database) and the projection is 12-month windows (+1y…+5y). Operating income is economic: it adds back acquisition-related intangible amortization ($7.99 billion, a non-cash charge from VMware, CA, and Symantec) and expenses stock-based compensation ($8.48 billion, 22% of free cash flow, a real cost). The normalized tax rate is set at 16%, the one the company itself publishes in its adjusted-measures reconciliation for the quarter and for the first three quarters of the fiscal year; the window's effective book rate is 5.4% and is distorted by the fiscal 2025 tax benefit. Year 1 starts from current guidance: the fourth quarter is guided at $34.80 billion of revenue and a 66% adjusted operating margin, an annualized pace of $139 billion; of the 69 points of year-1 growth, 56 are the mechanical carryover of that quarter and the rest a modest sequential advance. From there the path closes a constant fraction of the gap to the 10% terminal rate each year — the large first-year step is arithmetic from the starting point, not a jump in the curve. The terminal multiple is derived from the parts: 20 times after-tax operating income in semiconductors (80% of the terminal metric) and 18 times in software (20%), a 19.6-times blend, below the 22.9× the mechanical position-within-band rule would suggest. Cash accumulates the retained fraction of the flow after the dividend ($0.65 per share per quarter) and the real buyback, which returned to scale ($8.45 billion over the trailing twelve months versus $2.45 billion in fiscal 2025). The company dilutes slightly on net —stock-based compensation exceeds the buyback— so it does not carry a share-count path: that cost is already in the metric. Watch: customer concentration (42% of revenue from a single customer in the quarter ended May 3, 2026), the gross margin compression the company itself anticipates from the accelerator mix, and working capital consumption ($11.02 billion in the window).

Today's elevated multiple is the price of growth: if the business grows, the entry point cheapens on its own going forward (EV falls as cash increases). The exit multiple at 3 years is higher than the terminal at 5 years —at 3 years there is more growth still ahead—, so the value curve shows whether value creation is concentrated in the early or the later years. The required return is applied to the base scenario.

They are valued separately, with their own rationale, and are not incorporated into the base or the verdict (they are excess return). When assigning them value — in Editmode —, the total with optionalities updates live, without moving the base.

The verdict, the base CAGR, and the margin of safety are always calculated on the base; optionalities do not alter them (with optionalities at $0 they do not move).

Each card fixes a required annual return and answers: if the business is worth $616 in 5 years, what is the maximum that can be paid today to obtain that return? Since it now trades at $355, the margin of safety is how much cheaper the market is than that maximum. The three thresholds: 4% covers inflation (the floor), 10% is the long-term average return, and 15% is the level of a great investment.

With a target price of $616 in 5 years plus $18 of dividends collected (the dividend adds to the return, not to the price) and a required return of 4.5% annually, the maximum to pay today is $510. Against the current market price ($355), the margin of safety is 30.4% (trades below the maximum → there is margin) and the total return at that price would be 12.5% annually.

• Valuation by multiples, part by part. Each engine on its own multiple — semiconductors 20 times and software 18 times after-tax operating income, summed. Not a single multiple over the consolidated figure.
• The starting point was re-anchored. At $355 it trades at 42× after-tax operating income. The quarter moved the base metric abruptly: trailing-twelve-month revenue rose to $89.10 billion and the next quarter is guided at $34.80 billion.
• The multiple compresses as it compounds. The multiple on after-tax operating income falls from 42× to 10× over five years as the metric compounds, with the exit multiple held within each business type's band.
• Concatenated optimism (the method's risk). The base case stacks revenue that nearly triples with an economic margin held around 57%. The exit multiple is kept below what the mechanical position-within-band rule would suggest, precisely to avoid stacking a third favorable assumption.
• Partial margin of safety. At $355 the discount against the value brought to today is real, but does not reach the threshold of a great investment: against the maximum price that would admit a 15% annual return, the margin is -12%.

Each margin over sales, year by year: historical (solid line) → projection (dotted).

• ✓ ROIC exceeds the cost of capital (~10%)
• ✓ High incremental ROIC (compounding)
• ! CFROIC backs up the ROIC (73%, cash vs. accruals)
• ✓ Healthy balance sheet (low corporate debt)
• ✓ Durable competitive moat (multiple advantages)

• Strong and predictable cash. The free cash flow (~$33bn TTM, ~42% margin) is enormous and the capex minuscule (<1% of revenue); the predictability of software is high, that of AI semis depends on the cycle.
• High return on capital. ROIC ~27% (on total capital, including VMware's goodwill) — well above the 10% bar; ex-goodwill it is extraordinary.
• Acquisition amortization trap. The amortization of intangibles (~$8bn/year, VMware/CA/Symantec) is a non-cash charge from acquisition accounting → it is added back in the economic operating income; the market values on non-GAAP which also adds it back.
• Material stock-based compensation. The SBC (~$8.8bn TTM) is ~28% of free cash flow (11.6% of revenue) → material by the cash flow bar (§3): it is expensed in the metric (EV/NOPAT), never added back.
• Net diluter. The shares rose (4,232M → 4,876M) despite the buybacks: the stock-based compensation issuance exceeds the buybacks. The cost is already in the metric (no additional dilution is modeled, §6).

