Back to research
Bottom line
Fiber optic drones (FODs) are small attack drones that trail a hair-thin glass fiber back to the operator instead of flying on a radio link. The cable can't be jammed. That is the whole point: they keep working where electronic warfare, the jamming and signal attacks both sides use, knocks radio-controlled drones out of the sky. Russia and Ukraine now build them into standard practice. The open question is what they cost to keep flying.
For most of 2025, the bottleneck sat on Russia's side: the one factory that turns raw glass into optical fiber, at Saransk, had been bombed by Ukraine and knocked offline, and it supplied the whole country. Through 2026 the bottleneck became something bigger, the global price and supply of the specialized glass fiber these drones need. That shortage hits Ukraine at least as hard as Russia. And the AI boom is now buying up the same fiber, because data centers use it too. Russia, Ukraine, and the West all draw their cable from the same narrow set of Chinese suppliers.
The bigger shift is in the cost. These started as the textbook cheap, expendable weapon: precise, unjammable, a few hundred dollars each, and built to be used up. The fiber changed that. It is the one part that makes them unjammable, it now comes from a single source, and its price swings hard. The sticker price of a drone is still low. Getting one to hit its target has grown expensive, because so many fail on the way. The economics have turned against them, even as their range and use have grown.
A global fiber shortage
The Saransk plant is still down. There is no confirmed restart since Ukraine's strikes in April and May 2025, so it has been offline for more than a year. Ukraine's foreign intelligence service expects no output before the end of 2027, and Russia has pushed back its own deadline to make fiber domestically to 2028. The plant used to turn out about 4 million km of fiber a year. Today it makes almost none.
With its own factory dark, Russia now imports all of its fiber from China, because US and Japanese supply is blocked by sanctions. Prices followed. Ordinary telecom fiber roughly tripled in China over 2025. The specialized fiber these drones use, the same type AI data centers are buying, rose higher and faster. Ukrainian units that paid about $6 to $7 a km in 2025 were quoted $27 to $30 a km from China by May 2026, and as much as $50 a km for Western supply once China tightened. Sellers now demand payment in full, up front (see Exhibit 1).
Exhibit 1. The fiber these drones use has run far ahead of ordinary telecom fiber, the same supply AI data centers are buying.
Russia is trying to build the piece it is missing. A new plant at Saransk, due in March 2027, would make the raw glass blanks that fiber is drawn from, something Russia has never produced at home. It is a roughly 7 billion ruble project. Neither it nor the delayed domestic-fiber deadline closes the gap in 2026 and 2027.
How many drones Russia can actually build is genuinely unclear, and we would rather say so than fake precision. Russian industry claims 50,000 a month, a figure with no trade data behind it. A cautious estimate from the supply side lands far lower, on the order of 20,000 to 30,000 a year. But the demand side points higher: Russian reporting says the country bought close to 60 million km of fiber in 2025, roughly a tenth of world output, while Ukrainian intelligence puts Russian use at 15 to 20 million km, and still other reporting says the 50 to 60 million figure covers Russia and Ukraine together. Those estimates differ by three to four times, wide enough that averaging them would invent a number none of the sources support. We hold the range.
No longer cheap
The cable is now the most expensive part, and it's scarce.
By spring 2026, the spool of fiber for a single 35 km drone cost about $700, more than a Starlink terminal. On Ukraine's state marketplace, the price of a finished fiber drone roughly doubled by May 2026. One of the leading makers had built its name on cheap long-range fiber. It stopped scaling that line once the cable hit $40 a km, and redesigned its main drone so it can switch between fiber, satellite, and radio control. When a builder starts hedging away from the very thing that defined its product, that is the signal. Fiber used to be a free advantage. Now it's a cost the program has to manage.
And a lot of these drones fail before they hit anything. Ukrainian units that listen in on Russian communications say Russia loses roughly 1 in 4 to manufacturing defects and about 1 in 5 to a snapped cable or dead battery (a single source, so treat it as a claim). One Ukrainian maker pulled back every batch built on Chinese fiber in May 2026 after finding flaws inside the glass that inspection had missed. Winter makes it worse: cold turns the fiber brittle and hard to unspool, and one account put weather-related signal losses as high as half. The only failure-rate figure we have, under 1 in 10 cables breaking across 90 flights, comes from a manufacturer describing its own product.
