The White Space Is What Nobody Built

I typed four words into a patent database and it answered with 125,048 results. Drone, UAV, unmanned aerial vehicle. The defence subclass alone holds more than 100,000. In 2023 the world filed 19,700 drone patents. China filed 17,285 of them, 87 per cent. The United States filed 858. Russia filed 333. Then I searched for the one system everyone agreed was working. Fiber optic drone control, immune to radio jamming, first deployed in August 2024. Zero patents. I searched the Iranian programmes in English and in Farsi. Zero. I searched SYPAQ, whose cardboard airframe flies resupply missions into places nothing else reaches. No filings found. The 125,048 was the noise. The zero was the finding. A record of claims is not a list of what an industry built. It is a list of what an industry believed was worth defending. Most people read it as an inventory. It reads better as a confession, and it reads best where the page is blank. The probability that the most valuable position in a field sits unclaimed is not zero. Most competitors are pricing it at zero, and they price it at zero because the empty space looks like a mistake.

Nobody spends EUR 15,000 in legal fees and three years of examination on a position they think is worthless. Every patent is a bet written down in public. Read the connectivity column and the belief is obvious. Huawei holds 10,513 patents in that class. Qualcomm holds 7,896. Samsung holds 7,585. Read the consumer column: DJI holds 4,628. Counter-drone work, the defence of the sky against small aircraft, holds more than 25,000 patents, and the same names keep appearing: Anduril, Elta, IAI, Boeing. Filings across the field grew 16 per cent a year, from 16,800 in 2022 to 19,700 in 2023. I have built companies across twelve countries, and the pattern held in every one. Money goes where the moat can be described to a lawyer. Where the moat is a factory method, a supplier relationship or a week of improvisation, nobody files. So the record says two things at once. It shows where a crowd has gathered. It also shows, by its absences, which parts of the field everyone quietly agreed not to defend.

The empty spaces are not scattered. They arrive in three clumps. First, the cheap end. Low-cost and expendable airframes sit at the bottom of the table, and the paper and cardboard programmes sit at near zero. Second, the sanctioned end. Russian and Iranian combat aircraft are run as trade secrets, which is a supply-chain decision dressed up as a strategy: sanctions prevent international filing, and a copied airframe cannot be patented by the party that copied it. Third, the adaptive end. Field improvisation invented on a Tuesday and used by Friday is the hardest thing on earth to claim, because the person inventing it is busy. Ukraine filed four drone patents across 2022 and 2023. Four. In a war that rebuilt small-aircraft warfare in eighteen months. That is the shape of the thing. Gaps cluster around cost, around legal exposure and around speed. They do not cluster around importance. Nothing in a patent record measures importance.

ZALA Aero, the Kalashnikov subsidiary behind the Lancet, holds no drone patents. Kronshtadt, which builds Orion, holds none. The Orlan-10, one of the most-produced reconnaissance aircraft of the war, holds none. The Shahed-136, a delta-wing airframe built in the thousands, is unpatented. The argument for secrecy is that a competitor cannot copy what it cannot read. The object itself does not care. In 2026 the United States fielded LUCAS, a reverse-engineered Shahed built at about USD 35,000 a unit, with no licence to negotiate and no claim to answer. I built Entrade up to 200 wood-gasification machines across the UK and Europe. The advantage was never a document. It was the tolerance on a weld, the supplier who answered on a Sunday, and the fact that nobody else wanted to touch a machine that dirty. A trade secret survives exactly as long as the gap between what you know and what a competitor can see. That gap closes the day your product works in public. Secrecy is not protection. It is a delay, with no stated date.

The 125,048 was the noise. The zero was the finding.

Gaps get chosen by arithmetic long before they get chosen by ambition. A guided interceptor carries a price near USD 1 million. The small airframe it is aimed at costs about USD 3,000. That ratio does not improve with better engineering, and everyone in the market can read it off a spreadsheet. The published programmes at the cheap end: SYPAQ's Corvo PPDS in Australia at USD 1,000 to 5,000 a unit, Air Kamuy 150 in Japan at USD 2,000 to 3,000, PapyDrone-800 in South Korea at roughly USD 1,000, Skykrafts ahaDRONE in India at about USD 500. None of them holds a published patent. The only filing anywhere near the construction method is US10589858B2 from Imperial College, and it covers robotic foam deposition, a different technique entirely. A corrugated board airframe is die-cut, folded and flown. The material costs less than the shipping. That is why the money stayed away and the gap stayed open. It is also the only reason it will not stay open forever.

Some territory is fenced and signed. Counter-drone interception carries more than 25,000 patents. Traditional airframe design is DJI country, with 1,000 patents and the legal department to enforce them. Walking into either is a decision to spend a decade in someone else's shadow. The empty side of the map reads differently. Fiber optic drone control: patents found, none. Maritime air defence fired from an unmanned surface vessel: none. Machine learning that retrains on the frontline in days rather than weeks: none. Detection of a tethered drone, the counter to the counter: none. Nobody avoided those spaces by accident. Fiber optic control removes the radio link the incumbent defence industry sells everything around. A cheap expendable aircraft undercuts a product line carrying thirty years of margin. A model that updates itself in the field removes the cloud contract that pays for the rest. The gap is not where the market is heading. The gap is where the market's current revenue tells it not to look. That is why it is still there. It is the only reason it will not be there for ever.

None of this is really about aircraft. Pick the field you operate in and find the record of what has been claimed inside it. Count where the claims cluster. Then look at the space beside the cluster and ask the only question that matters. Is this empty because it is worthless, or because it is uncomfortable? In my own trade the answer is usually uncomfortable. I signed a EUR 20 million personal guarantee and the company folded on Christmas Day 2008. Capability was never the constraint. The thing I refused to look at was the thing that finished it. So run the search. Filing counts, unit costs, dated programme lists, the four-year-old survey nobody has updated. The information is public and almost nobody reads it, because reading a blank page means admitting you do not know why it is blank. The probability that your field's largest unclaimed position is invisible to you is not zero. Most operators are pricing it at zero. The white space is not empty. It is where the competition refused to go, and the reason is usually printed in their own filings.