DOX › analyze
DOX · Analyze
Date: 2026-08-12 | Price: $57.28 | Market cap: $6.08B | EV: $6.98B | FY ends Sept 30 First full file on this name. Prior mention: screen-dividend-software-ai-tailwind-2026-03-18.md, where DOX ranked #2 of 7 and "BEST INCOME PICK" at $66. It is now $57.28 — down 13.2% since, and −35.4% from its $88.61 high.
Headline: three of the four numbers that made this screen cheap are wrong, and the business is better than they say
DOX surfaced in the 8/12 screen sweep on a set of figures that are each corrupted in a different way. Correcting them is most of the analysis.
| Screen showed | Reality | Why the screen was wrong |
|---|---|---|
| Revenue −9.4% FY25, 3yr CAGR −0.3% | +3.1% pro forma constant currency | ~$600M of low-margin non-core activity phased out in FY24. Vendor CAGR divides a continuing-ops endpoint by a total-company base — [[pitfall-divested-segment-corrupts-multiyear-cagr]] |
| P/E 13.6x trailing | 10.1x normalized | TTM GAAP EPS $4.20 carries $207.8M of restructuring charges |
| Forward P/E 7.14x | 7.73x on the current fiscal year | Implies $8.02 EPS vs FY26 guidance of $7.41. Vendor forward is a year out — [[pitfall-vendor-forward-eps-is-the-wrong-fiscal-year]] |
| Earnings growth −57% | +5.5–6.5% guided non-GAAP | Same restructuring distortion |
The divestiture is the whole story on the top line, and it was a good decision. Amdocs shed ~$600M of low-margin revenue and its economics improved:
| FY2024 | FY2025 | ||
|---|---|---|---|
| Revenue | $5.00B | $4.53B | −9.4% as reported, +3.1% cc pro forma |
| Gross profit | $1.76B | $1.72B | −2.5% |
| Gross margin | 35.2% | 38.0% | +280bps |
| Operating income | $759.7M | $812.2M | +6.9% on 9.4% less revenue |
That is the correct shape for a deliberate portfolio pruning: less revenue, more profit. A screener sorting on revenue growth reads it as decline.
Phase 1 — Fundamentals
Reconstructing the real earnings base
Yahoo's TTM figures are distorted, so the quarters are rebuilt from the source:
| Quarter | Revenue | GAAP dil. EPS | Restructuring / unusual | Normalized income |
|---|---|---|---|---|
| Q4 FY25 (Sep-25) | $1,150.2M | $0.88 | −$79.7M | $160.4M |
| Q1 FY26 (Dec-25) | $1,155.9M | $1.45 | −$11.3M | $166.7M |
| Q2 FY26 (Mar-26) | $1,172.0M | $1.28 | −$10.4M | $146.0M |
| Q3 FY26 (Jun-26) | $1,174.9M | $0.59 | −$106.4M | $136.5M |
| TTM | $4,653M | $4.20 | −$207.8M | $609.6M |
Three earnings bases, three very different multiples:
| Basis | EPS | P/E at $57.28 |
|---|---|---|
| TTM GAAP (as reported) | $4.20 | 13.6x |
| TTM normalized | ~$5.67 | 10.1x |
| FY26 non-GAAP guidance (midpoint) | $7.41 | 7.73x |
FY25 non-GAAP diluted EPS was $6.99 (+8.5%); FY26 is guided to +5.5–6.5%, giving $7.37–7.44.
🚩 But the restructuring is not a one-off, and I will not treat it as one. Charges have now hit four consecutive quarters ($79.7M → $11.3M → $10.4M → $106.4M). A company that restructures every quarter is running a continuous cost programme, not absorbing a discrete event. The honest earnings power sits between normalized GAAP ($5.67) and company non-GAAP ($7.41) — non-GAAP also adds back equity compensation and intangible amortization, which are real costs. I anchor valuation nearer the FCF line, which cannot be adjusted away.