Annual levels from the official filing (10-K); the % over each bar is the year-over-year (YoY) growth vs the prior year.

Broadcom's growth decomposes by segment into volume by price: in semiconductors, custom-silicon units (artificial intelligence accelerators and networking) by generation price; in software, subscription licenses by price, with the VMware migration to enterprise-wide subscription. The engine is custom silicon co-designed with the large clouds, whose quarterly revenue was $16.7 billion and is guided to $21.7 billion, and whose contracted revenue reached $164.6 billion — signed revenue that pulls growth forward. The non-artificial-intelligence portion of semiconductors (wireless, broadband, storage) is mature and cyclical.

The % are the annual (year-over-year) growth: each year —historical and projected— vs the prior one; the TTM (trailing 12m) vs the TTM of a year ago, to avoid overlapping windows. The historicals are exact figures from the official filings; the projected years come from the year-by-year model (the intermediate years without their own series are interpolated between the anchors). The projected columns (+1y…+5y) are 12-month windows counted from the TTM close (2-ago-2026): the projection starts from the most recently reported data, not the fiscal year. The projected base is realistic and unbiased — the risk discount is applied at the end, via the required return. The rationale for each metric is in the (i).

• Real growth, not nominal. Yes — AI revenue jumped from $12.2bn to $20.2bn on custom silicon volume, with the contracted backlog ($164.6bn) as signed backing.
• Where the engine comes from. Semiconductors (~63% of revenue) grows +79% in the latest quarter from custom AI silicon; software (~37%) grows ~9% but contributes the margin. The growth engine is AI.
• Anchored in contracted backlog. The $164.6bn RPO (from $33.3bn) is signed revenue that leads semiconductor growth — more defensible than a trend extrapolation.
• Realistic deceleration. The base case decelerates from the current +48% toward ~+7% by the fifth year: the law of large numbers and the cycle bite; extrapolating today's ramp would be concatenated optimism (§2).
• Growth concentration. AI growth depends on a few programs from one hyperscaler; the same backlog that anchors it is concentrated in one dominant contract.

• Relevant in 10 years / recurrence. Yes — the AI compute infrastructure (custom silicon + networking) and the critical virtualization software are structural; the unknown is customer concentration, not relevance.
• Market share / cost. Leader in custom silicon (the hyperscalers' default co-design partner) and de facto standard in enterprise virtualization (VMware). R&D cost advantage at scale.
• Pricing power. Strong in both engines: the software segment margin (77%) evidences VMware's pricing power; in custom silicon, IP leadership sustains the price per generation.
• Operating leverage. High: asset-light (capex <1% of revenue), with AI revenue scaling strongly — the operating margin expands as it grows.
• Capital cycle / competition. AI attracts massive capital and Broadcom is a beneficiary, but the hyperscalers could internalize the design over time; the non-AI portion of semis (wireless/broadband) is cyclical.
• Customer concentration. The structural risk: a single customer = 42% of revenue (Q2 FY2026), the five largest ~45%. The dependence on a few hyperscalers for custom silicon is extreme.

Not all debt is equal: only the structural needs refinancing; the rest is operational (self-liquidating).

Structural debt is what is exposed to the contraction phase of the cycle; operational debt (leases, matched funding) self-liquidates with the business.

• ✓Leverage (net debt / EBITDA)Net debt / EBITDA 0.7x
• ✓Interest coverage (EBIT / interest)16.3x
• ✓Liquidity (current ratio)2.50×
• !Cash quality (CFROIC vs ROIC)CFROIC backs 73% of ROIC
• ✓Value creation (ROIC − 10% bar)+22pp
• ✓Malinvestment test (capex vs incremental ROIC)Productive investment (2.0x)
• !Float / working capitalConsumes cash $11 bn (positive WC)
• ✕Dilution (SBC % of revenue + shares)SBC 9.5% of revenue

A traffic-light interpreted by the method (not generic): float (negative WC) adds up, capex is judged by incremental ROIC (malinvestment test), and a lender is not subjected to corporate solvency. The (i) shows the derivation of each number.

