The decisive moment in Dmitry Korolev’s career may have arrived with a cup of coffee and an invoice. In Moscow’s Gorky Park, his team had made a drone deliver coffee. The machine flew. The pilot worked. Then the cafe owner asked the question that turns an engineering demonstration into a business: how much would a second location cost?
The calculation was fatal. Korolev later admitted that the team had not really thought about the product, the business or the economics. He called the result beautiful and useful to nobody. It is an unusually clean confession from a founder, and an excellent beginning for the company he would build next. The failed delivery did not persuade him to leave drones alone. It taught him to move the value somewhere else.
TraceAir would keep the aircraft and discard the spectacle. The drone became a sensor. The product became a shared, measurable picture of a construction site. In Korolev’s telling, the software did everything after reality had been copied: it compared what was there with what engineers had planned, calculated quantities and gave the people moving dirt a view they could use without waiting for a specialist to translate it.
First, make the race visible
Long before construction, Korolev was trying to make action in the air legible to people on the ground. He studied physics at Lomonosov Moscow State University from 2003 to 2008, wrote software and became a serious competition paraglider. AirTribune records him winning the overall classification at a 2012 pre-world championship event and racing internationally the following year. A professional profile describes an eighth-place finish in a 2012 World Cup.
Paragliding gave him a product problem disguised as a sport. Pilots understood the race because they were inside it. Spectators saw colored wings in a large sky and very little else. Korolev and his collaborators built AirTribune, a platform that combined trackers, live maps, competition pages and results. Put a tracker on every pilot, and a private aerial contest became a story the ground could follow.
The pattern would return. AirTribune translated motion into a map. TraceAir translated earthwork into a shared visual record. Both businesses began with an aerial system that outsiders struggled to read. Korolev’s durable instinct was not merely to fly. It was to reveal the hidden race.
“The drones are sensors that copy reality, and the software does everything else.”Dmitry Korolev
The aviation club gets a job
Korolev met Maria Khokhlova in a physics laboratory at Moscow State University. Khokhlova assembled a team of developers and designers for custom software work; Korolev borrowed their skills for projects where aviation met code. During the drone boom of 2014, the experiments grew more ambitious. Automatic landing was fun enough that future TraceAir co-founder Alexander Solovyev jokingly described the group as an aeromodelling club. Then deadlines arrived and, as Solovyev put it, the club was over.
The coffee project used up money but clarified the choice ahead. In early 2015, Korolev and Khokhlova went to Silicon Valley to reboot. They met Nikita Ushakov, freshly out of Columbia University, the day before leaving; he joined the trip. For three weeks, the group tried to understand what drones could become. Delivery had weak economics. Information for an industry had a chance. Mining and construction made the shortlist.
The useful detour
Construction won because the founders could get close enough to learn. An innovation executive at Morton, then a large Russian developer, offered them a place to test. Korolev returned from California and went to sites every day. He watched, flew drones and talked with foremen. The industry’s famous inefficiency stopped being an abstract statistic. It became a sequence of people who could not get the same information at the same moment.
On one early project, two surveyors walked the same route and produced models with different earth volumes. The client had hired TraceAir for an independent view and asked whether the startup could calculate the quantity. The founders’ answer was gloriously early-stage: yes, they would learn now.
A map people can argue over
The founders concluded that delays were not only failures of machinery or planning. They were failures of common evidence. Designers, developers, contractors and general contractors communicated through chains that accumulated errors. Advanced tools existed for engineers, but most people on a jobsite were not engineers. A field manager with a question often had to return to the expert, who had the software, the file and the authority to interpret both.
TraceAir’s wager was that a visual model could widen the circle. Regular drone scans could become georeferenced 3D views. Plans could sit on top of current conditions. Teams could inspect progress, calculate cut and fill, track utilities, flag drainage risks and revisit the history of a site. The clever bit was not any one measurement. It was giving several parties a durable object around which to coordinate.
Korolev once overheard a builder survey a site and say that builders change the terrain of the earth. The line stayed with him because it made construction both grand and accountable. The ground begins one way and ends another. It ought to become better. Software, in this conception, does not replace the people shaping it. It helps them see the consequences sooner.
“They really do change the terrain of the earth. The ground was one way, then it became another.”Dmitry Korolev
The product did not cross the border unchanged
TraceAir began in Russia in 2015 and moved part of the founding team to the United States in 2016. The accelerator then known as 500 Startups gave the company an American platform, a small investment and a crash course in sales. It did not make Russian product-market fit portable. Contractors organized work differently. Relationships formed differently. The founders had to rebuild the product and its position for a new market.
Korolev’s account of the transition is refreshingly unvarnished. The team struggled to find its first U.S. customers and lived through cash gaps. They worked construction networks through warm introductions and site visits. In 2018, he described a routine that started at six in the morning with the distributed development team, moved into local customer work, and ended around eight at night.
Even speech became part of product adaptation. A customer told Korolev that following his English demanded too much effort. He did not retreat into the easy explanation of accent or culture. He studied his pauses and learned where an English speaker expected them to fall. It is a tiny anecdote beside a transcontinental company move, but it captures the operating temperament: remove friction where the customer feels it, even when the friction is you.
Two surveyors can produce two answers. A useful product creates a repeatable reference.
The drone attracts attention; the measurement and decision justify the budget.
Expert software matters more when field and office can share what it reveals.
A larger market still demands local language, local relationships and local product discovery.
From a pilot to a platform
The company accumulated proof in stages. An early Morton pilot was reported to have saved the developer 30 million rubles. TraceAir joined 500 Startups, raised $2.5 million in 2018 and focused increasingly on U.S. homebuilding and land development. By May 2024, when it announced a $25 million Series B led by PeakSpan Capital, the company said it served the top 30 homebuilders across more than 40 states and Canada.
In September 2025, TraceAir announced another change: Ivan Lvov became chief executive as the business entered a multi-product phase. The company said its platform was used by 17 of the top 20 U.S. homebuilders and had nearly 300 active North American accounts. Its current leadership page names Lvov as CEO. Korolev’s public role is now most precisely described by the title that cannot be handed over: co-founder.
That arc gives the coffee drone its proper place. It was neither a silly prelude nor secret genius. It was an expensive way to learn that technical possibility and customer value live in different neighborhoods. Korolev kept flying long enough to connect them.
TraceAir’s own internal joke once imagined construction working like a video game: choose a building on a map and watch the progress bar fill. Reality remains less obedient. Dirt swells and compacts. Plans change. Contractors disagree. Weather arrives. The useful ambition is not to pretend those complications disappear. It is to make the changing ground visible enough that fewer of them arrive as surprises.