Lumentum Holdings LITE

EARNINGS BANNER — Q4 FY2026 reports 2026-08-11, 7 days away

Lumentum's fiscal year ends in late June. Fourth-quarter and full-year results are due 2026-08-11 on consensus EPS of $2.97 and revenue of $988 million — a 22.2% sequential increase on the March quarter's $808.4 million, and a figure that would complete a fiscal year of roughly $3.0 billion against $1.645 billion a year earlier.

The verdict below is set with that timing as an explicit input. The stock closed at $849.47, up 8.92% on the day, after a 696.3% twelve-month return — and, notably, down 13.0% over three months, which is a consolidation rather than a peak. This dive declines to take a new position seven days before a print on a 1.513-beta stock that has risen sevenfold, and every upgrade condition in Section 10 is written to be answered by that report.

Reference table

1. The quarter, and what the filing says is inside it

Quarterly net revenue, from inc_q, $M — the trajectory is the thesis:

Revenue has grown in every single quarter for seven consecutive quarters, from $308.3 million to $808.4 million — a 162% increase — and the gross margin has expanded from 16.6% to 44.2% across the same span.

The 10-Q's own disaggregation, quoted rather than paraphrased:

> "Net revenue increased by $383.2 million, or 90.1%, during the three months ended March 28, 2026 compared to the three months ended March 29, 2025, driven by a $232.5 million increase in Components products and $150.7 million increase in Systems products.

>

> The increase in Components products was primarily driven by the ramp of laser chip and laser assembly product shipments to support strong, broad-based demand across intra-data center, data center interconnect, and long-haul applications. Additionally, a slight increase in average selling prices of laser chip products contributed to the increase in Components revenue driven primarily by a shift to 200G lane speeds. The remaining approximately 12% of Components revenue growth was due to an increase in shipment volume of data transport products, encompassing line subsystems solutions for long-haul terrestrial networks and charge pump products used in sub-sea network installations.

>

> The increase in our Systems products was primarily driven by our cloud transceiver product lines which increased by more than $137.0 million due to higher shipment volume, partially offset by lower average selling prices. We also continued the initial phase of optical circuit switch shipments, which contributed more than $25.0 million in revenue during the three months ended March 28, 2026, and we remain on track for manufacturing expansion over the coming quarters to support future growth."

And for the nine months: "Net revenue increased by $843.4 million, or 72.4%... driven by a $560.3 million increase in Components products and a $283.1 million increase in Systems products... cloud transceiver product lines increased by more than $268.0 million... optical circuit switch shipments... contributed more than $38.0 million."

Four things in that disclosure deserve separating out, because they are what a components story usually cannot supply.

First, and most important: laser chip average selling prices ROSE. "A slight increase in average selling prices of laser chip products... driven primarily by a shift to 200G lane speeds." A components supplier raising price while nearly doubling volume is describing a supply constraint it controls. That is the single most valuable sentence in this filing and it is the direct empirical test of the knowledge base's "choke point" claim.

Second, and pulling the other way: cloud transceiver prices FELL. "Higher shipment volume, partially offset by lower average selling prices." The transceiver business — the larger of the two Systems drivers at $137.0 million of incremental quarterly revenue — behaves like a normal component business, winning on volume and losing on price. Both dynamics are present in the same company, and the mix between them determines whether the margin holds.

Third, optical circuit switches are a new category ramping from zero. "More than $25.0 million" in a quarter and "more than $38.0 million" across nine months, with manufacturing expansion under way. On $808.4 million of quarterly revenue this is 3.1% — small, growing from nothing, and the specific product the knowledge base's most technical claim names.

Fourth, the growth is broad rather than single-application. "Intra-data center, data center interconnect, and long-haul applications", plus "line subsystems solutions for long-haul terrestrial networks and charge pump products used in sub-sea network installations." Roughly 12% of Components growth came from data transport products unrelated to the AI build.

The customer concentration, stated by the company

> "During the three months ended March 28, 2026, two customers individually accounted for 26% and 12% of our total revenue, respectively."

Thirty-eight per cent of revenue in two accounts. This is the hardest single risk in this dive and it is disclosed rather than inferred. In an optical components business, revenue is won socket by socket in a customer's design cycle; a competitor winning the next generation removes the line rather than eroding it. Nothing in this archive discloses contract terms, duration or backlog for either relationship.

The reporting structure changed — and the payload has the old one

> "Prior to fiscal year 2026, we operated in two reportable segments consisting of Cloud & Networking and Industrial Tech. During the first quarter of fiscal year 2026, we implemented a re-organization, and we are now managed as a single, integrated enterprise... Accordingly, following the reorganization, we determined we operate in a single reporting segment."

The vendor's seg_prod reports "Optical Communications Segment" and "Lasers Segment" — a structure two reorganisations out of date, whose most recent row is fiscal 2023 and contains only "Lasers Segment" at $209.2 million. The block is unusable and no segment figure in this dive comes from it.

Geography — seg_geo is CLEAN and it contains a surprise

The block reconciles exactly to FY2025 revenue of $1,645.0 million — the three regional aggregates sum to the total and each region's country detail sums to its aggregate. Against eight broken seg_geo blocks in twelve names in the previous batch, this is worth recording as clean.

What it shows is a supply chain, not a customer base. Hong Kong at 24.2% and Thailand at 17.7% are ship-to destinations in a contract-manufacturing network, not end markets — the United States at 19.0% is the largest single national figure and understates where the demand originates. It also means the tariff and export-control risks the 10-Q discusses at length are not abstract: 60.8% of revenue ships to Asia Pacific.

2. The convertible-note machine — the most consequential thing in this file after the growth

Lumentum funds itself with convertible notes, four series are outstanding, and the conversion mechanics are actively producing shares right now.

The 2032 Notes, from the 10-Q:

> "On September 8, 2025, we issued $1,265.0 million in aggregate principal amounts of 0.375% Convertible Senior Notes due in 2032 ('2032 Notes') in a private placement... The net proceeds from the sale of the 2032 Notes was approximately $1,254.7 million, after deducting $10.3 million of debt issuance costs. Concurrent with the issuance of the 2032 Notes, we used $843.1 million of the net proceeds to repurchase $581.1 million aggregate principal amounts of the 0.50% Convertible Senior Notes due in 2026 and $102.0 million of the net proceeds to pay the cost of the 2032 Capped Call Options...

>

> The initial conversion rate is 5.3257 shares of common stock per $1,000 principal amounts of the 2032 Notes (which is equivalent to an initial conversion price of approximately $187.77 per share)."

The share price is $849.47. The conversion price is $187.77. The stock is 4.52 times the strike.

