LPTH: Backlog Tripled. Shareholders Funded the Cash.

The 60-second read

Five things to know

1. Demand has substance. The backlog is approximately three times the previous year-end level, with most scheduled for delivery within twelve months. It still depends on customer schedules and successful manufacturing.
2. The acquired perimeter matters. FY2026 includes twelve months of G5 Infrared and about five months of AML. Reported revenue growth is not a same-business organic growth rate.
3. Adjusted profit does not equal cash. The company excludes earnout remeasurement and stock compensation. Earnouts can reflect business success while still transferring value to acquisition sellers.
4. Use the new share base. June-end common shares exceed the share count implied by the provider market-cap snapshot. Our forward models use 81 M shares, clearly identified as an assumption.
5. No numerical company guide was verified. We separate management’s qualitative outlook, stale/thin Street estimates and our own assumptions. We do not manufacture a beat or miss.

Three scenarios, one probability-weighted number

Illustrative 12-month equity values; probabilities are Charged Alpha judgments, not statistical forecasts.

Signposts: what would change our mind

These are our forward tests, not management promises. This is our first LPTH packet, so there is no prior Charged Alpha signpost score to claim. Unconfirmed reporting dates are review deadlines rather than scheduled company events.

The tape

The reference price is $9.68 from the saved September 10 provider quote. That is a valuation input, not a verified causal reaction to this earnings release. The quote lists a 4.54% daily decline, but the relationship between the filing acceptance field, publication time and trading session has not been independently reconciled. We therefore do not describe that move as the market’s response to these results. This matters because a compelling story can otherwise acquire a false opening statistic before the financial work begins.

The capitalization check is more important than a one-day price move. Multiplying the reported June-end common shares by the reference price produces approximately $677.24M of common market value, already above the provider’s $607.80M field. The latter embeds a smaller share base. Neither figure is a fully diluted current capitalization. Preferred conversion rights, awards, warrants and stock earnouts create further claims. Our forward models use 81 M shares and retain the underlying components so the assumption can be challenged. A rising share price does not repair a stale denominator; only an updated capitalization bridge does.

The print

The authoritative current source is LightPath’s September 10 Item 2.02 earnings release for the quarter and year ended June 30, 2026. The detailed statement columns are, in order, Q4 FY2026, Q4 FY2025, FY2026 and FY2025, expressed in dollars. Exact Q4 revenue is $21.162352M versus $12.209793M. The calculated increase is 73.32%, while the rounded summary prints 73.8%; we use the detailed values when calculating growth. Revenue rose about 10.51% from the March quarter. This is a strong absolute result, but the composition of the growth determines whether it should command a platform multiple.

FMP’s newest quarterly and annual fundamentals were rejected. The Q4 row carries an annual-scale revenue field inflated to tens of trillions of dollars, and current cash-flow and balance-sheet fields have similar scaling defects. We have not divided those numbers by an arbitrary factor. The raw responses remain saved, while the current quarter and year were transcribed from the actual SEC statement and checked against their subtotals. Historical provider operating-profit fields also omit or misplace earnout marks, so the historical operating-income series uses SEC facts and the latest comparable filings. No earnings beat or miss is asserted because a pre-release consensus with verified basis was not established.

The current release also contains small wording and rounding inconsistencies: one sentence calls positive adjusted EBITDA a loss and another rounds the net loss differently from the statement. These are not reasons to reverse a sign silently or discard the filing. The reconciliation table reports positive $2.095863M adjusted EBITDA and negative $4.140865M net income; those are the values used here. Our numerical record preserves the original columns and the discrepancy notes. The future 10-K remains an important confirmation of the unaudited year-end statements, especially the capitalization and acquisition disclosures.

Products and demand

LightPath sells a combination of proprietary infrared materials, optical components, assemblies, camera modules and engineering. Its economic ambition is to move from supplying an optical part to solving a larger system problem. That can increase content per customer and deepen qualification relationships. The current release attributes margin improvement to product mix, throughput and manufacturing yields rather than pricing alone. The attractive demand driver is replacement of China-linked germanium supply in defense and other sensitive applications. Qualification cycles can provide durability, but they also make the timing of conversion uneven.

