The Rules Take Effect
On June 25, 2024, the Treasury Department and IRS published final regulations that lock a fivefold tax-credit multiplier, the IRS reported, to prevailing-wage and apprenticeship rules most construction firms have never tracked at this granularity. For contractors who spent months operating under proposed guidance, the compliance framework is now settled, and the penalty for missing a single apprentice ratio or wage determination is the loss of the credit itself.
Published in the Federal Register on June 25 and effective August 26, the regulations amend ten IRA-modified or newly created code sections. They define how taxpayers satisfy the prevailing wage and apprenticeship requirements that determine whether a project receives the base credit or the increased amount. The final regulations generally apply to qualified facilities placed in service in tax years ending after June 25, 2024.
The prevailing wage requirement incorporates Department of Labor wage determinations by reference, meaning the applicable rate shifts by county, trade, and project type. The apprenticeship requirement layers three distinct obligations, the Treasury's figures put the thresholds at:
| Construction start | Labor-hours threshold |
|---|---|
| Before 2023 | 10% |
| 2023 | 12.5% |
| 2024 and later | 15% (according to the Federal Register) |
Plus a daily apprentice-to-journeyworker ratio tied to each registered program's standards, and a participation mandate: any employer with four or more workers on the project must hire at least one qualified apprentice. Miss any of the three and the apprenticeship requirement fails.
Penalties are structured to make noncompliance expensive. For prevailing wage shortfalls, the taxpayer owes the wage differential plus interest at an elevated underpayment rate, plus $5,000 per affected worker, the Treasury's data shows, (tripled to $10,000 if the IRS finds intentional disregard, the IRS found). Apprenticeship failures trigger a $50-per-labor-hour penalty, the Treasury's figures put, rising to $500 per hour for intentional disregard, the final regulations' figures put. Both regimes include cure provisions: pay the correction and penalty, and the taxpayer is deemed to have satisfied the requirement. But the cure only works if the taxpayer catches the error first.
Recordkeeping obligations mirror the enforcement design. Taxpayers must maintain payroll records for every laborer, mechanic, and apprentice employed by the taxpayer, contractor, or subcontractor: hours worked in each classification, actual wages and fringe benefits paid, and the prevailing wage determination applied. The records must be sufficient to establish compliance on audit. No sampling, no summaries. The IRS has signaled it will verify.
Two narrow exceptions relieve the burden entirely. The Beginning of Construction Exception covers facilities where construction began before January 29, 2023 (60 days after Notice 2022-61 provided initial guidance). The One Megawatt Exception exempts facilities under 1 MW AC from both requirements. A transition rule also carves out all work performed before January 29, 2023, regardless of exception eligibility. Treasury has indicated further regulations are coming for sections 48 and 48E.
For EPC contractors, developers, and the compliance teams that serve them, the regulatory text converts a tax incentive into a labor-data problem: track every hour, every classification, every wage determination, across every subcontractor, on every federally funded project continuously.
Software Steps In
The prevailing-wage and apprenticeship rules that took effect in June 2024 didn't just add paperwork — they multiplied it. A single federally funded project can involve dozens of subcontractors, each filing certified payroll weekly on forms that vary by state, each trade classification tied to a wage determination that changes by county, each apprentice hour tracked against a ratio that differs by craft and jurisdiction. Doing this in spreadsheets means thousands of manual cross-checks per project. Miss one, and the tax credit evaporates.
AI compliance platforms automate the entire chain. Dili, a New York startup founded in 2023 by Brian Fernandez and Anand Chaturvedi (who went through Y Combinator's Summer 2023 batch), built its system around the specific mechanics of Davis-Bacon, IRA, and related federal labor standards. The platform ingests certified payroll in any format (WH-347, state forms, PDF, Excel, CSV), validates every line against live wage rates pulled directly from SAM.gov across all 50 states and counties, calculates fringe benefits automatically, and flags underpayments, misclassifications, and documentation gaps before they become audit findings.
The apprenticeship side runs in parallel. Dili monitors apprentice-to-journeyman ratios in real time, tracks registered apprenticeship hours against RAPIDS certification data, and integrates with state apprenticeship programs so compliance doesn't depend on a subcontractor remembering to update a spreadsheet. The company says its average payroll review takes five minutes (down from seven-plus hours manually) and that it has processed nearly 14,000 payroll reports covering 4.7 million labor hours and 18,000 workers across more than 700 projects.