The other series, also disclosed: the notes "rank equally with all of the Company's existing senior unsecured indebtedness, including the Company's outstanding 0.50% Convertible Senior Notes due 2026, 0.50% Convertible Senior Notes due 2028, and 1.50% Convertible Senior Notes due 2029, and do not contain any financial covenants, restrictions on dividends, incurrence of senior debt or other indebtedness, or the issuance or repurchase of securities by us."

How the dilution actually arrives:

> "Under each series of Notes, we are required to satisfy our conversion obligation with respect to converted Notes by paying cash equal to the principal amount of such converted Notes and paying or delivering... cash, shares of common stock or a combination, at our election, with respect to any conversion value in excess thereof."

So the principal is repaid in cash and the entire gain above par is settled in stock at the company's election. At 4.52 times the conversion price, that excess is roughly 3.5 times the principal.

And it is happening, in two documented forms:

> "As of April 30, 2026, we have received early conversion requests totaling $500.8 million aggregate principal amount of the Notes, which will be settled in cash and the conversion value in excess thereof will be settled in shares of common stock."

> "On April 7, 2026, we entered into privately negotiated exchange arrangements with certain holders of our 2026 Notes and 2029 Notes. Pursuant to these agreements, we issued an aggregate of approximately 5.7 million shares of our common stock in exchange for approximately $264.8 million and $209.8 million aggregate principal amount of the 2026 Notes and 2029 Notes, respectively, and the related conversion value in excess of the principal amounts thereof."

5.7 million shares against 77.8 million outstanding is 7.3% dilution in a single transaction, on 2026-04-07. And $500.8 million of principal has already requested early conversion. The capped call options bought for $102.0 million protect only the 2032 Notes — "specifically designed to mitigate the dilutive impact of the 2032 Notes, such that no dilution will occur until the capped call price is exceeded" — and the cap price is not disclosed in the extractable text of this archive.

The visible evidence in the payload confirms the direction. weightedAverageShsOutDil for fiscal 2025 reads 69.6 million; the 10-Q cover states 77.8 million outstanding at 2026-04-30. An 11.8% increase in ten months. And the balance sheet shows the mirror image: net debt of $2,087.5 million at 2025-06-28 against an implied trailing net debt of approximately $696 million — a $1.4 billion reduction achieved substantially by turning debt into equity.

The honest framing. This is not a scandal and it is not hidden — it is the ordinary and expected consequence of financing a company with converts and then having the stock rise sevenfold. The company got extremely cheap money (0.375%, no covenants) and the holders got the upside. What it means for a buyer today is that the denominator of every per-share figure is rising, and that consensus EPS estimates struck on a share count that keeps growing carry a source of downward pressure the revenue line does not.

3. Cash flow, balance sheet and the quality of the trailing numbers

Annual, $M:

Two cautions on the annual series. First, fiscal 2025's net income of $25.9 million against an operating loss of $192.2 million is produced by an income-tax BENEFIT of $198.0 million — a valuation-allowance release rather than operations. The vendor's effectiveTaxRateTTM of MINUS 60.5% and netIncomePerEBTTTM of 1.605 both reflect it. Every trailing profitability figure on this name is distorted by that item and none is used as a valuation anchor. Second, inc_a reports the FY2021 and FY2022 rows as very nearly identical — revenue $1,712.6M, net income $198.9M and EPS $2.79 in both, differing only by $1.1 million of operating expenses — which is a duplicated row and is flagged in Section 7.

Free cash flow is CLEAN in arithmetic and small in magnitude. freeCashFlowOperatingCashFlowRatioTTM reads 0.605, correctly below 1.0, and operatingCashFlowPerShareTTM $6.327 less capexPerShareTTM $2.498 equals freeCashFlowPerShareTTM $3.829 to three decimals. capexToDepreciationTTM of 0.702 is plausible for a company that is expanding manufacturing capacity. The resulting free-cash-flow yield is 0.41% — economically trivial against a $66.1 billion market capitalisation, and correctly computed. A company scaling from $1.6 billion to $3.0 billion of revenue while building capacity should not be generating meaningful free cash flow, and it is not.

Balance sheet at 2025-06-28 with the trailing implication:

Net debt of roughly $696 million against a $66.1 billion market capitalisation is 1.1% — economically immaterial, and it got there by converting into the equity. The notes carry no financial covenants, so there is no refinancing cliff and no maintenance test. The balance sheet is not a risk on this name; the share count is.

A related acquisition the payload does not record. From the 10-Q, Note 4: "On March 17, 2026, we acquired a manufacturing facility in Greensboro, North Carolina for $38.0 million in cash from a third party. The acquired business mainly included land and building, machinery and equipment and an assembled workforce. As part of the transaction, we also entered into a transitional supply agreement..." A capacity purchase consistent with the "manufacturing expansion over the coming quarters" language, and absent from every vendor field.

4. Valuation — priced in or room?

At $849.47 (market cap $66.09B, 77.8M shares). Lumentum's fiscal year ends in late June; the estimate rows are labelled by the fiscal year they cover.

Estimate coverage is good on the anchor years and collapses beyond. FY2027 rests on 14 analysts for EPS and 18 for revenue, with an EPS range of $15.275 to $20.667 — a 35.3% spread, which is wide and appropriate for a business growing 90%. FY2028 rests on 7 analysts for EPS. FY2029 rests on ONE analyst for both revenue and EPS and is excluded from every conclusion in this dive — a single-analyst estimate is not a consensus and printing a 19x FY2029 multiple would be a fabrication.

est.ebitAvg is REJECTED and it trips both documented signatures at once, more severely than any other name in this batch. ebitAvg is NEGATIVE in every forward year — −$20.6M (FY2026), −$39.0M (FY2027), −$60.3M (FY2028), −$84.0M (FY2029) — while ebitdaAvg is strongly positive at $405.1M, $765.8M, $1,184.3M and $1,648.0M, and netIncomeAvg is positive at $563.0M, $1,250.8M, $1,996.6M and $2,895.1M. A company cannot earn $1,996.6 million of net income on MINUS $60.3 million of operating profit. Separately, ebitdaAvg is exactly 13.500% of revenueAvg in every year from FY2024 to FY2029. Both the negative-EBIT signature and the fixed-ratio signature fire. All forward valuation uses epsAvg only, which validates well: the FY2026 consensus of $8.225 against three reported quarters totalling $5.14 plus a Q4 consensus of $2.97 gives $8.11, within 1.4%.

Enterprise value — CLEAN, with no possible omission. enterpriseValueTTM of $66,784.7M less market capitalisation of $66,088.8M implies net debt of $695.9M, consistent with the note repurchases and conversions documented in Section 2. minorityInterest is 0 and preferredStock is 0, so the noncontrolling-interest and preferred omissions this programme finds on most names are structurally impossible here. The multiple is usable, though at 22.3x FY2026 consensus revenue it is not doing much work.