FY2026 contains a full year of G5 Infrared, acquired in February 2025, and approximately five months of AML, acquired in January 2026. The year-earlier period contains only part of G5 and no AML. The 92.79% calculated revenue increase therefore cannot be called organic growth. AML contributed $1.1M of infrared-material sales to Q4 according to management, but that is not a complete same-perimeter bridge. Assemblies and modules represented roughly 44% of annual revenue in the rounded product table versus about 23% previously. Product profit is not separately disclosed; allocating the consolidated margin to each line would manufacture precision.

The $110.9M year-end backlog is impressive against $37.4M a year earlier, yet it is only $0.3M above the March quarter’s $110.6M. Both comparisons belong in the analysis. Management schedules $85.6M for delivery within the next twelve months, equivalent to about 77% of the backlog and greater than the latest year’s revenue. That schedule does not turn every order into an unconditional receivable. It depends on customer programs, qualification, yields and capacity. Our base forecast assumes $105M next-year revenue; it is an analytical forecast that also requires new business and successful conversion, not a company-issued numerical guide.

Profit and the seller’s claim

Gross profit rose to $8.35M in Q4 and gross margin reached 39.44%, compared with 22.04% a year earlier. That improvement is material even after allowing for the acquired perimeter. The operating statement then absorbs $8.12M of selling, general and administrative costs, $1.13M of product development, $0.46M of intangible amortization and a $3.40M acquisition-liability remeasurement, partly offset by a $0.46M disposal gain. The result is a $4.30M operating loss. The income statement should not be reduced to either the gross-margin success or the final loss alone.

An earnout remeasurement is a poor measure of this quarter’s factory economics because it can increase when the acquired business performs better. It is also a real claim on shareholder value. The right treatment is to remove the periodic mark when measuring a normalized operating margin, then explicitly include the cash or shares owed to the sellers in valuation. Calling the charge noncash today does not make the settlement free. For Q4, adding back the $3.40M mark and removing the $0.46M disposal gain leaves an approximately $1.36M economic operating loss while retaining stock compensation and depreciation.

The year tells a related but larger story. FY2026 adjusted EBITDA was $4.23M, yet reported net loss was $20.55M. The annual reconciliation excludes $15.64M of acquisition-liability marks and $2.86M of stock compensation, along with smaller debt-extinguishment, acquisition and currency costs. Removing only the earnout mark and disposal gain from reported operating profit still leaves about $4.47M of annual economic operating loss. We retain stock compensation as a cost of employing people. We also retain depreciation in an economic earnings route because an optics manufacturer must keep investing in real equipment.

Earnings quality

Quality improved at the factory gross-margin level but remains mixed at the shareholder level. The same acquisition growth that broadens the product portfolio introduces contingent purchase consideration, amortization and integration costs. The preferred security and equity raises change who owns the future result. A nine-check scorecard makes these issues visible instead of allowing one adjusted measure to decide the verdict. Several tests remain warnings because the company is loss-making, the comparison perimeter changed and no formal numerical earnings guide was verified.

Receivables ended the year at $15.11M and inventory at $18.01M. Using ending balances and a ninety-day quarter gives approximately 64 days of sales in receivables and 126 days of cost of sales in inventory. These are monitoring proxies, not audited turnover statistics: average balances, acquisition timing and program mix can materially change the interpretation. The operating cash-flow statement shows $5.58M consumed by receivables, $5.06M by inventory and $2.24M by prepayments and deposits, partly offset by $3.31M from payables and accrued liabilities. These movements show the financial cost of scaling the order book.

The G5 earnout also illustrates why classification needs attention. FY2026 includes approximately $3.81M of an earnout payment in operating cash flow and $3.54M in financing cash flow. Their combined cash payment is about $7.35M. A free-cash-flow metric based on reported OCF already includes the operating portion; subtracting it again would double count. Conversely, a financing-classified acquisition payment remains an economic outflow even though it is absent from conventional FCF. We show reported FCF, acquisition cash and the seller payments separately rather than rewriting them into a cleaner-looking synthetic cash number.

