Buildots

The problem: the gap between the plan and the site

Buildots was founded in 2018 by three graduates of the Talpiot program: Roy Danon (CEO), Yakir Sudry (CTO) and Aviv Leibovici (CPO). It was built around an old problem in the construction industry: at any given moment, it is hard to know how much of the work has really been completed. On a large project, many subcontractors work in parallel, and progress reports are collected by hand and based on estimates. According to Danon, it often turns out only too late that the project is drifting from the plan, and by that stage fixing it is already expensive.

How it works: a camera on the helmet and a vision model

The site manager walks around the site a few times a week with a 360-degree camera on their helmet. In some cases the footage is taken by drone or by laser scanning. The system compares the footage with the building's 3D model (BIM) and the schedule, and computer vision models determine which components have already been installed, from drywall to openings for electrical outlets. According to the company, it classifies hundreds of types of work from the video alone, and the footage becomes a progress report and risk alerts within 24 to 36 hours. Its delay-forecasting tool compares the planned pace of work with the pace actually measured, and flags activities that are likely to miss their finish date.

Customers and scale

According to the company, the system is deployed at more than 100 construction companies and project owners. They include Intel, the data center operator Digital Realty, and contractors such as STO Building Group, JE Dunn, Mortenson, Bouygues and HOCHTIEF, and in Israel Tidhar, Ashtrom, Azrieli and Mivne. The company claims the system cut delays by about half and saved an average of three months on a project's schedule. These are figures the company reports itself, and they are not audited. In 2026 it employed about 400 people, 260 of them in Israel.

The data center wave

The recent acceleration in the company's business is directly tied to the artificial intelligence boom. Data centers are projects where electrical, cooling and structural work sit right next to each other, and a small delay at one contractor rolls onward and pushes back the day the computing power starts to earn money. Danon told Calcalist that the company tripled its revenue in the year before the funding round and expects to grow at the same pace, but he declined to give the amount. According to him, the same problems also exist in building offices, schools and hospitals, but in data centers every deviation from the plan causes much bigger losses. In October 2025, the company acquired Genda, whose technology tracks workers' presence on site using a Bluetooth infrastructure.

The September 2026 round

On 14 September 2026, Buildots announced a $130 million funding round led by O.G. Venture Partners, founded by businessman Eyal Ofer, with participation from Lightspeed, Intel Capital, Qumra, Viola Growth, Poalim Equity and Avigdor Willenz. According to the announcement, the total the company has raised since it was founded stands at $297 million. Calcalist reported that the valuation in the round is estimated at about $1 billion, compared with about $300 million in the previous round, of $45 million, about a year and a half earlier. It also reported that the whole round was primary capital, with no sale of existing shares. The company itself did not publish a valuation.

What the story teaches

Buildots is an example of artificial intelligence that leaves the screen for the physical world, and therefore depends on data that cannot be downloaded from the internet. The company stresses that its models rely on eight years of footage from real construction sites, not on information collected online. This advantage is also the limitation: the success measures that are published, such as fewer delays and growth rate, all come from the company, and only time will show whether they hold across hundreds of projects at once.