Peer context — the set is close to useless. The vendor peers are Bilibili (a Chinese video platform), Bentley Systems, DocuSign, Entegris, F5, Fabrinet, Klarna, Kaspi.kz, Rubrik and Zebra Technologies. Only Fabrinet — Lumentum's own contract manufacturer, and the likely explanation for the $291.8 million of Thailand revenue — has any relationship to this business. Coherent, the company named alongside Lumentum in all three knowledge-base claims, is absent, as are Corning, Marvell and every other optical peer. No peer multiple comparison is drawn.

4a. What today's price assumes (the inversion)

At $849.47 — 103.3x FY2026 consensus, 45.7x FY2027, 22.3x FY2026 revenue — the price embeds:

• Revenue reaches $3,001 million in FY2026 and $5,673 million in FY2027. (Consensus; 16 and 18 analysts.) The first nine months delivered $2,007.7 million, so FY2026 needs approximately $993 million in the fourth quarter against a consensus of $988 million — essentially in hand, and reportable in seven days. FY2027's $5,673 million requires 89.0% growth on top, which is the assumption doing the work.
• Both largest customers stay and keep growing. (Our derivation; the 26% and 12% figures are the company's.) This is the most fragile assumption in the price. Nothing in this archive discloses contract length, and in a components business a design loss removes a revenue line rather than eroding it.
• Laser chip pricing power persists through capacity additions. (Company disclosure, half-weighted.) Average selling prices rose on the 200G transition. Every optical component in history has eventually met competing capacity; cloud transceiver ASPs, disclosed in the same paragraph, are already falling.
• The market keeps paying 40-50x forward earnings. (Our number.) At 35x FY2027E the stock is $650; at 60x it is $1,115. A 25-turn band on a business growing 89% is not analyst indecision — it is the honest width of a valuation for something at this stage.
• Convertible dilution is absorbed. (Our number; the terms are the company's.) $500.8 million of early conversion requests outstanding, 5.7 million shares issued in one April transaction, and a share count already up 11.8% in ten months. Consensus EPS is struck on estimates of a share count that keeps rising, which is a persistent downward pressure on the denominator that the revenue line does not face.

4b. The return bridge (why the multiple moves)

Expected return over the next twelve months decomposes as: EPS growth (+125.9%, FY2026E $8.225 to FY2027E $18.579) + multiple compression (assumed LARGE, 103.3x on the forward year to about 53.8x on the then-forward year, −48%) + no dividend ≈ +18%.

That decomposition should be read carefully because it is unusual. The base case assumes earnings more than double AND the multiple almost halves, and the two nearly cancel. That is the correct way to value a company at this stage: the multiple must compress, because 103x forward earnings on 82% revenue growth is a statement about the growth rate rather than about the business, and growth rates decelerate by arithmetic if nothing else — consensus itself has revenue growth going 82.4%, 89.0%, 54.7%.

The practical consequence, stated plainly: almost none of the expected return is re-rating and all of it is earnings. That is the return shape this programme prefers. What it does not do is make the entry safe — because if the earnings do not double, the multiple compresses anyway and there is nothing to catch the price. The bear case at $500 is exactly that scenario and it is 41.1% down.

If the multiple compressed to 35x FY2027E the price would be $650 (−23.5%). At 60x, $1,115 (+31.3%).

4c. Variant perception (where we differ, what would surprise)

• We are BELOW the street and modestly so. Our base of $1,000 is 7.6% below the consensus target of $1,082.67, with the median at $1,100. The street's LOWEST published target, $900, is already 5.9% above the current price, and 18 of 25 analysts rate the shares Buy or better with no sell ratings at all. We are not contrarian on this name; we are a shade more cautious and considerably more explicit about the dilution.
• We think the knowledge base's claim is unusually testable and we tested it. Three independent commentators say optical interconnect is a physical choke point in AI infrastructure and that Lumentum is on the short list. Section 6 checks that against the 10-Q and finds two pieces of primary corroboration: laser chip average selling prices RISING on the 200G transition, and optical circuit switch revenue appearing from nothing at more than $25.0 million in a quarter with manufacturing expansion under way. A price increase in a components business is the hardest available evidence of scarcity, and it is in a filed document. This is the closest this programme has come to primary confirmation of an expert claim, and it is why the conviction reads positive-moderate rather than positive-low.
• We think the convertible dilution is materially under-discussed. The share count rose 11.8% in ten months; 5.7 million shares — 7.3% — were issued in a single transaction on 2026-04-07; $500.8 million of principal has requested early conversion. Watchable number: shares outstanding on the 2026-08-11 10-K cover, against 77.8 million at 2026-04-30. Anyone modelling this company on a fixed share count is modelling a different security.
• We think the two-sided pricing disclosure is the key operating tension and nobody is framing it as one. Laser chip prices are rising; cloud transceiver prices are falling — both stated in the same paragraph of the same filing. Watchable number: gross margin, 44.2% in the March quarter against 28.8% a year earlier. If it keeps expanding, the components mix is winning; if it stalls above 40%, transceiver pricing is catching up.
• Positive surprise that would force a re-rate: optical circuit switch revenue above $75 million in the fourth quarter, which would make it a franchise rather than a pilot; fiscal 2027 guidance above the $5,673 million consensus; gross margin above 47%; or disclosure of a third customer above 10% of revenue, which would reduce concentration rather than increase it.
• Negative surprise that would break the thesis: the largest customer falling below 20% of revenue without a replacement; gross margin below 41%; fourth-quarter revenue below $940 million against the $988 million consensus; any indication that 200G laser pricing has peaked; or a large equity issuance beyond the mechanical convertible settlements.

Synthos fair values

All three anchors are multiples of the FY2027 consensus EPS distribution (mean $18.579, low $15.275, high $20.667, 14 analysts), cross-checked against FY2028 and against enterprise value to consensus revenue. The range is deliberately very wide, because a 35.3% analyst EPS spread on a business growing 89% does not support false precision.

