Applied Optoelectronics [AAOI]
Task 5 | analysis 2026-07-27 | migrated to the Criteria framework 2026-07-29 | framework v1.5.1 Spot $88.12 (scan, 2026-07-28). The 2026-07-27 build used $97.82.
2026-07-29 MIGRATION NOTICE — read this before anything below it
Migrated from the retired six-Gate / E[R]-vs-cash-hurdle framework onto the Criteria framework (
references/criteria.md,references/valuation.md).
- The position verdict is deleted. §3 previously read
INVESTMENT DECISION: WATCHLIST (short bias) — NOT Short, NOT Long. The memo emits an analysis, not a call.E[R]versus the 4.7% cash hurdle is deleted — includingE[R] = 0.30(−82.8%) + 0.45(−62.4%) + 0.25(+8.5%) = −50.8%. Replaced by the reverse-DCF implied-path test and a 12-month target, with sensitivity over the exit multiple, never over scenario probabilities.- Criteria are retired; every test is a named Criteria with a type, returning PASS / FAIL / INDETERMINATE.
- Revenue is now TTM, not last fiscal year. For AAOI this is a 3.7% correction — immaterial, unlike MU (136%) and SanDisk (76%) — and it is applied for consistency.
THE ACCOUNTING-QUALITY FINDING IN §2 IS PRESERVED IN FULL AND IS INDEPENDENT OF ANY REVENUE CORRECTION. A single private, unrated stocking distributor at 53.1% of revenue and 71.6% of receivables at ~264-day implied DSO, named in none of the 18 quarterly earnings releases examined (4Q21–1Q26), is a finding about disclosure and cash conversion. It does not depend on which revenue endpoint is used, it is not weakened by anything in this migration, and it is the single most valuable output of this memo.
Pre-computed inputs are taken from
reports/scan/AAOI_analysis.json(as-of 2026-07-28) and are used rather than recomputed.
| OLD (screen / scan / rescore path) | NEW (corrected) | |
|---|---|---|
| Revenue basis | last fiscal year, FY2025 (ended 2025-12-31), 210 days stale | TTM, four quarters to 2026-03-31 |
| Revenue | $455.7m | $472.6m — last-FY understated it by 3.7% |
| EV | $6.625bn | $6.625bn |
| EV / Sales | 14.54x | 14.02x |
| Demonstrated revenue CAGR | 26.9% (annual-FY) | 26.0% (TTM-consistent, FY2022 → TTM) |
| Implied-path margin | +5.5pp (scan, on a hardcoded 20% terminal margin) | +2.6pp |
AAOI is the name where the last-FY error did not matter, and saying so is as important as reporting the 136% and 76% errors on MU and SanDisk. The artifact scales with how fast revenue is inflecting within the stub period; AAOI's inflection is annual rather than quarterly, so both endpoints agree. What did not agree was the terminal margin: the 2026-07-28 rescore used a hardcoded 20%, which is corrected to the derived 18.45% below and costs AAOI 2.9pp of margin.
portfolio_book.json read at 2026-07-27:
| Item | Value |
|---|---|
| Positions | none — the book is 100% cash |
| Watchlist | MU (two-sided), ISRG (long bias) |
| Cash hurdle | 4.70% annual |
| Max single-name weight | 5.0% |
| Max sector concentration | 25% |
| Max pairwise correlation without disclosure | 0.60 |
| Position hard stop | −2.0% of book |
| Drawdown ladder | review −5%, de-gross −10%, stop −15% |
Marginal correlation (trailing 1-year daily log returns, 252 common days, Alpaca SIP):
| Pair | ρ |
|---|---|
| AAOI vs COHR | 0.632 |
| AAOI vs LITE | 0.627 |
| AAOI vs CIEN | 0.532 |
| AAOI vs MU (watchlist) | 0.386 |
| AAOI vs SPY | 0.383 |
| AAOI vs AMZN | 0.156 |
| AAOI vs MSFT | 0.053 |
Two things follow, and one of them answers the caller's question 3 directly:
Since the book is 100% cash, no position would be displaced. The 4.7% cash-hurdle test is retired and is not applied — under the current framework capital competition is resolved by slot competition inside the strategy, not by a return threshold inside the memo. The correlation and concentration facts above are retained because they are inputs a book genuinely needs.