Subcontractor visibility is the force multiplier. Prime contractors are liable for every tier below them. Dili gives each sub a drag-and-drop upload portal, runs automated reminders, and surfaces risk scores so the compliance team sees which subs need attention before Friday's payroll lands. The platform syncs with ADP, Acumatica, and major project-management tools, and exposes an open API for custom workflows. Former Department of Labor enforcement officials on Dili's team review edge cases and run look-back assessments, retroactive audits of historical payroll that the company says surface issues in days rather than months.
The metrics are striking: an 85 percent reduction in compliance labor, reviews 48 times faster, more than $1 billion in federal funding protected across the customer base. But the operational shift matters more than the numbers. Compliance managers who used to spend weekends chasing missing forms now spend five minutes uploading reports and the rest of the week managing exceptions. The platform doesn't replace judgment — it concentrates it where the risk actually lives.
Inside Dili's Sprint
Dili has moved fast since the IRS finalized its rules. In a LinkedIn post, the company reported 215 percent month-on-month growth, a figure that lines up with the regulatory tailwind: the final regulations apply to such facilities, creating immediate demand for software that can track wage determinations, apprenticeship ratios, and certified payroll across multi-state projects.
Headcount is scaling to match. The team stood at 15 as of the most recent Y Combinator job posting, and Dili is actively hiring for five roles spanning engineering, product, sales, and operations:
| Role | Base salary | Equity |
|---|---|---|
| DevOps Engineer | $180,000–$220,000 | 0.10–0.20% |
| Senior Full-Stack Engineer | $125,000–$200,000 | 0.10–0.30% |
| Product Manager | $160,000–$190,000 | 0.10–0.20% |
| Compliance Customer Success Manager | $100,000–$130,000 | 0.05–0.20% |
| Founding Business Development Representative | $65,000–$95,000 | 0.05–0.10% |
All roles are based in New York or remote within the U.S., with the BDR role specifying California as an additional option. The DevOps Engineer posting calls for a platform engineer to own infrastructure across AWS, Terraform, GitHub Actions, observability, security, and developer tooling, a signal that the engineering team is investing heavily in reliability as transaction volumes climb.
The hiring push reflects a broader product bet: Dili is positioning itself as the system of record for labor compliance on federally funded physical-industry projects. Its roadmap emphasizes moving beyond point-in-time audits to continuous monitoring for the life of a tax credit, a capability that matters because the IRA's prevailing-wage and apprenticeship requirements persist for the credit period (up to 10 years for some energy credits). With the IRS regulations now final, the compliance window is open and the data volume is growing. Dili's capital and the current hiring sprint are aimed at capturing that volume before incumbents can retrofit their platforms.
Incumbents Watch the Gap
The regulatory pressure from the IRS's June 2024 prevailing-wage rules is not only lifting startups. Two public companies that already sit on thousands of construction and energy job sites (Samsara and Procore) are layering AI into their existing platforms, though their entry points differ from the pure-play compliance automation that Dili sells.
Samsara (NYSE: IOT), which has put more than $1 billion into R&D since its founding, built its reputation on vehicle telematics and site safety. Its Connected Operations Platform now ingests over 100 billion miles of operator data and, as of Q4 FY25, construction delivered the company's highest net-new annual contract value mix for the sixth straight quarter. In June 2025, Samsara unveiled more than a dozen AI-powered tools aimed at physical operations: AI Safety Intelligence that flags risky driving in real time, a multicamera system giving drivers 360-degree visibility, weather overlays from the National Weather Service, a wearable with fall detection and one-year battery life, and Fault Code Intelligence that turns vehicle diagnostics into automated work orders. The company says its AI dash cams have prevented 380,000-plus accidents and cut crash rates by nearly 75 percent across 2,600 fleets. Customers such as Emery Sapp & Sons report a 25 percent drop in at-fault accidents and a 40 percent year-over-year decline in DOT-reportable incidents.
None of those features, however, directly automate certified payroll review, apprenticeship-hour tracking, or wage-determination mapping, the core compliance tasks the new IRA rules demand. Samsara's platform captures labor-adjacent data (hours behind the wheel, time on site, vehicle location) that could feed a compliance engine, but the company has not announced a prevailing-wage module. Its June 2025 launch was squarely targeted at safety, routing, and maintenance. If Samsara moves into wage compliance, it will likely do so by exposing its site-level telemetry to partners or by building on top of the HappyRobot integration now available in its App Marketplace, which automates 20 million-plus conversations a year for fleet operators.