• Bear ~$500 — 32.7x the FY2027 consensus LOW of $15.275, and 16.9x FY2028E. Cross-check: 6.9x FY2027 consensus revenue; 362% above the 52-week low of $108.15 and 39.2% below the 200-day average. The scenario: one of the two large customers reduces or loses its socket, transceiver price declines outrun volume, gross margin falls back below 40%, the AI infrastructure order book pauses, and a 1.513-beta stock that rose sevenfold gives back the last twelve months of multiple. −41.1%.
• Base ~$1,000 — 53.8x the FY2027 consensus MEAN of $18.579, and 33.9x FY2028E. Cross-check: 13.7x FY2027 consensus revenue; 5.0% below the 52-week high of $1,053.09; 7.6% below the street consensus of $1,082.67. Sensitivity, stated openly: 45.7x FY2027E gives $849 — spot exactly — and 60x gives $1,115. The base case is a modest multiple expansion on an earnings estimate that more than doubles, and Section 4b says which leg is doing the work. The scenario: fiscal 2026 completes near $3.0 billion, fiscal 2027 grows as modelled, both customers stay, optical circuit switches scale, and the multiple compresses less than the earnings grow. +17.7%.
• Bull ~$1,400 — 67.7x the FY2027 consensus HIGH of $20.667, and 47.4x FY2028E. Cross-check: 19.2x FY2027 consensus revenue; 33.0% above the 52-week high; 10.2% above the highest published street target of $1,270. The scenario: optical circuit switches become a several-hundred-million-dollar franchise, laser chip pricing power extends through the 400G transition, customer concentration falls as new hyperscalers qualify the parts, and the market pays a scarcity multiple for the only alternative to a duopoly. +64.8%.

Base is 17.7% above spot; asymmetry roughly 1.58:1 (41.1% down, 64.8% up), with no dividend. A 17.7% base is respectable and a 1.58:1 payoff ratio on a 1.513-beta stock seven days before a print is not enough to earn a Buy tier. The distribution here is genuinely wide and the resolution date is known — that combination argues for waiting, not for sizing down.

5. Knowledge base — nine raw hits, three survivors, unanimous, and CORROBORATED BY THE FILING

Raw hits: 9. After a case-sensitive entity re-run: 3. Name-level claims on Lumentum: 3, and ALL THREE are BULLISH. Discarded: 6.

The primary sweep ran the entity terms LITE, Lumentum and Lumentum Holdings, plus free text on optical transceiver, silicon photonic and co-packaged, across all 52,021 distilled claims. It returned just 9 hits across 7 channels — the second-smallest raw lane in this batch, after Kinder Morgan's 12 — which means the homograph problem that destroyed the sweeps on ALL (1,955 hits), TGT (441), CRH (313) and AEP (243) barely applies here. The six discards are jensen_huang (3), plus one each from quarantine_misattributed, mike_green, real_vision, dylan_patel, andreas_steno and jordi_visser — general commentary on optics and accelerated computing that names no company.

The three surviving claims, verbatim:

> 2026-04-15 · BULLISH · conviction 74 · horizon: thesis · entities: Lumentum, Coherent · channel: jensen_huang · no named speaker

> "Nvidia's multi-year investment in Lumentum, Coherent, and silicon photonics (COUPE) reshaped the supply chain into mainstream computing tech."

> 2026-03-02 · BULLISH · conviction 62 · horizon: thesis · entities: LITE, COHR · channel: andreas_steno · speaker: Andreas Steno · role: independent

> "Lumentum and Coherent own the efficient optical switches sending data between GPU racks — a choke point as electricity becomes the AI bottleneck; Nvidia is investing ~a couple billion in each."

> 2026-04-26 · BULLISH · conviction 75 · horizon: thesis · entities: COHR, GLW, MRVL, LITE · channel: jordi_visser · speaker: Jordi Visser · role: independent

> "Optical is necessary for inference and power efficiency — Coherent, Lumentum, Corning, plus Marvell (silicon-photonics leader, bought another company this week); transceiver market in shortage."

This is the strongest knowledge-base lane in this batch, on five grounds, each stated with its qualification.

Unanimous direction. All three bullish. No countervailing claim exists in the store.

Three genuinely independent channels. jensen_huang, andreas_steno and jordi_visser — unlike HPE, where two of three claims came from the same speaker. Breadth is a real 3.

Current. 2026-03-02, 2026-04-15 and 2026-04-26 — all within five months of this dive.

Two named speakers with role: independent, which is stronger sourcing than the speaker: null channel attribution that dominates this store and that produced the 2026-07-25 attribution failure.

And — uniquely in this programme so far — the mechanism is testable against primary documents, and it tests positive. The claims assert three specific things. We checked each against the 10-Q:

Two of three assertions are confirmed by the company's own filed document. The third — an Nvidia investment of "a couple billion" — appears in two of the three claims and in none of the filings available here, and it is recorded as unverified rather than accepted.

And the qualifications, because a strong lane still needs them.

All three name Lumentum alongside Coherent as a pair. Not one distinguishes between them, discusses relative market share, names a price, or mentions the customer concentration, the convertible dilution or the transceiver price declines. These are correct calls on an industry, expressed through two companies.

All three were written after the stock had already risen several hundred percent. The earliest is 2026-03-02; the twelve-month return through this dive is 696.3%. A bullish claim made partway up a sevenfold move is a good call about the industry and it is not a discovery of value.

And one attribution note. The jensen_huang claim carries speaker: null despite the channel name, which per the documented store defect means it is channel-attributed commentary rather than verified first-person speech. It is weighted as channel-attributed and its unverified Nvidia-investment assertion is treated accordingly.

Conclusion. Breadth 3, claim count 3, net conviction positive-moderate — the highest in this batch. Three independent, current, mostly named-speaker claims pointing the same way, with two of their three testable assertions confirmed by the company's own 10-Q. That is a real lane. It is also a lane about an industry rather than a security, written after a sevenfold move, and it says nothing about the price, the customers or the share count — which is exactly why the verdict is Watch and not Buy.

6. Data integrity — what we rejected and why

Six findings, plus an archive-level limitation stated first.

0. THE FILING ARCHIVE IS PROSE-ONLY for the financial statements — zero [TABLE] markers in the 10-Q filed 2026-05-06. The check was run and returned 0 tables. The consolidated statements of operations, balance sheet and cash flows could not be read as tables — the 10-Q's text explicitly says "See accompanying Notes" where each statement should be. However, unlike the KMI archive, this filing carries extremely rich narrative disclosure, and everything material in this dive is quoted from prose: the 90.1% revenue increase and its $232.5M/$150.7M split, the laser chip and transceiver pricing directions, the optical circuit switch revenue, the customer concentration, the entire convertible-note structure and terms, the Greensboro acquisition, the segment reorganisation and the share count. Each such figure is labelled filing-verified. What could NOT be verified against the filing is the quarterly income statement line-by-line, the balance sheet and the cash-flow statement, all of which come from the vendor and are labelled as such.

