LightPath's Defense Pivot Is Real. The Stock's Running From It Anyway.

# LightPath's Defense Pivot Is Real. The Stock's Running From It Anyway.

Let's get the obvious thing out of the way: LightPath Technologies (NASDAQ: LPTH) has been a stupid-money stock for most of 2026. Down about 10% year-to-date against a market that's up into the teens, down roughly 49% off its 52-week high of $18.94, and sitting at $9.72 as of this week. Yet over the past year it's somehow up roughly 300%. That's not a trend; that's a scar tissue situation.

Here's what I actually find interesting about this one, and it's not the stock price.

The business is doing the thing they said it would do, and it's working.

LightPath rebranded itself as a vertically integrated infrared optics and imaging play, and the fiscal 2026 results—reported September 10—back the story. Revenue nearly doubled, up about 93% to $71.7 million, with Q4 alone hitting a record $21.2M, up 73%. Gross margin expanded to 36% from 27%, because higher-value assemblies, modules, and cameras climbed to 44% of sales. Backlog ballooned roughly 197% to a record $110.9M, with $85.6M slated for delivery in the next 12 months.

That's not a company guessing at a pivot. That's a company executing one.

The counter-UAS money is the whole point.

In July, LightPath landed an $11M follow-on order for infrared cameras for counter-UAS work from a global technology customer. Then, in the weeks after the fiscal year closed, it booked another ~$24M in counter-UAS orders, some already moving to monthly delivery cadences. And it brought in Scott Payette, a former Anduril production leader, as VP and GM of G5 Infrared. That's a signal you don't fake: someone who built cameras at the company everyone's watching in defense wants to run this piece.

The geopolitics line is real, even if the timeline is shaky.

LightPath won a funded Department of War prototype agreement under the CLEAR program to develop its BlackDiamond germanium-alternative glass. Germanium is a China-controlled critical mineral, and there are germanium sourcing restrictions coming in December 2027. LightPath already exited China, selling its subsidiary LightPath Optical Instrumentation (LPOIZ) for $4.5M payable over five years. That's a coherent thesis: supply chain sovereignty is a defensible moat right now.

But let me be honest about the cracks, because a bull post without the faults is just a press release.

The skeptics aren't crazy.

The company isn't profitable. Full-year net losses sat around -$20.55M, operating expenses more than doubled to $45.5M from $22M—including a $15.6M noncash charge tied to G5 Infrared earnout accounting. Short interest is a fat 15% to 23% of float depending on who you read. Hedge fund ownership fell from 27 funds to 20. And insiders—via one fund, North Run Strategic Opportunities—have been dumping.

There are also real schedule problems. The Army's Next Generation Short Range Interceptor timeline slipped by months, the redesign of G5's cooled cameras onto BlackDiamond glass is behind schedule, and detector lead times stretched from about six months to 10-plus. Converting that record backlog into shipped revenue without the margins evaporating is the entire game here.

So where does that leave the tape?

LightPath is trading at roughly 9.5x sales against a sector average around 7.5x. The analyst targets cluster in the $14 to $15.80 range, implying maybe 45% to 62% upside—Piper Sandler initiated at Overweight with a $15 target. That upside is real, but it's not free: you're paying a premium for a company that still burns cash and is betting its whole valuation on defense programs that have a documented habit of slipping.

My read: LightPath is a story that's actually playing out, not a story that's just being told. The backlog, the margin expansion, the counter-UAS bookings, the Anduril hire, the China exit—all of it is the real deal. But the market is punishing it because it's still unprofitable, still dilutive, still dependent on a handful of defense programs that keep moving their dates.

If you own it, you're betting that a clean balance sheet ($93.2M cash, basically no debt) and a record backlog outlast the execution risk and the valuation. That's a defensible bet. It's not a safe one.

That's where the backlog story meets reality. Until then, a stock that's up 300% in a year but down 10% this one is telling you exactly what it wants: a conviction player who believes the pivot, or nobody at all. There's not much of a third option.