September 2026 AI Construction Roundup: A Robotics Surge, Procore's $845M Drone Bet, and Cheaper AI Takeoff

Welcome back to our monthly roundup of AI and construction-technology news worth a contractor's attention. The filter hasn't changed: does this actually move anything for the people estimating jobs, chasing bids, or running the work? We skip the hype, verify every item against a primary or reputable source, and give you a plain read on what it means. August was a heavy robotics month and a busy one for deals — here's what landed over the past five weeks or so.

Table of Contents

• Product Launches
• Funding and Deals
• Robotics on the Jobsite
• The Bottom Line

Product Launches

ConstructConnect adds Auto Scale and AI Auto Name to Takeoff Boost

ConstructConnect expanded its Takeoff Boost suite with two features slated for August 2026: Auto Scale, which reads a plan set and applies the correct page scale automatically, and AI-Powered Auto Name, which names project elements to cut down setup time. They join the suite's existing Auto Takeoff, Auto Count, Auto Link, and Auto Bookmark tools.

What it means for you: These are unglamorous features, and that's the point. Scale setup and project naming are the tedious first ten minutes of every takeoff — the part nobody bills for and everybody rushes. Getting scale wrong at the start quietly poisons every measurement downstream, so an auto-scale step that's actually reliable is worth more than its demo makes it look. If you already run Takeoff Boost, the sane test is to let Auto Scale set a sheet you've already scaled by hand and confirm the numbers match before you trust it on a live bid.

A new browser-based takeoff tool aims at the shops priced out of desktop software

A new entrant, ContractorCounter, launched a browser-based takeoff and estimating platform on August 28 with AI-assisted quantity measurement aimed squarely at subcontractors, trades, and small builders. It runs in the browser on iPad, Android, Mac, or Windows with no install, takes plain-English instructions ("count the doors," "measure the slab"), and pushes quantities straight into a bill of quantities for pricing. Pricing starts at $14 a month, with a $249 lifetime option — the founder's pitch is that "estimators shouldn't need a desktop licence and a training course to count doors and measure a slab."

What it means for you: Treat this one as a signal about direction more than a vendor recommendation — it's an early, self-announced launch, so kick the tires before you bet a bid on it. But the trend is real: AI-assisted takeoff is sliding down-market from enterprise seats toward the one- and two-person estimating shops that have historically made do with a PDF markup tool and a calculator. If you're a small sub who's been counting devices by hand because Bluebeam or PlanSwift felt like overkill, cheap browser-based options are finally worth a look. Just apply the same rule you'd apply to any of them: pilot it against a job you've already estimated the old way, and check the counts.

Funding and Deals

Procore buys DroneDeploy for $845M to feed its "digital coworkers"

On July 30, Procore announced an $845 million all-cash deal to acquire DroneDeploy, the reality-capture and visual-intelligence platform used in more than 180 countries. Procore's rationale is data: pair DroneDeploy's aerial capture and its library of labeled site imagery — including more than 100,000 tagged safety issues — with Procore's own trove of roughly 400 million photos and 126 million-plus drawings, and you have the fuel for the "digital coworker" AI agents Procore has been rolling out. The deal is expected to close later this year.

What it means for you: This is the clearest sign yet that the big platforms see proprietary jobsite data, not clever models, as the moat. For contractors already on Procore, expect reality capture and progress tracking to get pulled deeper into the core product over the next year — potentially useful if you're paying separately for drone documentation today. The quieter question worth asking your platform rep: as these acquisitions stack up, what exactly happens to the photos, drawings, and site data you've been feeding in, and how is it used to train the AI you're now being sold?

Fluxco raises $26M to attack the transformer bottleneck

Austin-based Fluxco closed a $26 million seed round co-led by 8VC and Congruent Ventures, with Trust Ventures and Overture Ventures joining. Founded in 2025, Fluxco runs an AI procurement platform for electrical grid equipment: it parses complicated utility transformer specifications, matches them against competitive bids from more than 150 global OEMs within days, and then handles leasing, logistics, and servicing. The company says it has already moved orders for 1,000-plus transformers across 34 projects and about $30 million in orders in its first four months.