1. est.ebitAvg is NEGATIVE in every forward year while netIncomeAvg is large and positive — REJECTED, and this is the most severe estimate defect in the batch. ebitAvg reads −$20.6M (FY2026), −$39.0M (FY2027), −$60.3M (FY2028), −$84.0M (FY2029) against netIncomeAvg of $563.0M, $1,250.8M, $1,996.6M and $2,895.1M. A company cannot earn $2.0 billion of net income on a $60 million operating loss. Separately, ebitdaAvg is exactly 13.500% of revenueAvg in every year from FY2024 through FY2029 — the fixed-ratio fabrication signature. Both documented signatures fire simultaneously. All forward valuation uses epsAvg only, which cross-checks to within 1.4% of the sum of three reported quarters plus the Q4 consensus.

2. seg_prod reports a segment structure TWO reorganisations out of date — REJECTED entirely. The vendor's most recent row is fiscal 2023 and contains a single line, "Lasers Segment" at $209.2 million, against consolidated revenue of $1,767.0 million — 88.2% of the business missing. Earlier rows use "Optical Communications Segment" and "Lasers Segment." The 10-Q states the company operated in "Cloud & Networking and Industrial Tech" until fiscal 2026 and now operates in a SINGLE reporting segment. The block cannot be reconciled to anything the company has reported in three years and no figure in this dive comes from it.

3. inc_a contains a DUPLICATED annual row. The FY2021 (2021-07-03) and FY2022 (2022-07-02) rows report identical revenue ($1,712.6M), identical cost of revenue ($924.0M), identical gross profit ($788.6M), identical net income ($198.9M), identical EPS ($2.79), identical share count (74.2M), identical EBITDA ($482.4M) and identical D&A ($167.1M) — differing only by $1.1 million of operating expenses. Two consecutive fiscal years cannot be identical to that precision. Neither row is used for anything in this dive; the FY2020, FY2023, FY2024 and FY2025 rows are used and are internally coherent.

4. inc_q contains a DUPLICATED quarterly row. The rows dated 2024-03-30 and 2024-09-28 both report revenue of $336.9M, gross profit of $77.9M, operating income of −$82.4M, net income of −$82.4M, interest expense of $5.5M and EBITDA of −$3.1M — identical in every field. One of the two is a repeat. Neither is used for any conclusion; the seven-quarter growth series in Section 1 begins at 2024-06-29 and the pre-2025 quarters are used only for context.

5. Derived leverage ratios CONTRADICT each other. debtToCapitalRatioTTM reads 0.5271 while longTermDebtToCapitalRatioTTM reads 0.0143 — a 37-fold difference between two ratios whose numerators should be nearly the same for a company financed almost entirely by long-dated convertible notes due 2026 through 2032. debtToEquityRatioTTM of 1.1144 is likewise inconsistent with debtToMarketCapTTM of 0.0497 given a book equity of $3.24 billion implied by bookValuePerShareTTM of $41.586. This is the "derived ratios contradicting the raw block" defect class. None of the leverage ratios is used; Section 3 uses the filed note principal amounts and the enterprise-value-implied net debt instead.

6. Trailing profitability metrics are distorted by a tax benefit and are not used. Fiscal 2025 net income of $25.9 million was produced on an operating loss of $192.2 million by an income-tax BENEFIT of $198.0 million. The consequences run through the payload: effectiveTaxRateTTM of MINUS 60.5%, taxBurdenTTM of 1.605, netIncomePerEBTTTM of 1.605 and ebtPerEbitTTM of 1.091 — all arithmetically consistent with a valuation-allowance release and all economically meaningless as measures of the business. priceToEarningsRatioTTM of 136.8x and priceToEarningsDilutedRatioTTM of 157.3x are not used anywhere in this dive.

Verified CLEAN — recorded because clean checks are findings:

• Share count — EXACT. The 10-Q cover states 77.8 million shares outstanding at 2026-04-30 against 77,800,000 implied by market capitalisation ÷ price. An exact match. (Note that this makes the vendor's FY2025 weightedAverageShsOutDil of 69.6 million 11.8% stale relative to the current count — correct for the year it labels, misleading for anything current, as Section 2 explains.)
• capitalExpenditure is CLEAN and free cash flow is computed CORRECTLY. freeCashFlowOperatingCashFlowRatioTTM reads 0.605, correctly below 1.0; operatingCashFlowPerShareTTM $6.327 less capexPerShareTTM $2.498 equals freeCashFlowPerShareTTM $3.829 to three decimals; capexToDepreciationTTM of 0.702 is plausible for a company adding capacity. The resulting 0.41% free-cash-flow yield is economically trivial and arithmetically right.
• Enterprise value — CLEAN, and no omission is possible. enterpriseValueTTM of $66,784.7M less market capitalisation of $66,088.8M implies $695.9M of net debt, consistent with the convertible repurchases and conversions documented in the 10-Q. minorityInterest and preferredStock are both 0, so this programme's two standard EV omissions cannot occur.
• seg_geo is CLEAN and reconciles exactly. Americas $480.9M + Asia Pacific $1,000.6M + EMEA $163.5M = $1,645.0M, exactly FY2025 consolidated revenue, with each region's country detail summing to its own aggregate.
• tech.max_dd_from_peak of −19.335% coincides exactly with pct_from_hi. Per the data contract this is correct, not a defect — the six-year peak falls inside the last twelve months, which for a stock up 696% is trivially true. Described as the current distance from the high, never as a maximum drawdown.
• tech versus quote. tech.last of $849.47 equals quote.price exactly. quote.yearHigh/yearLow ($1,085.68/$106.16) against tech.hi52/lo52 ($1,053.09/$108.15) — 3.1% and 1.9%. tech is used throughout.

Vendor composite rating — C+ / 2 overall, with priceToEarningsScore 1, priceToBookScore 1, debtToEquityScore 1 and discountedCashFlowScore 1. The rating block is uniformly negative and it is measuring trailing statistics on a company whose trailing statistics are meaningless — a $25.9 million net income produced by a tax benefit, a 20.4x book multiple and a 0.41% free-cash-flow yield. It is arithmetically defensible and analytically useless on a business growing revenue 90%, and it carries no weight in this verdict.

Non-equity tripwire — checked and passed. LITE is common stock, NASDAQ-listed. Price of $849.47 is not par-like; beta is 1.513, the highest in this batch; volume was 6.51M shares (~$5.5 billion of turnover, by far the heaviest in this batch); the 52-week band of $108.15 to $1,053.09 is a 874% range. No preferred stock, no noncontrolling interests, no dividend. Note the four series of convertible senior notes described in Section 2, which are NOT this security but which convert into it. This is common equity.

