Dili raises $21.7 million to automate the compliance paperwork slowing AI infrastructure buildout

Dili, a Y Combinator S23 company based in New York, has raised $21.7 million to automate prevailing wage tracking, apprenticeship monitoring, and certified payroll review for AI infrastructure projects. The startup's software helps customers secure federal tax credits tied to the Inflation Reduction Act, with the company claiming it has already helped secure more than $1 billion in federal funding.

Dili's strongest story isn't an unverified funding headline. It's that AI infrastructure projects are now colliding with labor compliance rules that can decide whether federal tax credits actually pay out. The AI infrastructure boom is usually told in chips and power. Nvidia orders. Utility interconnection queues. Data center campuses measured in gigawatts. Look closer and you'll find another obstacle sitting between a signed contract and a live facility: paperwork. Dili, a Y Combinator S23 company based in New York, is selling into that quieter problem. The startup builds AI software for prevailing wage tracking, apprenticeship monitoring, certified payroll review, and related compliance work across energy, infrastructure, data centers, advanced manufacturing, and construction. That is a long list of boring-sounding chores. According to Y Combinator's company profile, Dili says it has helped customers secure more than $1 billion in federal funding by running prevailing wage and apprenticeship compliance processes. That is real money. The reason the category matters is not hard to see. The Inflation Reduction Act tied larger clean energy tax credits to specific labor obligations. The IRS says taxpayers can generally multiply certain clean energy credit or deduction amounts by five if they meet prevailing wage and registered apprenticeship requirements. Miss them and you are not just cleaning up a form - you may be gutting the economics that made the project worth building in the first place, on a project that could run to hundreds of millions of dollars. That is a significant exposure. The compliance work got bigger after the IRA Prevailing wage and apprenticeship compliance is not new. The Davis-Bacon Act has been around for decades - long before clean energy tax credits became the centrepiece of American industrial policy. What changed is the volume of projects pulled into this kind of recordkeeping after the IRA expanded clean energy incentives and connected them to labor standards. The IRS guidance says taxpayers must keep records showing hours worked in each classification, actual wages and fringe benefits paid, applicable wage determinations, and other project details. For a developer, EPC, or subcontractor network, that becomes a weekly operating burden. Dili's product is built around exactly those chores. Its website says the platform ingests payroll files in formats such as PDF, Excel, and CSV, checks them against wage determinations from SAM.gov, flags underpayments or classification problems, tracks apprentice ratios, and generates audit-ready reports. You can understand the appeal immediately if you've ever seen a large construction project with dozens of subcontractors. One bad classification can travel a long way before anyone catches it. Frankly, this is a better use of AI than another chatbot sitting on top of a knowledge base. The work is structured. The rules are checkable. The stakes are high enough that a human still needs oversight, but the manual review itself is exactly the kind of repetitive process software should compress. Dili has also changed its centre of gravity. When TechCrunch covered the company in February 2024, it reported that Dili had raised $3.6 million to automate due diligence for private equity and venture capital teams. The founders, Stephanie Song, Brian Fernandez, and Anand Chaturvedi, had worked together at Coinbase, and the original pitch was about speeding up investment analysis. The current YC and company pages now put much more emphasis on PWA compliance and physical industries. That pivot is not cosmetic. It points the same automation idea at a market where a mistake can cost tax credits, delay a financing close, or trigger an audit. Data centers make the timing sharper The compliance market would be interesting even without AI data centers. With them, it becomes harder to ignore. Brookfield and NextEra Energy announced a $100 billion AI data center and energy complex at the Department of Energy's Paducah site in Kentucky, according to reports from The Wall Street Journal and the Financial Times. The project is expected to involve up to 1.8 gigawatts of utility capacity and 1.2 gigawatts of compute capacity, with completion targeted for 2032. Those figures are the visible part. Behind them sits the administrative work that follows federal land, clean energy incentives, construction labor, utilities, contractors, and local permitting. Power is still the headline bottleneck, of course. Permitting still slows projects. But if you're building against federal incentives, compliance becomes part of the construction schedule, not an afterthought. That's the opening Dili is trying to own. The company is not selling a glamorous layer of the AI boom. It is selling a way to keep the documents straight when capital projects get bigger, faster, and more exposed to federal rules. Bigger software companies may eventually notice the same market. For now, Dili has the advantage of being early in a narrow workflow that suddenly matters more than it used to. The honest version of this story does not need an unconfirmed funding number to work. Dili's real test is whether customers trust AI to read payroll, flag compliance risk, and do it before an auditor or tax equity partner does. If it can do that reliably, the boring paperwork becomes the product. In infrastructure, boring is often where the money is. Also read: Europe bets €10 billion on AI gigafactories but Nvidia still holds the keys https://startupfortune.com/europe-bets-10-billion-on-ai-gigafactories-but-nvidia-still-holds-the-keys/ • The EU is betting €30 billion on AI gigafactories but it's playing catch-up in a race it's already losing https://startupfortune.com/the-eu-is-betting-30-billion-on-ai-gigafactories-but-its-playing-catch-up-in-a-race-its-already-losing/ • Microsoft posted record fiscal 2026 revenue as Azure crossed $100 billion and proved its AI bet is paying off https://startupfortune.com/microsoft-posted-record-fiscal-2026-revenue-as-azure-crossed-100-billion-and-proved-its-ai-bet-is-paying-off/