The $50,000 Phone Call: How OpenSpace’s Jeevan Kalanithi Built a $902M Visual Intelligence Platform by Listening to Builders
OpenSpace CEO Jeevan Kalanithi turned a simple insight, that a picture is worth a thousand words on a jobsite, into a $902M visual intelligence platform now valued at the heart of ConTech’s spatial AI race. His playbook is unglamorous: live on jobsites, charge customers from day one, and build technology that removes work rather than adds it. For builders chasing margin and ConTech founders chasing PMF, his story is a masterclass in what actually wins in construction.
Most ConTech pitches start with a slide about how construction is broken. Jeevan Kalanithi’s started with a phone call from a superintendent named Tim, who said the company’s prototype had just saved him $50,000 in ten minutes by settling an argument over a covered-up inspection.
That was OpenSpace with two job sites on the platform, a clunky workflow only the three founders could operate, and zero revenue. Today the company has captured more than 40 billion square feet of jobsite imagery, employs around 350 people, and sits at the centre of one of construction technology’s hottest categories: spatial AI.
The path between those two moments holds the lessons that matter, both for builders evaluating where to spend their tech dollars next, and for founders trying to sell into an industry that, by most measures, has lost productivity over the last two decades.
This is the OpenSpace founding story, the field-first playbook behind it, and where spatial AI is heading next, told through Jeevan’s own words from the Bricks & Bytes interview.
The 3D Robotics lessons that became the OpenSpace thesis
Why a picture really is worth a thousand words on a jobsite
Before OpenSpace, Jeevan ran product and then served as president of 3D Robotics, a drone company that pivoted into construction with a tool called Site Scan. The idea was simple: fly above a job and see what is going on. It worked, but only partly.
Three things stuck with him. First, builders trust pictures more than documents. A superintendent will argue all day about a spreadsheet, but visual evidence ends the argument fast. Second, drones only solve the outside-and-above part of the problem. As one super told him, “Most of what we do is not even ground up.” The real action happens indoors, on innings two and three. Third, busy people will not adopt tools that add work. “Which exact person is going to go and fly the drone around on Friday?” was a question nobody on a packed jobsite wanted to answer.
Those three lessons became the OpenSpace thesis. Capture reality comprehensively, not just from above. Make it stupidly simple. And tap into the way builders already want to work, which is by looking at the thing and dealing with it.
The Tim story: $50,000 in ten minutes
How an early customer call proved the company was real
When OpenSpace was just three founders and two job sites, only Jeevan, Mike Fleischman and Philip DeCamp could actually run the software. Uploading videos required typing into a script. There was no app, no automated workflow, no polish. There was a working idea and not much else.
Then Tim called. He had been in a dispute with a drywall contractor who claimed to have given MEP an inspection date before covering up the work. Tim had OpenSpace running on the site. He sat at a computer, used the playback tool to rewind time on that exact location, and proved his version was correct. The avoided change order: roughly $50,000. The time to resolve: about ten minutes. The relationship with the drywall sub: still intact, because, as Tim put it to Jeevan, as long as you have the facts, people own up and move on.
For a three-person company, the maths were obvious. If a half-built prototype could save one superintendent $50K in ten minutes, the underlying behaviour change was real. The team raised price, signed a few hundred thousand in early sales, and the company moved from “interesting demo” to “thing people pay for.”
If we’re doing that, I can probably extrapolate that this concept is going to scale up.Jeevan Kalanithi, CEO of OpenSpace
The image-first workflow as a return to a better old way
Why a 60-year-old super told Jeevan the new tech felt familiar
One of the more interesting moments in the conversation was when a veteran superintendent told Jeevan that OpenSpace reminded him of how he used to run jobs in the 1960s. Back then, if an outlet was two feet off, he would just huddle with the crew responsible, point at the issue, and get it moved. Project keeps moving.
Now, the same fix involves filling out a form on a phone, routing it to someone off-site who has no spatial context, a string of emails, and three weeks of delay. The image-first workflow short-circuits all of that. Look at the problem, deal with it, move on. The technology is cutting edge, but the working pattern is older than most ConTech founders.
This matters more than it sounds. Global construction productivity grew just 0.4% annually between 2000 and 2022, compared with 2% across the broader economy and 3% in manufacturing. A lot of that gap is administrative drag: paperwork, communication latency, disputes resolved through written narrative instead of shared visual context. Tools that put image and location at the centre of the workflow attack exactly that drag.
Get the inside view on ConTech founders, before everyone else
Deep takes on the deals, products, and operators reshaping construction. Every week.
Join 2500+ ReadersThe Berkeley jobsite that rewrote the product
The customer feedback that forced a harder technical bet
Early on, when the team had a working prototype but no real capture method, Jeevan was on a Berkeley jobsite showing a PM the visual record of a podium before concrete was poured. The PM loved the output. Then he asked how long it took to capture.