7. Technicals

• Price $849.47. −19.3% from the 52-week high of $1,053.09; +685.5% above the 52-week low of $108.15. Position within the annual range: 78th percentile.
• Both moving averages are below the price. +3.3% above a 50-day average of $822.08; +39.8% above a 200-day average of $607.73. The 50-day proximity is the notable one — the stock is barely above its own quarter-average, which is what a consolidation looks like.
• RSI 57.3 — neutral, and remarkably so after a 696% year. MACD −18.63 — NEGATIVE, the only large negative MACD in this batch, reflecting the three-month decline.
• Current distance from the six-year peak: −19.3%, identical to the 252-day figure. Per the data contract that coincidence is correct.
• Relative performance, and the shape matters: 3-month −13.0% against SPY +7.6% and QQQ +7.7% — underperforming by 21 points in the quarter; 6-month +100.6% against SPY +11.1%; 12-month +696.3% against SPY +24.3% and QQQ +30.8%. This is not a stock at a peak; it is a stock that went up sevenfold, then fell 19% from its high and spent three months going sideways-to-down while the index rose. That is a materially better technical entry than the raw twelve-month number suggests, and it is the strongest single argument against the Watch verdict.
• Sentiment: 1 strong buy, 17 buy, 7 hold, 0 sell across 25 analysts; consensus target $1,082.67 (+27.5%), median $1,100, high $1,270 (+49.5%), low $900 — 5.9% ABOVE spot. No covering analyst has a target below the price and none rates it sell. As elsewhere in this programme, unanimity about the absence of downside on a 1.513-beta stock is treated as a caution rather than a comfort.

Today's move and what it does to the entry

LITE closed 2026-08-04 at $849.47, up $69.58 or 8.92%, from $779.89 — the largest single-day move in this batch. It opened at $854.00, traded $813.26 to $868.32, and closed on 6.51 million shares, approximately $5.5 billion of turnover. No filing is dated 2026-08-04; the most recent is the 8-K of 2026-07-30 announcing an executive retirement.

The honest read, and it cuts both ways. An 8.92% move without a company filing, seven days before results, on the heaviest turnover in this batch, is positioning rather than news. Against that, the stock is 19.3% below its high and down 13.0% over three months, so this is a bounce inside a consolidation and not a breakout to new ground. The entry is therefore better than the twelve-month return implies and worse than it was three days ago, and neither observation changes a verdict that rests on the 2026-08-11 print.

8. Insiders — eight filings, all director grants, no discretionary transaction

Every transaction is a director equity award at zero price. There is not one open-market purchase, not one open-market sale, not one tax withholding and not one transaction by any executive officer in the file.

Two observations.

The grants are tiny and that is the share price at work. Six directors each received exactly 108 shares on 2026-07-15 — approximately $91,700 each at the closing price. A uniform 108-share grant across six directors on one date is a fixed-dollar retainer converted at the prevailing price, and the smallness of the share number is a direct measure of how far the stock has run: at the 52-week low of $108.15 the same dollar value would have been roughly 840 shares.

The absence is the more interesting half. No transaction by chief executive Michael E. Hurlston, by the chief financial officer, or by any executive officer appears in the eight most recent filings. On a stock that has risen 696% in twelve months, the complete absence of executive selling is mildly reassuring and is weighted as such — mildly. Directors' holdings range from 1,313 shares (Thad Trent) to 21,116 (Ian Small), worth approximately $1.1 million to $17.9 million.

Separately, and not in the Form 4 file: the 8-K filed 2026-07-30 discloses that on 2026-07-27, Vincent Retort, Executive Vice President, Global Reliability & Quality, informed the Company of his planned retirement effective in October 2026, with a two-year consulting term during which "his outstanding equity awards will continue to vest... provided that he remains as a service provider." A quality-function executive retiring with a two-year vesting bridge at a company scaling manufacturing capacity is worth noting and is not, on this evidence, worth more than that.

9. Verdict, kill-criteria and flip conditions

Watch.

The case for the business — the strongest in this batch, and stated at full strength. Net revenue in the March quarter up 90.1% to $808.4 million, an increase of $383.2 million split $232.5 million Components and $150.7 million Systems. Operating income from MINUS $65.4 million to PLUS $175.6 million, a 21.7% margin. Gross margin from 28.8% to 44.2%. Seven consecutive quarters of sequential revenue growth, from $308.3 million to $808.4 million. Laser chip average selling prices RISING on the 200G transition — the hardest available evidence of genuine scarcity in a components business. Optical circuit switches from zero to "more than $25.0 million" in a quarter with manufacturing expansion under way. Consensus revenue of $3,001 million, $5,673 million and $8,772 million across three years. Net debt of roughly $696 million against a $66.1 billion market capitalisation, on notes with no financial covenants. And a knowledge-base lane of three independent, current, mostly named-speaker claims pointing the same way, two of whose three testable assertions are confirmed by the company's own filing.

Why this is Watch and not Buy — Tactical, as four specific objections.

First, price. 103.3x FY2026 consensus and 45.7x FY2027, 20.4x book and 39.3x tangible book, after a 696.3% twelve-month return. Our base of $1,000 is +17.7% on 1.58:1 asymmetry — respectable, and not enough on a 1.513-beta security.

Second, concentration. "Two customers individually accounted for 26% and 12% of our total revenue." Thirty-eight per cent in two accounts, with no disclosed contract terms, in an industry where design losses remove revenue lines whole.

Third, dilution, which is active. $1,265.0 million of 2032 Notes convertible at approximately $187.77 against an $849.47 price. $500.8 million of early conversion requests outstanding at 2026-04-30. 5.7 million shares — 7.3% of the count — issued in a single transaction on 2026-04-07. A share count up 11.8% in ten months. Every consensus EPS figure in Section 4 is struck on a denominator that keeps growing.

Fourth, timing. Results land 2026-08-11, seven days from now. Every upgrade condition below is designed to be answered by that report, and waiting a week for information that materially narrows a 41%-down / 65%-up distribution is not caution — it is the correct use of a known date.

And the honest counterweight, which is strong. The stock is 19.3% below its high and down 13.0% over three months while the index rose 7.6%. This is a consolidation, not a peak, and RSI at 57.3 after a sevenfold move says the advance has been orderly. If the 2026-08-11 print delivers $988 million and guides fiscal 2027 at or above $5,673 million, the FY2027 multiple falls toward 40x on an unchanged price and this becomes a Buy on arithmetic.