1. Fundamental conclusion. Revenue growth is real but its composition is misunderstood: FY2025's +82.8% was driven by a finite DOCSIS 4.0 cable upgrade cycle (+179%), not by AI (+31.7%), and that cycle has already plateaued (CATV +3.6% y/y in 1Q26). Cash conversion is deteriorating badly — eight-quarter operating cash outflow of $301M against a $216M net loss, with receivables at 180 days concentrated 71.6% in one private distributor. Operating income has not improved despite revenue nearly doubling. The company funds itself by issuing equity roughly quarterly.
2. Expectations conclusion. The Street is Buy-rated, ~$151 average target, FY2026 EPS ~$0.91, and 2Q26 revenue consensus of $194.1M sits at the top of the company's own guide. Consensus EPS requires roughly 35% gross margin; the company guides 29–30%, and the higher-margin CATV tranche is the one rolling off. ΔE (house − Street) is −13% to −26% on FY2026 revenue and materially larger on EPS.
3. Valuation conclusion. Probability-weighted value $48.15 against a $97.82 spot; scenario-weighted E[R] for a long −50.8%. No cell in the DCF sensitivity grid approaches the spot price. Crucially, the bull case — everything the bulls underwrite, granted in full — returns only +8.5%.
4. Portfolio conclusion. No position. The research conclusion (bearish) and the portfolio action (none) are different things, and they diverge here for reasons that have nothing to do with thesis quality — see Criteria 5 and 6.
Every Criteria is scored. The memo blocks on none of them.
| Criteria | Type | AAOI | Result |
|---|---|---|---|
| Quality Criteria | BINDING | Archetype INFLECTION (FY2025 operating margin −12.0%). The three INFLECTION tests: gross margin 30.0% — FAILS the level test (the standard asks ~50%, and this is the test that proves the unit economics work); operating-margin CHANGE +16.5pp — PASSES emphatically (−28.4% → −12.0%); revenue growth 26.9% annual-FY / 26.0% TTM-consistent — PASSES (above the ~18% threshold). One of three fails, and it is the one that matters most for this name. Corroboration all points the same way: accruals +0.159 (first positive year in five), F-score 4/9, GP/A 0.12, asset growth +113.6% driven by equity issuance rather than operations. Per criteria.md, F-score and gross-profitability levels are demoted to context on an INFLECTION name — but accruals are retained on both archetypes, and accruals here are the single worst reading |
FAIL |
| Valuation Criteria | BINDING | See §3A. Price requires 23.4% five-year revenue CAGR; demonstrated 26.0% TTM-consistent. Margin +2.6pp | PASS, narrowly |
| Liquidity Criteria | BINDING | $7.07bn market cap. The options chain was pulled and it is the binding constraint: implied vol 127–148%, realised 142.2% (252d), beta ~4.1, single sessions of ±45%. At those levels a defined-risk structure requires the thesis to be right by more than 30% merely to break even, and a −2%-of-book hard stop is unenforceable on a name that gaps 45%. Equity is tradable; the vehicle problem is real and priced | PASS on equity; every options structure priced FAILs |
| Downside Criteria | MEASURED | Permanent-loss case and its named cause — this is the most important line in the memo, and it is independent of every revenue and multiple question above. The cause is counterparty and cash-conversion risk concentrated in one undisclosed name: Digicomm International, a privately held, unrated stocking distributor, is 53.1% of revenue and 71.6% of accounts receivable (FY2025 10-K Note 8) at an implied ~264-day DSO against 180 days company-wide. AAOI is financing its own distributor with roughly nine months of working capital. Company-wide DSO went 106 → 180 days over five quarters; AR rose +418% against