About an hour for 50,000 square feet. The PM’s response: “Cool idea, but we’re never going to freaking use it.” He went on to describe what he actually wanted: walk the site like a normal site walk, with a camera on his hardhat, and let the software produce the rest automatically. That was harder to build, but it was also competitive moat. The team accepted the technical hit, and that is the version of OpenSpace that exists today: snap a 360 camera on, walk the job, and the image record assembles itself.
Where spatial AI goes next
Field agents, schedule automation, and the visual intelligence layer
OpenSpace now calls itself a visual intelligence platform, with spatial AI as the engine underneath. The thesis is that the generative AI wave most people read about is built around text and documents, which makes sense for the internet, but is a poor fit for builders who manipulate physical reality.
Spatial AI is the version trained on jobsite imagery, plans, specs, and location data, capable of pulling intelligence out of three-dimensional space rather than language. OpenSpace’s Progress Tracking product, strengthened by its November 2025 acquisition of Disperse, already feeds verified schedule updates into P6 and flags productivity issues. Their newer Auto Location feature estimates the indoor position of a smartphone without GPS, which opens the door to field agents that know where you are and what is around you.
The longer-term picture Jeevan describes is one where a junior superintendent can ask a question from inside a specific room and get back relevant issues, the right person to solve them, and the plan history of that exact space. With the industry facing a deep labour shortage, augmenting field staff so they perform like someone with fifteen more years of experience is not a luxury. It is one of the more credible answers to a problem that has nothing to do with software.
Founder lessons for selling into construction
Customer first, money first, simplicity always
Jeevan’s advice to ConTech founders is unromantic. Spend 80% of your early time on jobsites with the people who write cheques, not the people with “innovation” in their job titles. Charge customers from day one. If they will not pay, you have not found the angle yet. The flip side is the warning he gives about his own mistakes: OpenSpace at one point accumulated such a heavy feature backlog that the team felt buried and customers felt ignored. The fix was throwing the backlog out and trusting the team on the top three things customers really wanted.
The pattern repeats across most ConTech failure modes Bricks & Bytes has covered. Tech-first founders build elegant products nobody asked for. Industry-first founders build features nobody wants to pay for. The winners do field time, charge real money quickly, and stay ruthless about what they will not build.
OpenSpace vs the rest: how it fits the ConTech stack
| Tool category | Primary job | OpenSpace’s role |
|---|---|---|
| Project management (Procore, Autodesk Construction Cloud) | Workflow, documents, RFIs, scheduling | Integrates as the visual layer; not a replacement |
| BIM platforms (Revit, Navisworks) | Design coordination and model authoring | Compares as-built imagery to BIM models |
| Reality capture peers (HoloBuilder, StructionSite, Disperse) | 360 capture and progress documentation | Acquired Disperse in 2025; now market leader by scale |
| Drone capture (DroneDeploy, Skydio) | Outdoor and ground-up aerial documentation | OpenSpace Air integrates drone capture into the platform |
| Generic spatial AI startups | Computer vision applied to physical environments | Trained on the largest jobsite imagery dataset in the category |
OpenSpace is a visual intelligence platform for construction. A worker clips a 360 camera onto their hardhat, walks the site normally, and the software stitches the footage into a navigable visual record pinned to the floor plan. The output gets used for progress tracking, dispute resolution, and increasingly for AI-driven schedule and productivity insights. (Source)
OpenSpace has raised a total of around $199M across its funding history, including a $102M Series D in March 2022 led by PSP Growth at a $902M valuation, with an additional $9M added later that year. The platform has been used across more than 94 countries and has captured over 40 billion square feet of jobsite imagery. (Source)
Spatial AI is the branch of artificial intelligence focused on understanding three-dimensional physical environments, rather than text and documents. For builders, that matters because most jobsite decisions depend on what is actually there, not what a document says is there. OpenSpace combines visual data, plans, specs, and location to produce insights like schedule status, productivity analysis, and location-aware issue flagging. (Source)
In November 2025, OpenSpace acquired London-based Disperse, a progress-tracking platform combining 360 imagery with machine learning and human verification. The deal followed a June 2025 partnership and consolidates two of the largest players in image-based progress intelligence. Customers reported improvements in billing accuracy, schedule reliability, and issue resolution speed after the integration. (Source)
Spend most of your early time on actual jobsites talking to people who write cheques, not at innovation days. Charge customers from day one, because if they will not pay for the prototype, you have not yet found a real angle. And build technology that removes work from a busy person’s day, not technology that adds another tool to their stack.
OpenSpace pursued a venture-heavy strategy, raising close to $200M and building scale fast. HoloBuilder took a leaner path, raising around $8M before being acquired by Hexagon’s PPM division. Both built reality capture products; OpenSpace’s later acquisition of Disperse and pivot to spatial AI has positioned it as the category leader by capture volume and AI capability. (Source)
Related Articles
Breaking The $10M ARR Barrier with Kevin Halter, CRO of OpenSpace
Latest Construction Technology Funding Rounds, 11th May 2026