Pre-registered KILL criteria — what would take this to Avoid:

• Fourth-quarter revenue below $940 million on 2026-08-11 against the $988 million consensus, which would break a seven-quarter sequential growth streak.
• Gross margin below 41%, giving back a third of the 15.4-point expansion, which would mean transceiver price declines are outrunning laser chip pricing power.
• The largest customer falling below 20% of revenue without a disclosed replacement — the concentration is a risk in both directions, and losing it is the worse one.
• Fiscal 2027 guidance below $5.0 billion against a $5,673 million consensus.
• Optical circuit switch revenue failing to exceed $25 million in the fourth quarter, indicating the pilot has not scaled.
• Any equity issuance beyond the mechanical convertible settlements, or a share count above 85 million on the fiscal-2026 10-K cover against 77.8 million at 2026-04-30.
• A break below the 200-day average of $607.73, which on a 1.513-beta stock would have very little beneath it.

Pre-registered UPGRADE conditions — what would take this to Buy — Tactical, all resolvable on or shortly after 2026-08-11:

• Fourth-quarter revenue above $1.0 billion with fiscal 2027 guidance at or above the $5,673 million consensus. This is the cleanest trigger and it converts the base case from a multiple argument into an earnings one.
• Optical circuit switch revenue above $75 million in the fourth quarter, which would make it a franchise rather than a pilot and would directly confirm the knowledge base's most specific claim.
• Gross margin above 47%, demonstrating that laser chip pricing power is outrunning transceiver declines.
• Disclosure of a third customer above 10% of revenue, which would reduce the single hardest risk in this file.
• A price below approximately $700, which is 37.7x FY2027E and would put the base case above 42%.
• Confirmation in a filed document of the Nvidia investment that two of the three knowledge-base claims assert and that no filing in this archive mentions.

Where LITE fits in the Synthos Framework Portfolio. No position today, with an explicit calendar trigger: revisit within three days of the 2026-08-11 print. The technology sleeve would take this name at a target weight of 1.5% on either of the first two upgrade conditions, entered in halves. Sizing note, and it is the operative discipline: this is a 1.513-beta security with a 41%-down bear case, two customers at 38% of revenue and a print in seven days — the correct response to a wide distribution with a known resolution date is to wait for the resolution, not to size around it. Logged as a tracked Synthos call (Watch) as of 2026-08-04 at $849.47, with the fair-value anchors, kill criteria and upgrade conditions all gradeable.

Single biggest risk: two customers are 38% of revenue. The 10-Q states it without qualification: "During the three months ended March 28, 2026, two customers individually accounted for 26% and 12% of our total revenue." In an optical components business, revenue is won socket by socket inside a customer's design cycle, and a competitor winning the next generation removes the line rather than eroding it. Nothing in this archive discloses contract duration, backlog or qualification status for either relationship. Against a market capitalisation of $66.1 billion built on consensus revenue reaching $8,772 million by June 2028, that is a great deal of value resting on two purchasing decisions. The mitigating facts are real — laser chip prices are rising, optical circuit switches are a new category ramping from zero, the growth is broad across intra-data-centre, interconnect, long-haul and sub-sea applications, and three independent expert voices identify a genuine physical bottleneck that the filing corroborates. This is, on the evidence in this file, the best business in this batch. It is also priced at 45.7 times a forward estimate, financed with notes converting at a fifth of the current price, and reporting in seven days — and the discipline this programme exists to enforce is that a great business bought at the wrong moment is still a bad decision.