revenue +249%; AR plus inventory of $505.2m stands against TTM revenue of $507.0m. And Digicomm is named in none of the 18 quarterly earnings releases examined (4Q21–1Q26) — independently re-verified for this migration: zero matches for "Digicomm" across all 18 Ex-99.1 files on disk. A stocking distributor holds inventory it has not sold; sell-in is not sell-through. Realistic permanent impairment: the bear case at $16.85 (−80.9% from $88.12). A going-concern leg must be argued explicitly and it is not argued here — net cash is +$0.446bn — but the financing dependence is stated: eight registered offerings in fifteen months and share count +193% | Scored, blocks nothing |
| Catalyst Criteria | MEASURED | 2Q26 earnings, 6 August 2026 (confirmed). Consensus revenue $194.1m — the high end of the company's own $180–198m guide — and consensus EPS +$0.07 against a company guide of −$0.03…+$0.03. The second CATV-plateau test lands with 3Q26 | PASS |
| Momentum Criteria | MEASURED — entry timing only, never a selection veto | 12-1 momentum +495.0%, cross-sectional percentile 93.5th (quintile 5); 6-1 +302.0% (98.4th); RSI-14 38.1; 39.5% of the 52-week high ($223.10); below the 200-day. The realised 12-month range was $19.49 → $223.10, or +1,045%. The prior document scored this as Gate 6 — FAIL as positioned today and reverted the whole name to Watchlist on it. That veto is deleted. Momentum governs when to enter a position the analysis already justifies, never whether to own — and a name below its 200-day at 39.5% of its high, with 12-1 in the 93rd percentile, is a timing observation and nothing more |
Scored, blocks nothing |
| Peer Spread Criteria | MEASURED | Named same-end-market peers with their own multiples: COHR (EV/EBIT 444.7x, demonstrated growth 20.5%), FN (38.1x, 14.8%), LITE, CIEN (225.5x, 9.5%). AAOI's own EV/EBIT is INDETERMINATE — EBIT is negative, so the multiple does not exist. On sales, 14.02x = 86th percentile of AAOI's own trailing three years. Correlations: COHR 0.632, LITE 0.627, CIEN 0.532 — all three above the 0.60/0.53 disclosure band; and notably MSFT 0.053, AMZN 0.156, i.e. AAOI does not trade as a hyperscaler-capex derivative | Partly INDETERMINATE |
| Consensus Criteria | MEASURED — no longer an admission test | House FY2026 revenue $742.9m vs Street $850m–$1.0bn (−13% to −26%). The sharper disagreement is margin: consensus FY2026 EPS of ~$0.91 requires roughly a 35% gross margin, against a company guide of 29–30% for 2Q26 and no full-year GAAP print above 30.0% in nine years. Sourcing caveat retained: Alpha Vantage's quota was exhausted before the original build and again on 2026-07-29, so consensus is from public aggregators, and the FY2026 revenue figure is a range because sources genuinely disagree. Blocks nothing | Scored (qualified) |
| Short Mechanism Criteria | MEASURED — acted on by nothing on this fork | Decelerating growth? Yes, in the growth engine — CATV revenue $64.5m (1Q25) → $66.8m (1Q26), +3.6% y/y after +179% in FY2025. Exhausted margin runway? No — operating margin is still negative with room to expand. One of two legs, and the first leg is unusually well evidenced | Scored |
| Sub-sector Criteria | MEASURED | Optical components / CATV-HFC | Tagged |
Note on what changed and what did not. Three of the six old gates that produced WATCHLIST are gone as
blocking tests: Valuation Criteria (E[R]) is replaced by the implied-path test, Liquidity Criteria's FAIL is reclassified as a
vehicle problem rather than an admission test on the underlying, and Momentum Criteria's FAIL is deleted outright
because momentum cannot veto. What survives untouched is the accounting-quality evidence, and it survives
because it never depended on any of them.
