Thermal cameras see heat, and their lenses have long been made of germanium, a crystal mostly supplied from China. LightPath makes a substitute glass and turns it into lenses and cameras.
The company · 2 · What exactly does it sell?
Assemblies and modules are the biggest part
Each $1 of sales, fiscal 2026
44%30%22%4%
Lens assemblies and modulesHeat-camera lensesOrdinary lensesEngineering work
It is paid per part, and the parts are getting bigger: a lens sells for dollars, a camera system for thousands.
The company · 3 · How does it make money?
Bigger parts pay more, but it still loses money
41.4%kept per $1 after making it
-$2.2Moperating loss last quarter
$132.9Morders not yet delivered, a record
Because glass can be pressed into a finished lens in one step, it costs less than grinding crystal for hours. Selling whole cameras lifts what each order is worth.
Is it a good business? · 4 · What gives it an edge?
A Navy licence, and it is narrow
2021
Naval Research Laboratory licence, to it alone
SPEIR, NGSRI
Navy and Army camera programmes its glass is qualified into
2
other companies make its standard grades
The licence covers only the newest glass recipes; Umicore has pressed this kind of glass in volume for years.
Is it a good business? · 5 · What is the catch?
Umicore sells the same glass, and germanium is back
UmicoreBelgian, same glass, same claims
Syntec Opticstargets the same military buyers
27 November 2026China's export ban suspended until
Listed rivals: UMI.BR · OPTX · TDY
Resumed germanium sales ease the shortage that made the switch urgent.
Is it a good business? · 6 · What does its record show?
More kept per part, not lower costs
Sales, $M
35.6
32.9
31.7
37.2
71.7
Jun 2022Jun 2023Jun 2024Jun 2025Jun 2026
Per part27.2% → 36.0%kept per $1 after making it, for the year
Own growth3.6% → 38.7%growth without bought companies
Overhead34 centsoverhead per $1, above fiscal 2021's 31
Bought companies doubled sales, but office and selling costs fell back only to where they were five years ago.
Is it a good business? · 7 · How much room is there to grow?
Only the border towers can be counted
About $1Bplanned on US border camera towers
About 830 → 2,300towers by 2034
$20M to $40Mof camera content a year
Nobody has signed for the towers: in May management said the funding had not been released.
Is it a good business? · 8 · Who is in charge?
Growth bought with deals paid in shares
Since 2020Shmuel Rubin in charge
Since 2021losses every year
24.8%held by insiders in October 2025
Nearly all that insider stake was one fund, which has since sold most of it.
The story right now · 9 · Why does the military need new glass?
A US law bans Chinese germanium in military gear
15 March 2027Pentagon glass report due
18 December 2027germanium banned in US military gear
Chinasupplies most germanium
US law requires military equipment to stop using germanium from China. Camera makers need another lens material, and LightPath's glass is one of the few.
The story right now · 10 · How fast is it really growing?
Over half the growth was bought
Buy companiesPaid partly in shares.
Sales double102% in the release.
Strip out the purchases32.7% for nine months.
Each share owns lessMore shares were issued.
Growth from its own business is real but far below the headline. Paying for companies in shares means each share owns less.
The story right now · 11 · What changed lately?
Sales rose 38% and orders hit a record
$20.8Mquarterly sales
41.4%kept per $1, from 29.8%
$132.9Morders on hand, a record
-$2.2Moperating loss
Strong quarter
It answered our test: profit per part held for a second quarter, and no purchase was paid in shares. The shares still fell about a fifth.
The story right now · 12 · What is everyone arguing about?
Does one licence protect the growth?
Many investors think
A pure play on bringing optics home
We think
Over half the growth was bought; one licence protects it
Against us: profit per part has risen, and the company's own cash-profit measure has been positive four quarters running.
The price and our view · 13 · What is a share worth?
No single value here: it is priced as a bet
1.3×if it works
24%kept if it fails
Multiples of the company's value when the bet was last struck, 9 Nov 2026.
We put no single value on a share here, because the business does not yet earn a lasting profit. Instead we ask what the shares pay if the plan works, and what is left if it fails.