Cash flow — the strongest part of the case
| FY2022 | FY2023 | FY2024 | FY2025 | FY2026E | |
|---|---|---|---|---|---|
| OCF | $756.7M | $822.6M | $724.4M | $749.1M | — |
| Capex | −$227.2M | −$124.4M | −$105.5M | −$104.0M | — |
| FCF | $529.5M | $698.3M | $618.9M | $645.1M | $710–730M guided |
| Buybacks | −$508.5M | −$489.5M | −$563.1M | −$551.3M | — |
| Dividends | −$186.1M | −$199.5M | −$212.0M | −$224.4M | — |
FCF 3yr CAGR +6.8% against a −0.3% reported revenue CAGR — the divergence is the divestiture again, and it favours the owner. Capex has fallen from $227M to $104M; this is now a genuinely capital-light business converting ~14% of revenue to FCF.
FCF yield: $720M / $6.08B = 11.8%. Haircutting for real restructuring cash (~$70–100M/yr), call it $640M sustainable → 10.5%. Either figure is exceptional.
Capital allocation — aggressive and consistent
Buybacks have run $490–563M every year for four years, retiring stock relentlessly:
| FY2022 | FY2025 | Q3 FY26 | Change | |
|---|---|---|---|---|
| Diluted shares | 123.65M | 109.66M | 105.69M | −14.5% |
Combined return of capital is ~$790M/yr against a $6.08B cap — a ~13% total shareholder yield (9.1% buyback + 3.96% dividend). 🚩 Note this exceeds FCF: $551M + $224M = $775M vs $645M FCF in FY25, funded by drawing cash from $573M (FY22) to $325M (FY25). Sustainable at the guided $720M FCF, but there is no longer a cash cushion to over-distribute from.
Balance sheet
Debt $826.4M against $325.0M cash; debt/assets 13.2%, D/E 32.7%. Interest expense $42.4M is covered ~19x by operating income. No refinancing pressure.
🚩 Book value is not what it appears. Equity $3.43B includes $2.89B of goodwill — tangible book is roughly $0.54B, so BVPS $31.17 corresponds to a tangible ~$5. Any book-based model (Graham, P/B 1.84) is measuring acquired intangibles, not assets. Weighted down accordingly.
Dividend — a 13-year record, and the reason to own it
| Current | $0.569/qtr, $2.276/yr | Yield | 3.96% |
| Consecutive annual raises | 13 (2013→2026) | Latest raise | +8.0% |
| 5yr dividend CAGR | ~9.6% | 10yr dividend CAGR | ~11.3% |
| Payout — GAAP TTM | 54% | Payout — FCF | 31% |
Never cut, never frozen, raised every single year since initiation. The FCF payout of 31% is the
number that matters: the dividend is covered three times over and has ample room to keep compounding
near 8–10%. This clears the dividend-grower overlay in analysis_notes.md comfortably.
Phase 1 — Moat (qualitative)
What Amdocs actually is: the dominant vendor of BSS/OSS — the billing, charging, CRM, order-management and network-operations software that carriers run their businesses on. When you get a phone bill, an Amdocs system very likely produced it. It is sold with a large managed-services attachment (record levels this quarter), which is why 27,000 people work there.
The moat is switching cost, and it is close to the strongest kind. Replacing a carrier's billing stack is a multi-year, nine-figure programme with a live risk of mis-billing millions of subscribers. Carriers defer it almost indefinitely. Contract terms run 5–10 years — the newly signed Liberty Latin America deal is a 10-year agreement.
Market structure: Amdocs, Ericsson and Nokia together hold ~38–42% of global OSS/BSS software revenue; the top five vendors hold ~54–59% of cloud OSS/BSS. Amdocs is the #1 pure-play.
🚩 The adversarial case — two real threats
1. Customer concentration is severe. AT&T alone was 24% of revenue in FY23 and 27% in FY22. AT&T and T-Mobile together are approximately 40% of annual revenue. This is the single largest risk in the file and it is not diversifiable: two customers, in one consolidating industry, in one country. A renewal lost on price, or a scope reduction at either, moves the whole thesis. BR's ~65% proxy-revenue concentration is regulatory and sticky; DOX's 40% is commercial and negotiable.
