Pulley Raises 64 Million Dollars as Permitting Delays Hold Back Construction
Pulley has raised $64 million to address an expensive part of commercial construction that happens before the first machine arrives on site. The financing announced on November 18 combines a $41 million Series B led by Jump Capital, with Catalyst Investors participating, and a $23 million Series A led by CRV and Susa Ventures. Pulley says annual recurring revenue has grown fivefold over the past year. It has not disclosed the revenue base behind that multiple.
The headline sounds like a funding story for another AI software company. The more useful question is whether permitting has become a category of industrial infrastructure in its own right. A factory can have financing, a site and a contractor, yet still wait months for a permit application to clear. That delay consumes time on the project’s critical path, and there is no practical way to recover it by pouring concrete faster later.
The application is part of the construction schedule
Pulley combines a team of more than 75 licensed specialists with AI agents and a database of regulatory requirements. It checks plans, identifies likely objections, develops approval strategies and handles interactions with the authorities that review applications. The company also says it is authorized to conduct private plan reviews through programs covering more than 2,000 cities and counties. Those arrangements can shorten a review in jurisdictions where they apply, although they do not erase zoning requirements or every approval a developer needs.
The company reports that its customers have collectively saved 11,750 weeks of project time across $21 billion in commercial construction in all 50 states. The comparison matters as much as the large number. Pulley has not published enough detail in the announcement to independently establish how each saved week was measured or what the same projects would have experienced without its service. Still, the reported scope makes clear that the company is selling a repeatable process to organizations managing many sites, rather than a consulting engagement for one troublesome building.
Its stated customer base includes Amazon, Starbucks, JLL and Equinix. These businesses have different buildings, but a shared problem: delays multiply when the same operator tries to open stores, warehouses, offices or infrastructure in dozens of jurisdictions. A local rule that is manageable for one architect becomes a portfolio problem when nobody knows it until a reviewer sends the drawings back.
Pulley says customers typically receive only one round of comments per project, while more than 40% of projects are approved without comments. It also says review comments have fallen 65% in 2026 compared with 2025. These are useful operating indicators, though fewer comments do not by themselves prove a matching reduction in total calendar time. Application complexity, project mix and local agency workloads can change from year to year.
The larger prize comes before the permit
The next products move Pulley upstream, into decisions normally made before an application exists. The company plans tools that compare potential sites using zoning rules, utility availability, local sentiment and other constraints. It is also introducing AI-powered drawing review intended to catch building-code problems while designs are still being developed.
This is where permitting knowledge could become more valuable than an application checklist. A developer that learns late that a site needs an expensive redesign has lost more than review time. It may have signed a lease, committed to equipment, or promised an opening date. A reliable early warning would let the company reject the site or modify the design before those costs are locked in.
There are limits. A model can flag a code conflict and a specialist can negotiate a clearer submission, but neither can manufacture grid capacity, make a controversial rezoning politically acceptable, or guarantee the decision of a local official. The new site-selection tools may therefore prove their value by helping customers abandon unsuitable projects earlier, as much as by speeding up the projects they pursue.
Investors are paying for regulatory memory
The interesting asset here is accumulated knowledge of how different jurisdictions actually interpret their rules. Building codes are published; the practical sequence of submissions, exceptions, reviewer expectations and corrections is harder to package. Pulley wants to capture that experience from one project and apply it to the next, with licensed experts checking what automated systems find.
That creates a plausible advantage for a company operating across thousands of local authorities. It also creates a test. If the system really learns, a growing project base should produce fewer revisions and more predictable approval dates, even after adjusting for project size and complexity. A headline count of AI agents or jurisdictions served is less informative than that outcome.
The funding gives Pulley room to expand beyond permit processing into site selection and design compliance, but the business case will be settled by schedules. Developers need to know when a building can actually begin, and how much confidence to place in that date. Reducing the uncertainty before construction starts may be worth more than another incremental gain once construction is underway.