The price and our view · 14 · How could we be wrong?
New glass at size, and more shares coming
Glass nobody has made at sizeBiggest
The bought plant must melt larger glass first.
More shares to comeShowing now
About 3.0 million still due from preferred.
No forecasts, erratic resultsWatch
Four of six reports missed forecasts badly.
How firm the orders areUnknown
The two largest disclose no units.
About $110.8M of orders is due within a year, and late suppliers already held sales back.
The price and our view · 15 · How will we know?
February shows whether profit per part and orders hold
38% or more kept per $1, a third quarterGood sign
Orders on hand of $120M or moreGood sign
Under 36% kept per $1, or orders under $105MWe turn negative
FEB 202715
Second-quarter fiscal 2027 results, 15 February. Companies report every three months; each report is a test of the story.
Our view moves on these business events, not on a price line, since we put no single value on a share here.
The price and our view · 16 · So, what do we think?
A US law pushes the military to its glass
Priced for successthe kind of bet this is
NeutralMedium conviction
LightPath melts a glass that replaces Chinese germanium in military heat cameras, and orders reached a record. But over half the growth was bought, it still loses money, and the price already pays for success. Neutral.
↑ ↓ to move · C for contents
What it does
More about the company
Who it is
Thermal cameras see heat, and their lenses have long been made of germanium, a crystal mostly supplied from China. LightPath makes a substitute, a glass it melts itself in Florida and Texas, and turns it into lenses, lens assemblies and, increasingly, whole cameras.
Because it is glass rather than crystal, it can be pressed into a finished lens in one step instead of being ground for hours, and one pressed lens can do the work of several.
How it makes money
It is paid per part, and the parts are getting bigger: a lens sells for single dollars, an assembly for hundreds, a camera system for tens of thousands and sometimes hundreds of thousands. No customer is large: the biggest was 9% of fiscal 2025 revenue. What brings them back is a qualified place inside a named military programme.
Growth and profit
What growth costs
Growing has not yet made LightPath cheaper to run. Office and selling costs took a bigger share of each sales dollar for four years while sales went nowhere, then fell back when bought companies doubled sales, but only to about where they were five years ago. Most of last year's improvement came from earning more on each part made, not from spreading costs.
Its record
The numbers
Is it growing?Fast
Sales grew 93% in the year to Jun 2026, faster than the 17% in the year before.
Is it profitable?Losing
Operating margin is −27%, down from −11% in the year to Jun 2023.
Does growth pay off?Unclear
Running costs take 63¢ of each sales dollar, up from 41¢ in the year to Jun 2022.
Does it earn good returns?Poor
Shareholders' money lost 25% a year, down from −11% in the year to Jun 2022.
How long can it keep going?Long runway
It used $16.5M of cash in the latest year, so the $93.2M it holds lasts about 5.7 years at that pace.
Is the balance sheet safe?Net cash
It holds $93.2M of cash against $187K of debt, $93.0M more cash than it owes.
Are shares being diluted?Diluting
The share count rose 98% in 4.0 years, and pay in shares costs 4.0% of sales.
Why it wins
The evidence
The protection is a licence, and it is narrow. The US Naval Research Laboratory licensed its newest glass recipes to LightPath alone in 2021, and that glass now sits in named programmes, including a Navy ship camera through L3Harris and an Army weapon sight through Lockheed Martin.
The licence covers those recipes and nothing else. LightPath's own filing says two other companies make its standard grades, and Umicore has pressed this kind of glass in volume for years.
Who could beat it
The rivals in full
Umicore, a Belgian company outside the banned nations, sells the same kind of glass and makes the same claims for it. Syntec Optics, an American rival, targets the same military buyers. And germanium itself is back on sale: China suspended its blanket export ban to the United States until 27 November 2026, easing the squeeze that made the switch urgent.
Each rival
Umicore N.VUMI.BR
The integrator buys GASIR chalcogenide optics from Umicore instead. Umicore makes its own germanium and its own chalcogenide, moulds asphero-diffractive lenses in medium to large volume, and says GASIR beats germanium on transmission range and athermalisation - the same sentence LightPath writes about BlackDiamond, from a Belgian company that is not in a covered nation.