The old §Valuation Criteria computed E[R] = 0.30(−82.8%) + 0.45(−62.4%) + 0.25(+8.5%) = −50.8% for a long and compared
it to a 4.7% cash hurdle. Both are retired.
| Parameter | Value | Held fixed / basis |
|---|---|---|
| Spot | $88.12 | scan, 2026-07-28 |
| Shares | 80.24m | EDGAR |
| Net cash | +$0.446bn | EDGAR |
| EV | $6.625bn | fixed |
| Revenue (t=0) | $472.6m TTM (to 2026-03-31) | fixed |
| Years | 5 | fixed |
| WACC | 10.0% | fixed |
| Terminal EBIT margin | 18.45% | semis/optical cohort median of positive operating margins (n=22 of 25). AAOI's own FY2025 operating margin is −12.0%, so the max(own, cohort median) rule reduces to the pre-profit branch, and it is named as such. This is a generous assumption on a company that has not printed a positive annual operating margin in the last decade |
| Exit multiple | 42.8x EV/EBIT | GROWTH_MATCHED, n=16 (scan) |
RESULT — what the price requires: a 23.4% five-year revenue CAGR. Margin = demonstrated − required = 26.0% − 23.4% = +2.6pp. → PASS, narrowly. On the scan's annual-FY growth measure of 26.9%: +3.6pp. On the 2026-07-28 rescore's hardcoded 20% terminal margin: +5.5pp. The correction to a derived terminal margin costs AAOI 2.9pp, and it is the only correction on this name that moved anything.
A PASS here is not an endorsement, and the framework is explicit about why. The Valuation Criteria asks one question — is the price achievable? — and 23.4% compound revenue growth is achievable for a company that has demonstrated 26.0%. It says nothing about whether the revenue converts to cash, which is precisely what the Downside Criteria evidence puts in doubt. Under the old framework a cheap-enough valuation could not rescue a failed mechanism; under the new one "cheap cannot rescue a failure" in the Quality Criteria remains the rule, and the Quality Criteria FAILs on gross margin. These two Criteria disagreeing is information, not an inconsistency.
Implied compression, as a number. EV/EBIT compression is INDETERMINATE — FY2025 EBIT is negative, so today's EV/EBIT does not exist. On a sales basis the exit multiple is 42.8 × 18.45% = 7.90x EV/Sales against today's 14.02x = −43.7%.
Exit-multiple identification. The scan anchor is GROWTH_MATCHED with n=16 and its constituents genuinely bracket AAOI's 26.9% growth (CI 15.0%/12x … ANET 27.1%/53x … AVGO 24.4%/64x). 5 of the 16 are semiconductor or optical companies — a materially better sector fit than MU (1 of 15) or Ciena (4 of 29). Restricting the same screen to the semis/optical cohort gives exactly n = 5, the minimum, at a median of 64.1x (FN 38x, ANET 53x, AVGO 64x, MPWR 83x, AMD 198x), which would put the required CAGR at 13.8% and the margin at +12.2pp. Both are reported; the lower, more conservative anchor is used as the base.
Sensitivity — over the EXIT MULTIPLE, never over probabilities:
| Exit EV/EBIT | Basis | Required CAGR | Margin (26.0% demonstrated) | Result |
|---|---|---|---|---|
| 12.3x | anchor-set minimum (CI) | 58.3% | −32.3pp | FAIL |
| 30.9x | anchor-set p25 | 31.7% | −5.7pp | FAIL |
| 42.8x | anchor-set median — base | 23.4% | +2.6pp | PASS |
| 64.1x | anchor-set p75 and the sector-restricted growth-matched median (n=5) | 13.8% | +12.2pp | PASS |
| 198.0x | anchor-set maximum (AMD) | −9.2% | +35.2pp | PASS |
The flip point sits between the anchor set's 25th percentile and its median — this is a genuinely knife-edge PASS, and it is stated as one rather than presented as a result. Note also that the old build's own scenario table anchored exit multiples on AAOI's own trading history of 1–2x sales pre-2024, with a bull case at 7.0x. That anchoring is now explicitly disallowed as a 5-year exit multiple — a 1–2x sales multiple was set when AAOI was a flat-revenue business, and applying it to a 26% grower is the same defect as anchoring a 30% grower on 1–7% growers. The old §6 scenario returns (−82.8% / −62.4% / +8.5%) are therefore retired as valuation outputs and retained only as Downside Criteria evidence.
