Buildots Patent Portfolio Analysis

One pulse, a hundred and thirty million

On 14 September 2026 Buildots announced a $130M Series E. Lightspeed Venture Partners, Human Capital, Qumra Capital, O.G. Venture Partners and TLV Partners are named on the round. The company was founded in Israel in 2018. By the standards of a Series E this is a substantial single event, and it arrives into a business that is visibly staffing for scale: of roughly a hundred open roles, forty are go-to-market and twenty-one are software or AI.

That hiring mix is the first clue about what stage this is. Customer success leads for North America and Canada, delivery project management, a go-to-market AI engineer. This is not a company still deciding what it is building. It is a company selling something already built. Which raises the obvious question about the asset being sold, and whether the intellectual property record behind it is as heavy as the cheque.

Three ticks on the arc, one solid

Here is the entire public record. Three patent families. The first was filed on 2 April 2020, the most recent published filing on 14 December 2022. One per year across 2020, 2021 and 2022. The titles are plain: construction project tracking, system and method for assessing imaged object location, interactive GUI for presenting construction information at construction projects. Two offices carry them, the United States and Europe.

One tick is solid. One of the three families is granted, which means it is enforceable now rather than someday. The other two are published and pending. The last of them entered the public record 45 months before the Series E was announced. Anything filed since sits inside the eighteen-month publication lag and simply cannot be seen. What you can see is small, early, and consistent.

Three families is not a portfolio. It is a position.

Two dots, and one of them repeats

Two inventors are named across the whole record. Yakir Sudry appears on three of three. Roy Danon appears on three of three. There is no third name anywhere in the file, no widening pool as the company grew, no later family with a different team attached. Every idea Buildots has put on the public record between April 2020 and December 2022 came from the same pair, and was written down the same way.

That has two readings and both are true at once. Concentration like this usually means the technical thesis was settled early and held, which is why the filings read as one continuous argument rather than three separate bets. It also means the inventive record has no redundancy. There is no second generation of named contributors behind these two, at least not yet in anything published.

A small circle among very large ones

Set Buildots against everyone else filing in the same CPC groups over the last five years and the scale shifts hard. State Grid China: 4,720 families. Zhejiang University: 2,399. Tsinghua University: 1,891. Beihang University: 1,826. Southeast University: 1,808. Tianjin University: 1,332. Buildots: three. On this scale the company is barely a mark on the page.

Volume is not the same as overlap, and none of those organisations is doing what Buildots does. But the arithmetic matters for one reason. In a technical space this densely filed, the value of three families rests entirely on how the claims are drawn. Count cannot carry it. Scope has to. A narrow claim in a crowded class protects very little; a well-placed one can protect a great deal.

Five organisations found their way back

The arrows here run inward. Flyreel cites these filings nine times. Slate six. Dalux Aps six. Doxel three. Topcon two. Twenty-six citations in total from five distinct organisations, pointing back at three families written by two people between 2020 and 2022. That is the only external measure of the record's weight available, and it is the one that does not depend on how Buildots describes itself.

Citations are the closest thing patent data has to an independent vote. They record that other people filing in this territory had to reckon with this work. Twenty-six of them against three families is a high ratio, especially for a record whose newest public entry is from December 2022. The filings are being read. That is what the money bought.

Twenty-six citations against three families is the whole argument.

What the next publication will settle

Everything behind this question is now four years old or more. One granted family, enforceable today. Two pending, both filed before the end of 2022. Two inventors, no third. A citation pattern that says the work landed. And a company hiring forty go-to-market people, which is the signature of a firm monetising rather than exploring.

The gap between December 2022 and September 2026 is the part nobody outside the company can read. It either contains new families working through the publication lag, or it contains nothing. The next eighteen months of publications will say which, and that single fact determines whether the picture above is a foundation or a ceiling.