2. A scaled competitor was just assembled — in May 2026. NEC completed its $2.9B acquisition of CSG Systems and is merging it into Netcracker to build a unified AI-native OSS/BSS platform. That converts two sub-scale competitors into one with real breadth, explicitly positioned on the AI-native architecture argument. This is recent, material, and is very plausibly a large part of why the stock de-rated 35% while its numbers held. It does not break the switching-cost moat on installed accounts, but it raises the cost of winning new logos and strengthens the counterparty across the table at every renewal — including the two that are 40% of revenue.
The AI question — applying the field test
[[pattern-headcount-revenue-divergence-tests-ai-deflation]] says management AI commentary is
worthless and only one disclosure separates monetizing AI from being eaten by it: the sign of
revenue growth against headcount growth.
DOX passes. Headcount is 26,969, down from roughly 31,000 in FY2023 (~−13%), while revenue grew ~3% cc and gross margin expanded 280bps. Revenue up, headcount down — productivity is being retained, not handed back to clients. The continuous restructuring charges are the cash cost of exactly this conversion, which is why I treat them as recurring rather than one-off.
The aOS (Amdocs Operating System) agentic platform, built with AWS and NVIDIA partnerships, is the offensive expression of the same shift. Q3 commentary reports early traction. Genuine optionality, not yet a demonstrated revenue line — I underwrite none of it.
Evergreen rating
Moderately evergreen. Carriers will need to bill customers for as long as there are carriers, and the incumbency is deeply entrenched. But the end market is structurally low-growth and consolidating — fewer, larger carriers means fewer, more powerful buyers. This is a 3–4% grower with a strong moat, not a compounder. Own it for the cash return, not the growth.
Phase 2 — Valuation
Type overlay: dividend grower. Graham is weighted down for goodwill; DYT is applied with the price-cause correction; FCF-based valuation carries the most weight.
| Model | Output | Weight | Note |
|---|---|---|---|
| FCF yield | $63–80 | High | $640–720M sustainable FCF at an 8–9% required yield |
| Multiple on FY26 non-GAAP | $67–82 | High | 9–11x $7.41 |
| Bogle expected return | ~10.0%/yr | Medium | 3.96% yield + 6.0% EPS growth, no re-rating assumed |
| Graham IV | $54 GAAP / $63 normalized | Low | √(22.5 × 5.67 × 31.17) = $63.06, but BVPS is 84% goodwill |
| DYT | $76 (corrected) | Low-Med | See below |
🚩 DYT correction. The 3.96% yield is 89% above the 2.09% five-year average, which reads as deep value. But per [[pitfall-dyt-inverts-when-price-caused-the-yield]], the yield can rise two ways and here it is mostly the wrong one: the dividend grew 8%, while the price fell 35%. Roughly two-thirds of the yield expansion is the price decline. Reverting to the 2.09% average would imply $108.90, which is not credible; a normalized 3.0% yield gives $75.87, and that is the version I carry.
Fair value: $65–82, central ~$73
At $57.28 the stock sits ~21.5% below the central estimate and below the low end of the range.
The range is deliberately below where a 38%-gross-margin, 21.6% operating-margin software business with a 13-year dividend record would normally trade. 40% customer concentration and a newly-scaled competitor justify a permanent discount — I am not modelling a re-rating to a normal software multiple, and neither should the thesis depend on one.
Bogle says you earn ~10%/year if the multiple never moves at all. That is the load-bearing number: at 7.7x forward, you are paid to wait, and any re-rating is upside you did not underwrite.
Zones
- Entry $50–62 — in zone at $57.28. Strong buy below $52 (52-week low $49.74).
- Trim $82 — deliberately dollar-form. ≈11x FY26 non-GAAP EPS $7.41. Per
[[pitfall-multiple-trim-inherits-the-broken-vendor-field]], the site computes a multiple-trim as
mult × (price ÷ vendor P/E), and DOX's vendor GAAP EPS of $4.21 is depressed by restructuring — so aTrim 11x fwdtag would render $46, below both spot and the entry zone, inverting the call. Re-set this judgment at each/analyze.