Put those together. The key part costs several times what it did a year ago, comes from one contested supplier, and gets thrown at missions where a quarter to a third of the drones fail before impact. This is the exquisite-versus-attritable pricing trap in miniature: cheap, expendable weapons and expensive, long-lived platforms follow different math, and treating one like the other misleads buyers. Count only the drone's sticker price and it looks cheap. Count what it actually takes to land a hit, including the failures, the winter losses, and the scarce cable, and it starts to look like the pricey systems it was supposed to replace. These drones stay decisive where jamming makes radio control impossible. What no longer holds is the idea that they are cheap.
Where things stand
Range. They fly farther than they used to. Reliable range now runs about 10 to 50 km, up from 5 to 25 km a year ago. Both sides have put 30 to 50 km models into service. Ukraine approved a near-50 km drone in March 2026; Russian models run from 23 to 42 km. Claims of 50 to 65 km are unverified; the credible Russian figure is around 42 km.
Who's ahead. Not equal yet. General Syrskyi, then Ukraine's top commander, said in January 2026 that on fiber drones Ukraine is "only catching up" to Russia. Fiber still makes up no more than about 15% of Ukraine's drones at the front. Even so, the numbers are large: Ukraine took delivery of 374,000 fiber drones under 2025 contracts, and more than 92% of that total again in just the first four months of 2026.
How common. It depends on where you look. In one Russian frontline sector, analysts counted 111 fiber drones out of 560 drones flown per day, about 1 in 5. Ukrainian units reported anywhere from 10 to 50% near Kherson, and up to about 90% in parts of the Kharkiv front.
Where it's spreading. Russia stood up a dedicated drone command, Rubicon, and grew it from about 1,450 people in early 2025 to roughly 5,000 a year later, with its standard units more than tripling in size. The technology is moving well beyond this war. Hezbollah reportedly made fiber drones a main weapon, blamed for 8 of 12 reported Israeli military and contractor deaths after the ceasefire. A US military branch tested several fiber-drone vendors in the field in January 2026. And the same idea has reached the water, on both Russian and Western sea drones.
Key judgments, indicators, and gaps
We assess, with moderate confidence, that the bottleneck has moved from one Russian factory to a worldwide fiber shortage, and that the shortage squeezes Ukrainian and Western programs as much as Russian ones. Worth watching: the price of drone fiber, how hard AI data centers pull on the same supply, whether Saransk restarts, and whether the 2027 plant stays on schedule. What we still can't see: trade data doesn't reveal how much imported fiber goes to the military, so every supply estimate rests on a guess about that split.
We assess, with moderate confidence, that once you count the cost of each successful strike, the economics got worse through 2026 even as range and use rose, and that makers are hedging away from fiber rather than betting more on it. Worth watching: the price of a fiber spool against the cost of a satellite or radio link, how many drones adopt switchable control, and any failure-rate figure that doesn't come from the manufacturer. What we still can't see: the numbers that would settle it, how many drones pass quality control and how often cables really break in the field, mostly aren't published. Marketing claims (50,000 a month, 65 km range, under $700) have consistently run ahead of what holds up in combat.
We judge, with high confidence, that fiber drones are here to stay wherever jamming blocks radio control, that they'll keep spreading to more armies and armed groups, and that the effort to counter them has grown from improvised fixes into formal doctrine and dedicated units. This is not guaranteed forever. It weakens if drones get smart enough to hit targets on their own, without any link back to a pilot, or if fiber stays so costly and scarce that units fall back on jam-resistant radios. The practical question for Western militaries is supply: can they get the fiber to fly these drones from somewhere other than China, at a price that still makes sense once you account for how many are lost?
Sources and confidence. Built from public reporting through August 2026. Russian state media, Russian manufacturers, and pro-Russian open-source trackers are treated as claims, not confirmed fact. Some Ukrainian intelligence and manufacturer figures come from a single source. Where numbers conflict, such as Russian fiber demand at about 60 million km versus 15 to 20 million km a year, we keep all of them rather than split the difference. The fast-moving figures are dated: fiber prices as of May 2026, Saransk offline as of spring 2026, Rubicon size as of spring 2026.