Cash, funding and dilution

The cash increase is the most easily misread achievement in this print. Cash rose from $4.88M to $93.20M, an $88.33M increase. The cash-flow bridge is negative $10.23M of operations, negative $12.79M of investing, positive $111.02M of financing and positive $0.32M of foreign-exchange effects. Net public equity proceeds were $112.34M and private equity proceeds were $7.89M. Those are funds contributed by investors, not cash earned from selling optics. The stronger balance sheet reduces immediate funding pressure, but its shareholder cost is visible in the share count.

The March 10-Q sets the remaining G5 settlement at $9M: $6.3M cash and $2.7M stock, payable in the first half of January 2027. AML contingent consideration is capped at $3M in stock. Our 81M share denominator is a gross modeling allowance, not GAAP diluted shares or a certified current cap table. It combines June common stock with earlier gross award counts and an explicit preferred proxy. All purchaser warrants from the Series G financing were exercised cashlessly by March 31. The 2.975M performance-option awards are included within the 4.145M option total; their $10.53 strike is above the $9.68 reference, and vesting depends on demanding targets. No exercise cash is credited. RSAs are not additive because restricted issued shares may already be in common. SBC remains an economic expense. Operating leases are not also capitalized because operating margins retain rent.

The actual Series G certificate defines optional conversion using liquidation preference, which includes unpaid accrued dividends, divided by the conversion price. The stated-value floor is 3.019535M shares. The $8.906686M balance-sheet amount reflects redemption accounting; dividing it by $2.15 gives a4.142645M upper modeling proxy, not a proved present conversion amount. Certain fundamental transactions, liquidation or mandatory conversion during the five-year guaranteed term trigger a dividend make-whole; ordinary optional conversion alone is not listed as that event. The March filing reports7,610 preferred converted into 3.539379M common while$10.4M of temporary equity was reclassified, underscoring that book value and issued conversion shares are different measures. Exact remaining accrued dividends, waivers and award movements require the next filing.

Seller consideration also makes dilution price-dependent: $5.7M of potential stock equals 0.589 M shares at $9.68, but1.9M at $3.00. The 81M allowance is not a maximum. We show75/78/81/84M sensitivity and treat potential claims above 84M as adverse evidence. These counts must be reconciled without adding already issued shares twice.

Three routes to value

All three routes share one capitalization bridge: $93.20M cash, less $0.85M of loans and finance leases and $6.30M agreed G5 cash consideration, yielding $86.06M adjusted net cash. They also share 81M assumed common-equivalent shares. The models do not subtract the preferred carrying amount after assuming conversion, and they do not turn a noncash earnout charge into a free acquisition. This consistency is more valuable than a third decimal place in the target. Every operating forecast below is our assumption; management’s release provides qualitative direction rather than a verified numerical earnings guide.

We weight the cash-flow route 25%, the revenue route 25% and the economic-earnings route 50%, producing approximately $5.17; the published base is rounded to the nearest $0.50, giving $5.00. The very wide route spread is information. A revenue multiple can reward future strategic relevance long before the associated equipment, working capital and taxes deliver cash. The DCF is deliberately demanding about those reinvestment costs. The earnings route gives management credit for building scale but requires depreciation and employee compensation to remain real expenses. Our SELL assessment reflects the gap from the $9.68 reference price, not a claim that the company’s technology lacks value.

The model capitalization implies about $698.02M of operating value today. Holding our $210M FY2030 revenue and 25× earnings multiple fixed, matching that value requires an EBIT margin of approximately 29.9%, versus our 16% base. Higher revenue, a larger terminal multiple or a lower discount rate can also close the gap, so the implied-margin calculation is a conditional test rather than a forecast. A bull can reasonably expect greater platform value; the burden is to explain which variable improves and why. With very high uncertainty, we would require roughly a 30% margin of safety to the $5.00 base—about $3.50—plus improving cash evidence before considering a BUY.

