Spend a day with a concrete estimator and you'll understand why the job breeds a particular kind of fatigue. There are structural drawings to decode, cross-references that snake from foundation plans to obscure detail callouts three sheets away, rebar lap splices to calculate by hand. Days disappear into a single bid. And then, often as not, you lose the work anyway.
That operational reality puts a hard ceiling on how many projects a typical concrete subcontractor can even chase. Which is why Rudus—a startup so new its two founders are barely out of their Y Combinator cohort—believes it has found a wedge. The pitch is straightforward, almost blunt: feed us your structural PDFs, let the AI do the detective work, and reclaim 70% of your estimation time.
If the numbers hold, a contractor could theoretically bid three to five times as many jobs without expanding headcount, potentially unlocking 50% more revenue per estimator. That's the kind of leverage that gets attention in a fragmented, low-margin industry.
Narrow by Design
Rudus emerged from Y Combinator's Spring 2026 batch and formally announced its platform in late May, though precise customer traction remains undisclosed. What the company has made clear is its refusal to generalize. "Built for this trade and only this trade," the founders wrote in their YC launch post—a pointed jab at the general-contractor-oriented tools that dominate the takeoff software landscape.
The tool itself tackles the specific drudgery of concrete estimation. Upload a set of drawings and the platform auto-sorts them: foundation plans here, section details there, footing schedules over there. The AI then identifies concrete elements—slabs, footings, walls, columns—chases down cross-references to pin down dimensions and detailing, and expands those elements into full assembly line items. Concrete volumes, formwork areas, rebar quantities, development lengths, lap splices. The output is structured for how concrete subs actually price work, exporting cleanly to industry standards like HCSS HeavyBid, Sage Estimating, B2W, or plain Excel.
A demo video from mid-April showed features like automatic sheet naming, footing autocomplete with pattern matching, and a natural-language query function—"ask the plan set anything"—that cites sources back to the original drawings. Useful, if it works reliably at scale.
The Founders
Rishi Pankhaniya and Sahil Goel don't come from concrete. Pankhaniya's resume runs through Airbnb (reportedly the first new-grad hire in over two years) and Apple. Goel's background spans Amazon, NASA's Jet Propulsion Laboratory, Northrop Grumman. But Goel did study both Computer Science and Construction Management at Cal Poly San Luis Obispo, a dual track that perhaps explains the product's trade-specific granularity.
The company has courted industry validation early—Pankhaniya posted a public thank-you to the American Society of Concrete Contractors for a "warm welcome" in early May. Whether that translates to meaningful adoption is another question. The onboarding flow promises demos using a contractor's own drawings on Day 1, calibration runs on past jobs by Day 2, live use within two weeks. Pricing? Undisclosed. The website offers only a "Request Demo" button.
A Market Suddenly Crowded
Rudus is hardly alone in promising AI-powered speed gains. ConstructConnect announced "Takeoff Boost" in late April at Google Cloud Next 2026—an AI-assisted service built on Google's Gemini Enterprise. Established players continue iterating: Procore with its Estimating/Takeoff tool (formerly Esticom), Autodesk with Forma Takeoff and the older Autodesk Takeoff in Construction Cloud.
Then there's the swarm of newer entrants. Togal.AI claims up to 98% accuracy. BuildVision AI touts 8-minute average takeoffs for concrete work. CivilTakeoff is using Claude vision models. There are a half-dozen others, all marketing variations on the same promise: upload drawings, get quantities faster.
What Rudus is betting on is depth over breadth. Where competitors cast wide nets across multiple trades or aim at general contractors, Rudus is solving for the idiosyncrasies of concrete estimation—rebar assembly expansion, development length calculations, the specific output formats that concrete subs feed into their bidding software. Whether that vertical specialization creates a moat or simply accelerates feature parity with better-funded competitors is an open question. The team is two people. No live customer case studies have surfaced yet.
An Industry Ready, Maybe
Still, the backdrop is favorable—perhaps more than the founders expected when they started building. A Dodge Construction Network survey published early last year suggested AI is nearing a "tipping point" among contractors. A 2025 RICS report found that a majority of project managers and quantity surveyors expect AI to deliver greater value than current tools allow.
For an industry still running on manual takeoffs and Excel macros, tools that genuinely compress estimation cycles could matter. Whether Rudus becomes one of those tools, or gets absorbed into the noise of an AI land rush, depends on execution in a market where everyone is suddenly promising the same thing.
The concrete contractors will decide soon enough. They always do.