Terminal value far exceeds 60% of EV (AAOI earns nothing today), so the reverse DCF is mandatory as the primary long-horizon output.
Anchor caveat, stated plainly. AAOI's own distribution is strongly bimodal — a pre-2024 regime at 0.2–1.5x and a post-2025 regime at 4–64x — so the three-year median of 3.90x is a gap midpoint, not a central tendency. The base anchor used is the trailing-one-year median of 10.32x, the same rule applied to Ciena in this batch. The anchor is identified but fragile: the one-year window is entirely a re-rating period, and that is disclosed rather than hidden inside the multiple.
| Anchor (AAOI's own EV/Sales) | Multiple | On consensus mid $0.925bn | vs spot $88.12 | Range $0.85bn – $1.00bn |
|---|---|---|---|---|
| Own 3-yr p25 — pre-2024 regime | 1.99x | $28.46 | −67.7% | $26.60 – $30.32 |
| Own 3-yr median — gap midpoint | 3.90x | $50.57 | −42.6% | $46.92 – $54.22 |
| Own 3-yr p75 ≈ hold-spot | 7.58x | $92.96 | +5.5% | $85.87 – $100.05 |
| Own 1-yr median — base case | 10.32x | $124.48 | +41.3% | $114.84 – $134.12 |
| Own 1-yr p75 / 3-yr p90 | 26.21x / 23.53x | $307.64 | +249.1% | — |
12-MONTH TARGET: $124 — +41.3% ABOVE SPOT (range $115–$134 on the consensus revenue range). This target is reported alongside a FAILING Quality Criteria and the accounting-quality evidence in §2, and it does not supersede either. A 12-month multiple-and-estimate target and a judgement about earnings quality are different instruments answering different questions, and the framework requires both to be shown.
The multiple required merely to hold spot on consensus mid-point revenue is 7.16x — 48.9% below today's 14.02x, and almost exactly AAOI's own three-year 75th percentile. On the house FY2028 base revenue of $0.879bn it is 7.53x.
Sanity band. Street average target ~$151.30 (+71.7% to spot) and a longer-dated ~$118 by May 2027 (WallStreetZen, single source). The house 12-month target of $124 sits 18% below the Street average and 5% above the longer-dated figure. The divergence from the Street average is explained: the Street embeds FY2026 revenue at the top of the range and a multiple above AAOI's one-year median. The external targets are a check on the output, never a calibration target for the model — and both are thinly sourced, which is recorded.
The prior §3 read INVESTMENT DECISION: WATCHLIST (short bias) — NOT Short, NOT Long. It is deleted.
Whether this analysis justifies a position is a question about a particular book — its slots, its
optical-complex exposure (COHR 0.632, LITE 0.627, CIEN 0.532) and above all its tolerance for a security with
142% realised volatility, beta ~4.1 and single sessions of ±45%.
What is established, as findings:
Momentum Criteria — FAIL as positioned today reverted the entire
analysis to Watchlist because the stock had not yet closed below its 200-day. That is a timing observation
being used as an ownership test, and it is exactly the failure mode criteria.md names momentum as the
test case for.What would refute the bearish evidence: a properly disclosed Digicomm relationship with terms and ageing;
DSO reverting toward 106 days; an announced 800G hyperscaler qualification with a named customer; CATV revenue
re-accelerating above +20% y/y; or a quarter of positive operating cash flow without new equity issuance.
Each is dated in 06_Catalyst_Calendar.md.
These are no longer conversion triggers for a position verdict; no verdict is emitted. They are the dated, falsifiable observations that would move the Criteria scores, listed so a book can act on pre-committed evidence rather than on the emotion of the moment.
Fundamental (confirmatory — the Downside Criteria evidence is already on the record): - F1. 2Q26 CATV revenue below $70M (i.e. still flat-to-down sequentially), confirming the cycle plateau. - F2. 2Q26 gross margin at or below 29%, or 3Q26 guided below 29%. - F3. 2Q26 DSO above 175 days, or the Digicomm receivable concentration rising above 71.6% in the 10-Q. - F4. Another 424B5 filed, confirming the financing cadence.