Wall Street context

The available Street marks are optimistic, but they are not fresh responses to this release. FMP’s retained news records show Piper Sandler initiating at $15 on August 12, Canaccord at $16.50 on May 11 and Lake Street at $15 on May 8. All predate the September 10 print. A provider consensus snapshot reports Buy and a target around $15.50; the count in a ratings aggregation should not be confused with the number of analysts supplying next-year estimates. We show firm and date so a reader can distinguish a live analytical disagreement from the mere repetition of old targets.

The retained annual estimate snapshot lists approximately $103.81M FY2027 revenue from three contributors and $139.50M FY2028 revenue from one. Its EPS and net-income figures imply a share basis far below the latest reported common count, and the snapshot does not establish a pre-earnings timestamp. We do not treat those EPS values as comparable to our diluted model or use them to declare an earnings surprise. The revenue path is useful as a rough external reference, with the thin contributor count visible. Our $105M/$140M forecasts happen to be close, yet our valuation is lower because the conversion from sales to economic profit, cash and value per share is more demanding.

The proper comparison is therefore not simply our target versus a consensus target. It is the required margin, capital intensity, dilution and duration of the growth runway behind each target. Full broker-model assumptions are unavailable in the retrieved records, so we cannot claim that analysts omitted any particular liability. Our own model makes those claims explicit and makes room for a bull outcome around $10. A future post-print revision with a reconciled share count would improve the comparison. Until then, a target table is context, not a substitute for a cash-flow model.

Management and allocation

Management has delivered a broader product platform, higher revenue, stronger gross margin and a much larger backlog. Those outcomes support the strategy of moving toward integrated optical systems. The next test is whether the company can scale the acquired platform without repeatedly asking common shareholders to finance the next phase. FY2026’s large net equity proceeds were rational in the sense that they replaced funding stress with capacity to invest. They also enlarged the denominator by nearly 63% over the year. A capital-allocation scorecard must hold both observations at once.

The March 10-Q documents the acquisition architecture in useful detail. G5’s initial fair-value consideration was about $27.13M, with cash, equity and contingent consideration; the first cash earnout was subsequently split across operating and financing cash flow. AML’s cash price was $7M, with up to $3M of stock contingent consideration. Acquired backlog and product know-how can be worth paying for, but a rising earnout is an additional price paid for success. The future question is not whether the charge is adjusted away. It is whether the combined business earns enough after the full purchase price and funding dilution.

Five questions frame the next review. How much next-quarter revenue comes from the same owned perimeter rather than new purchases? What share of the scheduled backlog has firm delivery milestones and acceptable expected margins? How much cash capital expenditure is required before cash conversion improves? What is the complete common-equivalent share count after preferred dividends, warrants, awards and both acquisition settlements? Finally, how much China divestment cash is collectible, and what operating profit leaves with that business? The current agreement calls for $4.5M over five years; we do not credit that face amount as immediate cash or count a divestment as organic growth.

Risks and counterarguments

The strongest bull case is that proprietary germanium-free materials and an integrated domestic production footprint become a scarce qualification asset. Long design cycles, defense requirements and existing customer programs could support a higher-value system business than historical lens sales suggest. In that outcome, current development and equipment spending creates future margins rather than simply consuming cash. A small revenue base also gives successful programs room to change the trajectory quickly. Our analysis does not assume that the recent gross-margin improvement must reverse; it asks how much durability and conversion the present price already requires.

The strongest bear case is a sequence in which backlog remains impressive while receivables, inventory and capacity spending absorb the cash. Acquisitions can keep reported revenue rising and make same-business comparisons harder. Equity issuance then funds the next opportunity, while stock compensation and seller consideration absorb more of the upside. None of these risks requires the products to fail. The distinction is between a valuable operating business and a sufficiently valuable claim per common share at today’s price. That is why our economic earnings route retains labor and asset-use costs and why the share denominator is visible.