Syntec Optics HoldingsOPTX
An American defence customer buys moulded glass optics from Rochester rather than Orlando. Its own 10-K names the same germanium constraint and the same night-vision and infrared-camera buyers, and it is the only other listed pure-play telling this story.
Germanium itself, re-suppliednot listed
Nothing gets substituted at all. China suspended the blanket export ban to the United States on 9 November 2025 until 27 November 2026, so a commercial buyer can buy germanium again, and the price pressure that made BlackDiamond urgent eases.
Teledyne FLIRTDY
The buyer takes a complete cooled or uncooled camera from a vertically integrated supplier that already earns a 19.4% operating margin, rather than one LightPath is still learning to build. LightPath names Teledyne FLIR twice in its own competition section, as a standard-product rival and as a prime that is both competitor and customer.
How big it could get
The market and who has signed
The market
Only one of LightPath's markets has a published count: American border surveillance towers. Homeland Security plans to go from about 830 camera-carrying towers to 2,300 by 2034, spending about $1B. At one camera a tower and the company's own price band, that is roughly $150M to $300M of camera content, about $20M to $40M a year. LightPath's orders on hand equal about a ninth of the programme's spend.
Who has signed
The orders are real but mostly unnamed. Motorola Solutions, its largest customer, placed camera orders of $18.2M, $22.1M and $11.0M. A counter-drone supplier ordered $13.0M of assemblies, and L3Harris $2.2M of Navy development cameras. Nobody has signed for the border towers: in May management said the funding had not been released.
Who runs it
Leaders, pay and the cash
Who runs it
Shmuel Rubin has run LightPath since 2020 and turned a lens maker into a maker of camera systems, mostly by buying companies; it has lost money every year since 2021. Albert Miranda has run the finances since 2021 and paid for the deals by selling shares. Insiders owned about a quarter in late 2025, nearly all one fund that has since sold most of it. The yearly bonus rides mostly on operating profit and paid nothing for two years; the chief executive did get a bonus for closing a deal.
What they did with the money
It bought its growth partly with shares. G5 Infrared cost $20.3M in cash plus shares and an earnout of up to $23.0M; Amorphous cost $7.0M plus shares. In June the company raised $50M selling stock at $14.00, and a director's fund sold the same number of shares at the same price, having converted them at $2.15.
The last report
The latest report
LightPath's quarterly revenue rose 38% to $20.8M, with a record $132.9M of orders
What it did to the thesis Strong quarter
It answered the test we set: the margin held above our bar for a second quarter and is guided to stay there, orders hit a record and no purchase was paid in shares. Late suppliers held sales back. The view moved to Neutral and stays there.
The report
In the quarter to 30 September 2026 revenue rose 38% to $20.8M; without Amorphous, bought in January, it rose about 30%. Cameras and optical assemblies were half of sales, and gross margin reached 41.4% against 29.8% a year earlier.
Orders on hand rose 20% in three months to a record $132.9M, $110.8M of it due within a year. It still lost $2.2M after operating costs, and cash fell by $10.3M, mostly into materials bought ahead of deliveries.
On the call the finance chief said margins should stay above 40% while cameras and assemblies are over half of sales. Late suppliers kept sales slightly below the company's plan.
What's happening now
The background
The shares closed at $6.16 on 16 November, down by more than a third since late September with no company news, while the September quarter brought the margin and orders the case was waiting for. The law's dates are fixed: an interim report on the Pentagon's glass strategy is due by 15 March 2027, and germanium becomes a banned material in American military equipment on 18 December 2027.
Where we disagree
Both sides in full
The market
A pure play on bringing optics home, whose revenue doubled. Ten analysts say buy and none hold or sell.
We believe
The law is real. But more than half the growth was bought, and the protection is one licence wide.
The argument
Nobody quotes the table. The quarterly filing gave like-for-like growth of 32.7% for nine months on the same day the results release led with 102%.