Phase 3 — Tensions
Fundamentals vs. Moat. Fundamentals sees expanding margins, growing FCF and a 13-year dividend record. Moat sees 40% of revenue in two customers and a competitor that just doubled in scale.
Resolution — [[pattern-margin-intact-while-volume-falls-defers-the-damage]] is the right lens,
and DOX survives it. That pattern warns that stable margins can mask an incumbent that has not yet
fought the price war. The test is to read margin and volume together. Here both are moving the
right way: gross margin +280bps and revenue +3% cc and managed services at a record. Amdocs is
not buying volume with price. The NEC-Netcracker threat is therefore prospective rather than already
in the numbers — which is precisely why it belongs in the break triggers rather than the verdict.
Sentiment dissents, and it should be recorded. Analysts have kept their ratings but cut targets hard: Stifel $88 → $71, Barclays $111 → $92, BofA $100 → $97, Wolfe downgraded to Peer Perform, KeyBanc initiated at Sector Weight (neutral). Consensus target $81.21 (+41.8%) is therefore a falling number, not a stable one. The direction of revision agrees with the moat analyst, not with the multiple.
Phase 4 — Verdict
🟢 ACCUMULATE · conviction [7.0] · in zone
The case for: a 7.7x forward multiple on a business with a 38% gross margin, 21.6% operating margin, 11.8% FCF yield, ~13% total shareholder yield, 13 consecutive dividend raises at a 31% FCF payout, 0.42 beta, and a switching-cost moat that carriers structurally refuse to test. Bogle gives ~10%/year with no re-rating whatsoever. Three of the four metrics that made it screen cheap were vendor artifacts; corrected, the business is growing modestly with expanding margins and passing the AI-deflation test that most of its labour-based peers fail.
Why not higher than [7.0]: two customers are ~40% of revenue, and in May 2026 NEC merged CSG into Netcracker to build a scaled AI-native competitor. Growth is 3%. Restructuring has run four straight quarters, so "adjusted" earnings deserve suspicion. This is cheap for reasons that are real, even if they are over-priced.
Why not lower than [7.0]: the reasons are priced twice over. At 7.7x forward with a 31% FCF payout, the market is charging for concentration risk that has not yet produced a single lost contract, while the company just signed a 10-year Liberty Latin America deal and posted record managed services.
Against BR — the comparison that prompted this file
| BR [8.5] | DOX [7.0] | |
|---|---|---|
| Forward multiple | 14.6x | 7.7x |
| Yield / streak | 2.55% / 20yr | 3.96% / 13yr |
| Total shareholder yield | ~4% | ~13% |
| Revenue growth | 7–9% | 3% |
| Customer concentration | Diversified | 🚩 ~40% in two |
| Moat type | Regulatory (SEC/NYSE framework) | Switching cost |
| Competitive change | Stable | 🚩 NEC+CSG+Netcracker merger |
BR keeps the top spot. DOX is materially cheaper and pays you far more to wait, but BR compounds
2–3x faster with a diversified base and a regulatory moat that no merger can assemble against it.
DOX is a legitimate income holding at a deep discount — a complement to BR, not a replacement.
Note the field lesson from [[consulting-it-services]]: diversification value comes from revenue
model, and DOX's contracted 5–10yr BSS licensing plus managed services is a genuinely different
model from BR's regulated proxy toll — they are not duplicates of each other.
Break triggers
- AT&T or T-Mobile renewal lost, or scope materially cut
- Non-GAAP EPS growth guided below 4%
- Restructuring charges continuing past Q4 FY26 (a fifth consecutive quarter)
- Dividend raise streak broken
- FCF below $650M
- NEC-Netcracker-CSG taking a named tier-1 Amdocs account
Next event: Q4 FY26 and FY27 guidance, mid-November 2026. That print carries the annual dividend raise (the 14th), FY27 growth guidance, and the first clean read on whether restructuring finally ends.