The analysis itself has limits. The FY2026 year-end release is unaudited, formal numerical company guidance was not verified, product-group profits are unavailable, and the exact current fully diluted cap table is not established. Cash location and the treatment of future preferred dividends can also affect common value. We have not applied a mechanical Beneish or Altman score: acquisition accounting, changing classifications and missing standardized inputs would make a precise-looking score misleading. Those omissions are explicit evidence gaps, not zero values. The wide $3–$10 scenario range and very high uncertainty rating reflect them.

What could change the call

The most useful near-term catalyst is a clearer operating-to-cash bridge in the next filing. Another backlog record would help, but only if it comes with acceptable conversion, receivable collection and manufacturing economics. The next annual report should clarify the June-end preferred balance, contingent consideration, share instruments and acquisition accounting. These are not peripheral footnotes: they determine whether our 81M gross share allowance adequately captures the economic claims and whether the cash bridge belongs to ordinary shareholders. The research will be updated when a verifiable filing changes those facts.

A higher fair value would require evidence that the platform can support either more revenue, a stronger sustainable margin, or less reinvestment than our current base. A lower share count than the model could also help, but it must come from a verified capitalization reconciliation rather than a provider field. We would reconsider the SELL if cash conversion improved enough to move the DCF toward the other routes and the price offered a credible margin of safety. A falling price alone is not evidence that the business is improving; similarly, a rising price is not proof that the contingent claims have disappeared.

Sources and method

The financial base contains nine quarters from Q4 FY2024 through Q4 FY2026 and three complete fiscal-year cash-flow statements for FY2024–FY2026. Historical values use SEC company facts and the corresponding retained 10-Q/10-K and earnings-release tables. Fourth-quarter flow values are derived as the fiscal year minus nine months when necessary; weighted-average shares and EPS are never derived by subtracting annual averages. Historical cash and restricted-cash tag differences are retained in the source record. The current release’s actual column headers and dollar units govern current-period values.

LightPath is a U.S. issuer with Nasdaq-listed Class A common shares and one vote per common share. The FY2025 annual report names BDO USA, P.C. as auditor beginning in 2025 and MSL as the prior auditor. It describes a single consolidated operating segment and reports no variable-interest-entity or off-balance-sheet arrangement. These facts are taken from that annual filing and should be refreshed when the next annual report is available. The preferred security is a separate senior economic claim, even though the operating business is consolidated. There is no ADS conversion applied to LPTH.

The data audit preserves a few distinctions that are easy to lose in a presentation. Revenue growth calculated from detailed dollars can differ from a rounded company summary. A product reclassification does not create economic growth. The latest SEC operating-income series includes earnout charges that some provider history excludes. Preferred accretion and retrospective accounting can affect historical common-share EPS, so we prioritize reconciled net income rather than building a valuation on a superficially comparable EPS series. A blank disclosure is unavailable, never zero. We retain source URLs, hashes, retrieval dates and raw statement rows so each calculation can be revisited.

Our figures label dollar amounts in millions unless they are per-share values. Gross margin is gross profit divided by sales; free cash flow is reported operating cash flow less cash capital expenditure. The working-capital-day proxies use quarter-end balances divided by quarterly revenue or cost of sales, multiplied by ninety. The three valuation routes use the same assumed dilution and common-value bridge. Scenario probabilities, discount rates, terminal growth, margins, capital intensity and valuation multiples are analytical assumptions. This packet is prepared before publication links exist; the final delivery step will add verified video and podcast links without inventing IDs.

Capital-rights source: Series G certificate, February 14,2025, SHA256 13dae3b84b1db990ab3a0337b9155436228fc7c02711061c60d3006ab2ccbf88. Read alongside March 2026 Form10-Q Notes 11–12 and the June 2026 release. The March filing reports approximately $24M unrecognized award compensation, including about $8.4M scheduled for FY2027; performance conditions and future expense recognition matter. Neither the option-count allowance nor our economic margin forecast erases this employee cost.