The case against us is fair: like-for-like growth rose from 3.6% to 38.7% for the year, gross margin from 27.2% to 36.0%, and the company's adjusted cash-profit measure has been positive four quarters running.
What it's worth
How we value it
No fair value: the September 2026 quarter still ran an operating loss of $2.17M after stock pay, and nine years hold almost no operating profit, so a value built on earnings would rest on a guess; the bet prices it instead.
The bet
Even if this works, the shares look expensive. The best case we can defend is one analyst's sales forecast for 2028, close to double today's sales. Valued like the cheaper of its two listed rivals, that is worth about what the company costs today. Management's much larger goal is five years away, and most of it is unsigned.
If it fails, shareholders keep the cash raised this year and an old lens business that barely grew for nine years, worth a fraction of today's price.
The bet
Seen as a bet
Even if this works, the shares look expensive. The best case we can defend is one analyst's sales forecast for 2028, close to double today's sales. Valued like the cheaper of its two listed rivals, that is worth about what the company costs today. Management's much larger goal is five years away, and most of it is unsigned.
If it fails, shareholders keep the cash raised this year and an old lens business that barely grew for nine years, worth a fraction of today's price.
What could go wrong
Every risk
The risks
Glass nobody has made at size
The larger camera redesigns wait on the acquired plant melting glass that big for the first time; management compares it to its own three-year ramp.
More shares to come
About 3.0 million more shares are still due from preferred stock convertible at $2.15, after 8.7 million were issued that way last year.
No guidance, erratic results
The company gives no guidance, and in four of its last six reports earnings per share missed forecasts by 100% or more.
What we still do not know
How firm the two largest orders are. They disclose no units, and their firmness rests on the company's own description.
How fast $132.9M of orders becomes revenue. About $110.8M is due within a year, and late suppliers already held the September quarter below plan.
What changes our view
Every route to a new view
Every route to a new view
→ Attractive
Gross margin stays at 38% or better through the June 2027 quarter, with like-for-like growth above 15% and orders on hand above $110M. The business, net of cash, is valued at about 5 times a year's sales or less; it reached about 4.5 on 16 November.
→ Unattractive
Gross margin falls below 36%, or orders on hand below $105M, at a quarterly report. Another acquisition is announced, part paid in shares, before today's orders have been delivered. The March 2027 glass-strategy report is missed, the germanium ban is put off, or a rival joins a named US military programme.
The next test
It supports the case if
Gross margin at or above 38%, revenue above the September quarter's $20.8M, backlog at or above $120M, and no new acquisition paid in stock.
It pressures the case if
Gross margin below 36%, backlog below $105M, revenue below $20.8M, or another acquisition part-paid in stock.
We would also be wrong if
Like-for-like growth falls below 15% for two quarters after Amorphous laps in January.
The verdict
Scores and how we weighed it
The scores
Growth
3 of 4
Business quality
3 of 4
Financial position
3 of 4
Execution record
3 of 4
Per share capture
2 of 4
Potential
2 of 4
Risk profile
3 of 4
Valuation
2 of 4
Evidence quality
4 of 4
How we weighed it
ValueMixed
No fair value is drawn; the price now asks much less of the future, though it still assumes profits never yet earned.
BusinessMixed
A real policy tailwind and a margin now above the level the case needed, with growth partly bought.
GrowthPositive
Like-for-like growth picked up, orders reached a record and most of them are due within a year.
RisksNegative
The glass is unproven at size, suppliers ran late and held sales below plan, cash went into stock ahead of deliveries, more shares are due and there is no guidance.
RivalsMixed
Between Teledyne and Syntec Optics on sales, so the group settles little.
The groupMixed
Optical names move together on AI spending news; it ranks last in its group.
Latest newsPositive
The September quarter beat the margin and order tests the thesis set; on the call the finance chief expects the margin to hold at its new level, but supply delays held sales below plan.
In plain words
Selling new shares
A company can raise money by creating new shares and selling them. The company gets cash, but the same business is now divided into more slices, so each existing share owns a little less of it. This is called dilution.