Berkshire requires a 25–50% discount to intrinsic value before buying.
Buffett Quality Checklist
✓ROE >15% consistently (≥7 of last 10 years)
✓Free cash flow positive (≥8 of last 10 years)
✓Conservative leverage — Debt/Equity below 1
✓Revenue growing at CAGR >5%
✓EPS growing at CAGR >5%
10-Year Financial History — SEC EDGAR 10-K Filings
Year▲
Revenue▲
Net Income▲
FCF▲
Owner Earnings▲
ROE▲
Net Margin▲
LT Debt▲
Cash▲
2016
$1.1B
$184.2M
$152.9M
$182.5M
16.6%
16.3%
—
$567.9M
2017
$1.6B
$423.2M
$616.3M
$428.6M
25.5%
25.7%
—
$859.2M
2018
$2.2B
$328.1M
$479.3M
$332.0M
15.3%
15.3%
—
$650.0M
2019
$2.4B
$859.9M
$947.3M
$877.0M
29.7%
35.7%
—
$1.1B
2020
$2.3B
$634.6M
—
—
19.1%
27.4%
—
$893.2M
2021
$2.9B
$840.9M
—
—
21.1%
28.5%
—
$620.8M
2022
$4.4B
$1.4B
—
—
27.7%
30.9%
—
$671.7M
2023
$5.9B
$2.1B
—
—
28.9%
35.6%
—
$1.9B
2024
$7.0B
$2.9B
—
—
28.5%
40.7%
—
$2.8B
2025
$9.0B
$3.5B
—
—
28.4%
39.0%
—
$2.0B
Warren & Charlie
Buffett / Munger — quality, moat & valuation
Arista Networks, Inc. (ANET) — Investment Memo
🐂 The Bull Case (Warren's voice)
Moat is real and widening. EOS software creates switching costs that deepen as hyperscalers automate – ripping out Arista APIs would cost years and billions. Net margin expanded from 16.3% to 39.0% while revenue grew 8×; that’s pricing power in action.
Economics are exceptional. Zero debt, ROE 28.4%, net income grew 17× on 8× revenue. Incremental returns on capital are rising – a textbook franchise.
Attractive entry only if FCF is proven. The DCF at $73.6B (~$58.58/share) assumes 8% FCF growth, 10% discount, 3% terminal. If FCF mirrors net income (highly likely from pre-2020 data), that intrinsic value is conservative. Berkshire would find the business interesting below $44/share (25% margin of safety) and ideally$29/share (50% safety). At current prices (likely well above $100), it’s a WATCH, not a BUY.
🐻 The Bear Case (Charlie inverts)
Hyperscaler vertical integration. The single structural threat: Amazon, Google, or Microsoft builds its own switches at cost. Arista’s pricing power evaporates. Given reseller concentration (two resellers flagged in 10‑K), a single defection could slash revenue 30-50%.
Accounting alchemy on 39% net margin. Networking hardware peers (Cisco, Juniper) run 15-25% net margins. 39% is either a monopoly moat or aggressive revenue recognition / one‑time gains. FCF data is missing after 2019 – the most important number. If cash conversion is poor (e.g., deferred revenue buildup or receivables stretching), the moat is thinner than it looks.
Management trust deficit. No share count disclosed, no insider ownership shown, no straight talk on FCF. Brilliant operators hiding the cash flow book – that’s a permanent impairment risk if the hidden numbers are worse than NI.
Most likely threat: Hyperscaler build‑out over the next 5–7 years. Arista’s 29% revenue growth in 2025 may already be decelerating as custom silicon gains traction.
💰 Valuation & Margin of Safety
Intrinsic value estimate: $58.58/share (DCF, 8% FCF growth, 10% discount, 3% terminal). This assumes FCF equals net income – a leap of faith given missing data.
25% margin of safety entry: $44/share
50% margin of safety entry: $29/share(Buffett’s ideal – rarely reached for quality)
Current price: Not provided, but given market context (ANET trades >$100), it is expensive by ~70-100% vs intrinsic. No margin of safety.
Verdict: WATCH
The business is structurally brilliant, but the missing FCF data and 39% net margin in a hardware business are unresolved red flags; we cannot pay a premium for trust we don’t have. If the stock ever drops below $44/share and Arista releases clean cash flow statements, this becomes a BUY. Until then, patient inversion wins.
Data sourced from SEC EDGAR XBRL filings (10-K only). For educational purposes — not investment advice.