• Net debt declining. Gross debt of $59.42 billion versus $65.14 billion at the close of fiscal 2025 and cash of $23.98 billion: net debt fell to $35.4 billion and cash flow covers it several times over.
• Enormous cash generation. Trailing-twelve-month operating cash flow ($40.65 billion) covers a minuscule capital expenditure ($1.25 billion) and leaves $39.40 billion of free cash flow.
• Revenue concentration. The balance sheet risk is not solvency but dependence: one customer was 42% of revenue in the quarter ended May 3, 2026, versus 29% a year earlier.
• Growth is consuming working capital. Current items absorbed $11.02 billion of cash over the trailing twelve months (receivables up $6.54 billion over three quarters, inventory up $2.25 billion): it is the cash cost of the buildout.

Who owns the shares — the alignment and whether there is a controlling shareholder.

• Capital allocation. Hock Tan is a serial high-return allocator: disciplined M&A (VMware), priority deleveraging and a growing dividend. Coherent as long as the return on capital holds.
• Alignment / skin in the game. Founder Henry Samueli chairs with 1.8%; Hock Tan has a modest direct stake but 96% of his compensation is variable, tied to long-term total shareholder return.
• Integrity / communication. It communicates customer concentration and the deleveraging transparently in the filing; the AI revenue guidance lives in the press release, not in the 10-K.
• Acquisitions / diworsification. Growth is by acquisition (VMware the latest) + organic in custom silicon; VMware's integration is already mostly done (restructuring −61%), with no signs of dispersion.

• The reference price is the last close before the results release: the company reported fiscal 2026's third quarter on September 2, 2026 after market close, so the price being compared against the value has not yet priced in a quarter with revenue growing 86% or the +93% guidance for the next.
• The stock fell while its artificial intelligence line tripled: over that same period the segment's quarterly revenue went from a low single-digit billion figure to $16.7 billion. The market compressed the multiple across the group of artificial intelligence infrastructure suppliers while the business was accelerating.
• The market discounts customer concentration and the durability of the custom silicon programs, which are real risks; the discrepancy is that the already-contracted revenue ($164.6 billion of remaining performance obligations, with roughly 30% recognizable within twelve months) is a fact presented in the filing, not an extrapolation.

The source of the discount is both calendar and perception: calendar, because the reference price precedes the results release, and perception, because the group's multiple compressed while the business was accelerating. The first corrects itself in the next trading session and deserves to be stated plainly; the second is the part that can sustain a return. The disconfirming test —customer concentration— is also the reason the discount falls short of the required margin of safety.

The annual return (CAGR at 5 years) in each scenario, with the total period return below — the margin of safety made visual: upside range wide, downside range narrow.

Even in the bear scenario, the return holds at -5%/year (-23% total): the margin of safety protects the downside. The bull (+23%/year, +181% total) exceeds it comfortably — a favorable asymmetry, with a narrow downside range and a wide upside range.

• Extreme customer concentration. A single customer was 42% of revenue in the quarter ended May 3, 2026. The loss or scaling back of a custom silicon program hits revenue and contracted revenue disproportionately.
• The base case builds in a very large buildout. Revenue nearly triples over five years from a window that has already grown 48.7%. It is what guidance and contracted revenue support, but it leaves little room for an execution stumble.
• Sustainability beyond what is contracted. Artificial intelligence revenue depends on a few programs from one large cloud; a digestion of cloud capital spending or a delay hits the line that supports the projection.
• Gross margin compression from mix. The company itself discloses that selling custom accelerators and networking equipment will likely raise operating margin but compress gross margin.
• Debt inherited from VMware. $59.42 billion of gross debt; covered by cash flow, but a risk if rates or the credit rating move against it.

Each thinker's analytical framework applied to our data.

• Buffett / Graham Quality + margin of safety

  A wide moat and ROIC 32% above the 10% bar, but the margin is limited (+30%) → excellent business at a fair price.

• Peter Lynch Growth at a reasonable price (GARP)

  A hyper-growth growing 27% at a multiple/growth of 1.5 → reasonable for its growth.

• Joel Greenblatt Cheap and high-return (Magic Formula)

  Earnings yield 3% (EBIT/EV) + ROIC 32% → falls outside the Magic Formula.

• Howard Marks Perception vs reality + cycle

  The price discounts 14% vs our 27%: perception is more pessimistic than reality, with an identified source of the discount.

• Seth Klarman Capital protection (bear scenario)

  Bear-scenario floor -6%/yr, bull-scenario ceiling +22%/yr over 5y: reasonable asymmetry, without an ample cushion.

• Pat Dorsey Moat strength (Five Rules)

  A wide moat, stable; sources: switching costs, intangibles, efficient scale, network effects, cost advantage → passes the Five Rules.

• Aswath Damodaran Expectations implied by the price

  Justifying the price requires discounting 14%, within what we project (27%) — the story squares with the numbers.