What it means for you: Anyone building data centers, substations, or anything that ties into the grid knows transformers are the long-lead item that can stall a whole schedule — quoted in months, sometimes north of a year. Fluxco is a narrow bet on one brutal procurement pain point, and it's early, but the shape of it matters: AI turning a dense spec into a matched, sourced order is exactly the kind of procurement work that eats preconstruction and purchasing hours across every trade. Electrical contractors and anyone chasing energy-infrastructure work should watch whether "spec in, sourced quote out" tooling shows up for the equipment they wait on too.

Robotics on the Jobsite

Autonomous excavators go live with real GCs — and the money keeps coming

August was the month jobsite robotics stopped being a demo reel. Bedrock Robotics, founded by Waymo alumni, put fully autonomous excavators to work on real jobsites for two established contractors — Sundt Construction and Zachry Construction — in an announcement dated August 19. Bedrock raised a $270 million Series B back in February, with Nvidia's venture arm and Tishman Speyer among the backers. A day earlier, Gravis Robotics landed a $200 million Series A from SoftBank, and FieldAI (backed by Nvidia and Bill Gates) has now pulled in $405 million across two rounds. By Construction Dive's count, the earthmoving-robotics field has raised close to $1 billion, and among contractors surveyed, robotics use grew 45% year over year.

What it means for you: "Autonomous excavator on a Sundt and Zachry site" is a different sentence than "pilot program," and that's the news. This is still limited to large, repetitive earthwork on big-budget jobs, so most GCs don't need to react this quarter. But if you self-perform sitework or dirt, it's worth understanding how these machines get priced and scheduled — by the hour, by the yard, as a rental with a supervisor — because that's the model that will eventually show up in a competitor's bid on a big pad or highway job. The labor story here is augmentation of a thin operator pool, not replacement, at least for now.

Burns & McDonnell puts AI robots on utility-scale solar

Burns & McDonnell and robotics firm Gritt announced a partnership on August 25 to deploy AI-powered robotics on utility-scale solar construction (PV Tech has more). Gritt's approach is to bolt intelligent, self-learning systems onto standard construction equipment so it can autonomously handle repetitive installation work — placing and assembling solar arrays, pouring concrete, and laying rebar. The two say they've spent the past year testing the technology at multiple real utility-scale sites, and they're framing it as complementing crews on repetitive tasks rather than replacing them.

What it means for you: The detail worth noting is "attaches to standard equipment." A robotics play that upgrades machines you already own is a very different adoption curve than one that asks you to buy a fleet of purpose-built robots. For contractors doing renewables and other large, repetitive-array work, the near-term watch item is whether an owner's engineer like Burns & McDonnell starts writing this kind of automation into how projects get planned and priced — which is how a pilot quietly becomes a bid expectation.

The Bottom Line

If August had a theme, it's that AI in construction is getting physical and getting consolidated at the same time. On one side, autonomous machines moved from staged pilots onto real GC jobsites and kept raising nine-figure rounds. On the other, the platform giants are buying up the data and the drones to feed their agents, while procurement and takeoff tools chip away at the office side of the business — including, finally, for the small shops that couldn't afford the enterprise stuff.

None of it changes the fundamentals of a good bid. The contractors getting value out of this wave, in our experience building specification-analysis tools, still aren't the ones chasing every launch — they're the ones picking one painful workflow, piloting a tool against a job they already know the answer to, and measuring before they scale. A robot on a solar array and an AI on a takeoff both live or die on the same question: does the output survive a second look from someone who knows the work? Ask that first, and most of the noise sorts itself out.

See It on Your Own Documents

If specification review is the workflow eating your preconstruction hours, we'd like to show you what purpose-built AI does with your actual bid documents — the non-standard clauses, the buried requirements, the risk hiding on page 400.

• Book a 15-minute intro call
• Request a demo

---

Every item above was verified against a primary or reputable source as of publication. Sources are linked inline. Last updated: September 1, 2026.