Provenance & disclosures

• Traceability: 3 name-level knowledge-base claims name Lumentum out of 52,021 distilled claims (raw hits 9, surviving a case-sensitive entity re-run 3, entity matches 3, discarded 6; breadth 3 independent channels, net conviction positive-moderate — the highest in this batch). All three are BULLISH, all are dated within five months of this dive, and two carry NAMED SPEAKERS with role: independent. They are: jensen_huang, 2026-04-15, conviction 74, speaker: null, entities Lumentum/Coherent — "Nvidia's multi-year investment in Lumentum, Coherent, and silicon photonics (COUPE) reshaped the supply chain into mainstream computing tech."; andreas_steno, 2026-03-02, conviction 62, speaker Andreas Steno, role independent, entities LITE/COHR — "Lumentum and Coherent own the efficient optical switches sending data between GPU racks — a choke point as electricity becomes the AI bottleneck; Nvidia is investing ~a couple billion in each."; jordi_visser, 2026-04-26, conviction 75, speaker Jordi Visser, role independent, entities COHR/GLW/MRVL/LITE — "Optical is necessary for inference and power efficiency — Coherent, Lumentum, Corning, plus Marvell (silicon-photonics leader, bought another company this week); transceiver market in shortage." Uniquely in this programme, the mechanism these claims assert was TESTED against the primary filing and two of its three testable assertions were CORROBORATED: optical circuit switch revenue of "more than $25.0 million" in the March quarter with manufacturing expansion under way, confirming the "optical switches between GPU racks" claim; and "a slight increase in average selling prices of laser chip products... driven primarily by a shift to 200G lane speeds", confirming the "choke point" and "shortage" claims by the hardest available evidence, which is price. The third assertion — an Nvidia investment of "a couple billion", appearing in two of the three claims — is NOT FOUND anywhere in the extractable text of this archive and is recorded as UNVERIFIED rather than accepted. The qualifications are stated in Section 5 and stand: all three name Lumentum alongside Coherent as a pair rather than as a differentiated pick; none discusses the customer concentration, the convertible dilution, the transceiver price declines or the price; and all three were written after the stock had already risen several hundred percent. The 6 discards are general optics and accelerated-computing commentary from jensen_huang (3) plus one each from quarantine_misattributed, mike_green, real_vision, dylan_patel, andreas_steno and jordi_visser, none of which names this company. At 9 raw hits this is the second-smallest sweep in the batch and the homograph problem that destroyed the ALL (1,955), TGT (441), CRH (313) and AEP (243) sweeps barely applies. All quotes are verbatim from the stored claim text.
• Data as-of: fundamentals — revenue growth and its disaggregation, product-level pricing direction, optical circuit switch revenue, customer concentration, the complete convertible-note structure and terms, the Greensboro acquisition, the segment reorganisation and the share count through 2026-03-28 (share count at 2026-04-30), all filing-verified from the prose of the 10-Q filed 2026-05-06; annual and quarterly financial statements from the vendor payload, unverified against the filings for the reason in Section 6 finding 0 · estimates 2026-08-04 · prices 2026-08-04, quote timestamp 1785873601 = 2026-08-04T20:00:01Z ($849.47, +8.92%; 50-DMA $822.08; 200-DMA $607.73; RSI 57.3; MACD −18.63) · knowledge-base claims 2026-08-04. Lumentum's fiscal year ends in late June; "FY2026" means the year ending 2026-06-28 and the estimate rows are labelled by the fiscal year they cover. All figures come from the Synthos vendor data file for LITE or from the SEC filings in the LITE archive; no figure comes from memory, recall or external retrieval.
• Filing archive contents, with the archive-level limitation stated first: the check for preserved tables was run on the 10-Q filed 2026-05-06 and returned ZERO [TABLE] markers, so the consolidated statements of operations, balance sheet and cash flows could not be read as tables and come from the vendor. The narrative disclosure is nonetheless exceptionally rich and every material qualitative and product-level figure in this dive is quoted from it. The archive also holds: 10-K filed 2025-08-19 (fiscal year ended 2025-06-28); 10-Q filed 2026-02-04 (December 2025 quarter); 10-Q filed 2026-05-06 (March 2026 quarter — source of the 90.1% revenue increase and its Components/Systems split, the laser chip and cloud transceiver pricing directions, the optical circuit switch revenue, the "two customers 26% and 12%" concentration disclosure, the 2032 Notes terms and conversion price, the $500.8 million of early conversion requests, the 5.7 million-share April exchange, the capped call description, the Greensboro acquisition and the single-segment reorganisation); 8-K filed 2026-05-05 (Item 2.02, third-quarter results — the Exhibit 99.1 press release and the Q3 FY26 conference-call slide presentation are referenced but not in this archive); 8-K filed 2026-06-01 (no extractable item content); 8-K filed 2026-07-30 (Item 5.02 — Vincent Retort, Executive Vice President, Global Reliability & Quality, notified the company on 2026-07-27 of his planned retirement effective October 2026, with a two-year consulting term during which outstanding equity awards continue to vest).
• Where the filings contradicted, corrected or CONFIRMED the vendor (detailed in Section 6): est.ebitAvg, negative in every forward year (−$20.6M to −$84.0M) while netIncomeAvg runs $563.0M to $2,895.1M — arithmetically impossible — with ebitdaAvg simultaneously fixed at exactly 13.500% of revenueAvg in every year from FY2024 to FY2029, so both documented estimate signatures fire at once; seg_prod, which reports a structure two reorganisations out of date, whose most recent row (fiscal 2023) contains a single line at $209.2M against $1,767.0M of revenue — 88.2% missing — against a 10-Q that states the company now operates in a single reporting segment; inc_a, whose FY2021 and FY2022 rows are identical in eight fields and differ only by $1.1M of operating expenses; inc_q, whose 2024-03-30 and 2024-09-28 rows are identical in every field; derived leverage ratios that contradict each other, with debtToCapitalRatioTTM of 0.5271 against longTermDebtToCapitalRatioTTM of 0.0143 for a company financed almost entirely by long-dated convertible notes; and trailing profitability metrics distorted by a $198.0 million income-tax BENEFIT in fiscal 2025, producing an effectiveTaxRateTTM of −60.5% and a netIncomePerEBTTTM of 1.605. Three corporate actions are absent from every vendor field: the $1,265.0 million 2032 Note issuance of 2025-09-08 with its $581.1 million concurrent repurchase of 2026 Notes and $102.0 million of capped calls; the 2026-04-07 privately negotiated exchange issuing approximately 5.7 million shares for approximately $474.6 million of principal; and the $38.0 million Greensboro, North Carolina manufacturing facility acquisition of 2026-03-17. Where vendor and filing AGREED — recorded, because clean checks are findings: the share count is an EXACT match (10-Q cover 77.8 million at 2026-04-30 against 77,800,000 implied); capitalExpenditure is CLEAN with freeCashFlowOperatingCashFlowRatioTTM of 0.605 correctly below 1.0 and the per-share subtraction exact to three decimals; enterprise value is CLEAN and no omission is structurally possible, minorityInterest and preferredStock both being zero; seg_geo reconciles EXACTLY to $1,645.0M of FY2025 revenue with each region's country detail summing to its aggregate; and tech.max_dd_from_peak of −19.335% coincides with pct_from_hi correctly.
• Estimate coverage: 13 analysts on FY2026 EPS and 16 on revenue; 14 on FY2027 EPS and 18 on revenue — the anchor for all three fair values, with an EPS range of $15.275 to $20.667, a 35.3% spread that is wide and appropriate for a business growing 89%; 7 on FY2028 EPS and 17 on revenue; ONE analyst on FY2029 for both revenue and EPS, which is not a consensus and is excluded from every conclusion in this dive.
• Peer note: the vendor peer set — Bilibili, Bentley Systems, DocuSign, Entegris, F5, Fabrinet, Klarna, Kaspi.kz, Rubrik, Zebra Technologies — is substantially unusable, containing a Chinese video platform, two payments companies and several enterprise-software names. Only Fabrinet has any relationship to this business, and it is Lumentum's contract manufacturer rather than a competitor — the likely explanation for the $291.8 million of Thailand ship-to revenue in seg_geo. Coherent, named alongside Lumentum in ALL THREE knowledge-base claims, is absent, as are Corning and Marvell. No peer-multiple comparison is drawn.
• Fair-value caveat: the $500 / $1,000 / $1,400 anchors are multiples of the FY2027 consensus EPS distribution — 32.7x the low of $15.275, 53.8x the mean of $18.579, and 67.7x the high of $20.667 — cross-checked against FY2028 EPS (16.9x / 33.9x / 47.4x) and against enterprise value to FY2027 consensus revenue (6.9x / 13.7x / 19.2x). Stated arithmetic, not a discounted cash flow. The base is sensitivity-disclosed: 45.7x FY2027E gives $849, which is spot exactly; 60x gives $1,115. The range is deliberately very wide because a 35.3% analyst EPS spread on a business growing 89% does not support false precision, and no greater precision is claimed than the estimate distribution permits. Section 4b states explicitly that the base case assumes earnings more than double AND the multiple almost halves, and that the two nearly cancel. The base of $1,000 is 7.6% BELOW the street consensus of $1,082.67, whose lowest published constituent of $900 is already 5.9% above the current price.
• Timing — a verdict input, not a footnote: fourth-quarter and full-year fiscal 2026 results are due 2026-08-11, SEVEN DAYS after this dive, on consensus EPS of $2.97 and revenue of $988 million — a 22.2% sequential step on the March quarter's $808.4 million. Third-quarter results were released 2026-05-05 and beat consensus EPS by 4.4% ($2.37 against $2.27), the fifth consecutive beat. The stock rose 8.92% on 2026-08-04 on approximately $5.5 billion of turnover with no company filing dated that day. The most recent insider filings are dated 2026-07-16 and are director equity awards of 108 shares each; there is no open-market transaction and no executive-officer transaction of any kind in the file.
• Accessibility note: no information in this dive is conveyed by colour. All emphasis is carried by bold text, table structure and explicit labelling.
• Not investment advice. Independent research, educational and informational only, never personalised. No recommendation to buy, sell or hold any security is made to any person.
• Version: 2026-08-04-full.