Technical (a Momentum Criteria observation — ENTRY TIMING ONLY; it can neither compel nor forbid ownership, and the prior document's use of it as a blocking test is deleted): - T1. A weekly close below the 200-day moving average (~$88.89), or - T2. A weekly lower-high / lower-low sequence with a 50/200 bear cross.
Vehicle feasibility (Liquidity Criteria, options limb — priced on the live chain, and the hardest to satisfy): - V1. 3-month realised volatility below 70% (currently 141.0%), and - V2. Front-month IV below 80% (currently 130–148%), so that a defined-risk put spread costs a defensible fraction of spot.
If F1–F3 confirm but V1/V2 do not, the evidence is confirmed and the vehicle is still absent. A vehicle that cannot be filled or justified is not a vehicle — but that is a statement about instruments, not about the company, and the prior document's step from "no viable vehicle" to "WATCHLIST" is exactly the conflation the current framework forbids.
| Vehicle | Cost / mechanics | Verdict |
|---|---|---|
| Outright short equity | Borrow est. 8–15% p.a. (15% SI, not PB-verified); unlimited upside risk; +45% single-day precedent | The only viable route for negative exposure, and only at ~1% of book. Not a recommendation |
| Long put outright — AAOI 18-Sep-26 $80 put | mid $10.14, IV 132.3%, delta −0.259. 10.4% of spot for eight weeks | Rejected. Naked long premium at 132% IV; the variance risk premium makes this systematically expensive and the skill's default is a spread |
| Put spread — 18-Sep-26 $90/$70 | Long $90 ≈ $15.3, short $70 $6.50 → net ≈ $8.8 debit, max value $20, max gain ~$11.2 | Rejected today: needs a −28% move by 18 Sep merely to reach max value, versus an implied ±29.7% already priced. No edge. |
| Jan-27 put spread — $100/$70 | net ≈ $33.55 − $16.04 = $17.5 debit on a $30 wide spread | Rejected. Paying 58% of the spread width. |
| Collar / call overwrite | n/a — no underlying position | n/a |
Conclusion on vehicle: no options structure is investable at these levels. At IV ≈ realised ≈ 130–140% the options market is efficiently priced and simply expensive; no structure priced above expresses a directional view at an acceptable cost. This is a Liquidity Criteria finding about instruments — it is recorded so a future user does not re-derive it, and it says nothing about whether the underlying is worth owning.
| Component | Weight | Score | Rationale |
|---|---|---|---|
| Fundamental trajectory (Quality Criteria) | 25% | 0.85 | Evidenced, four-factor corroboration |
| Variant vs consensus (Consensus Criteria) | 25% | 0.70 | Real and mechanism-linked, but consensus is thin (6 analysts) and the FY2026 revenue figure is not primary-sourced |
| Catalyst and timing (Catalyst Criteria) | 20% | 0.80 | Dated, 10 days out, resolves the specific disagreement |
| Valuation / payoff (Valuation Criteria) | 15% | 0.85 | Large gap; bull case only +8.5% |
| Balance sheet and risk | 10% | 0.35 | $439.7M cash, no net debt — the company cannot be forced. Bad for a short. |
| Technical / implementation (Liquidity Criteria) | 5% | 0.10 | Fails |
| Composite | 0.72 | MEDIUM |
Volatility tier: HIGH (136.1% realised vs the >45% threshold). Grid (Medium × High) → 1.0% of book. Capped at 1.0% and not raised, because the grid's "High vol" bucket was calibrated for 45–60% names and AAOI is at 142.2% (scan, 252-day). (Sizing is a constraint offered to the strategy, not a verdict. The prior build carried this arithmetic into the WATCHLIST call; that step is deleted.)
Consensus positioning: this view is against a Buy-rated Street with a ~$151 average target, and that consensus strengthened through 2025 as the stock rallied. The disagreement is decomposed to a specific, falsifiable cause (gross margin required vs guided), which is the standard the skill sets for a contrarian view — but going against a strengthening consensus in a 15%-short-interest name is real crowding risk and is named here rather than absorbed silently into the volatility tier.