Procore (NYSE: PCOR), the construction management software incumbent, took a different route. It rolled out an expanded Procore AI experience powered by Datagrid, a startup it acquired to embed AI agents directly into the platform. The new agents review RFIs, submittals, and contracts, document-heavy workflows that today consume hours of project-engineer time. Procore emphasizes that the AI is not a separate login; it works inside the existing project dashboard, surfacing insights and automations where superintendents and project managers already work. The Datagrid integration also syncs project financials and runs agents across RFIs and budgets.
Again, the announced capabilities center on document review and cost control, not on the certified-payroll and apprenticeship records that the IRS final regulations now require for tax-credit eligibility. Procore's platform already stores subcontractor pay applications and lien waivers, so the data substrate for wage compliance exists. But as of the announcement, the company had not released a dedicated prevailing-wage or apprenticeship-tracking agent. The gap is notable: Procore's customer base includes the EPCs and general contractors that must now prove, project by project, that every craft worker was paid the correct Davis-Bacon rate and that apprenticeship ratios were met for the life of the credit.
Both companies have the distribution, the data gravity, and the R&D budgets to close that gap quickly. Samsara's site-level telemetry could timestamp worker presence; Procore's document cloud already holds the payroll packets. What neither has done — at least publicly — is productize those assets for the specific compliance burden the IRA rules created. Until they do, the pure-play automation layer that Dili and a handful of other startups are building remains the only tool purpose-built for the new regulatory regime.
Money Follows the Mandate
The regulatory tailwind from the final regulations has arrived as the broader AI funding environment hits historic highs. Global AI funding climbed 59 percent quarter-over-quarter in Q2 2024 to $23.2 billion — the highest quarterly level on record — and then surged another 51 percent in Q1 2025 to $66.6 billion across 1,134 deals. The average AI deal size reached $28.9 million in 2024, up 55 percent from $18.6 million in full-year 2023, and 32 mega-rounds of $100 million or more closed in Q2 2024 alone. The United States captured 76 percent of total AI funding and 49 percent of deals in 2024, concentrating capital in the same jurisdiction where the prevailing-wage and apprenticeship rules now apply.
M&A activity has tracked the funding surge. Industrial software M&A transactions reached 86 in 2024, a 37 percent increase from 2023, with European startups accounting for over a third of exits, 136 deals, up 7 percent year-over-year. Strategic buyers have been aggressive: Nvidia completed seven AI acquisitions in 2024 versus one in 2023; Databricks and Snowflake each acquired five AI startups since 2023; Siemens agreed to buy Altair for roughly $10 billion enterprise value in October 2024; Cisco closed its $28 billion Splunk acquisition in March 2024; and HPE moved to acquire Juniper Networks for approximately $14 billion. Thoma Bravo's $5.3 billion take-private of cybersecurity AI firm Darktrace in October 2024 underscored private-equity appetite for applied AI. Industrial software, the category that includes compliance automation for energy and infrastructure, has traded at a premium to broader SaaS and attracted more than $20 billion in venture capital over the past three years.
The compliance-automation niche sits at the intersection of two of these trends: regulatory-driven demand and industrial-software premiums. Dili, a Y Combinator-backed (S23) startup founded in 2023, has grown to 15 employees in New York and is hiring across sales, engineering, product, and operations while supporting customers on projects that have secured over $1 billion in federal funding. Its focus on a rule set that took effect for facilities placed in service after June 25, 2024, makes it a prototype for the category: a narrow, high-stakes workflow that incumbents such as Samsara and Procore are now embedding via AI agents and acquisitions.
Not every exit will be a premium strategic sale. Market observers note a parallel track of distressed transactions — companies running out of cash and selling at one to two times revenue — as VCs decline to fund further rounds. Reverse acquihires have also proliferated: Google's $2.7 billion Character.AI arrangement and Amazon's Covariant acquisition let giants absorb talent and technology without triggering full antitrust review. For compliance-automation founders, the playbook is narrowing: build audit-ready, policy-driven controls that turn regulatory adherence into a product capability, and the acquirer set expands from pure-play construction-tech strategics to the industrial-software platforms and private-equity funds already paying premiums for that exact asset class. With the IRA tax-credit regime locking in prevailing-wage requirements for the credit life, the compliance workload — and the M&A rationale — is structural, not cyclical.
What This Story Leaves Out
The IRA prevailing-wage and apprenticeship rules govern a specific slice of federal climate and infrastructure spending: tax-credit projects (Sections 45, 45Y, 48, 48C, 179D, and others) and certain direct-pay programs where the Treasury's June 2024 final regulations apply. This story tracks the AI tools built to automate compliance with those rules: wage determinations from SAM.gov, apprenticeship ratios, certified payroll review, and audit-ready reporting for EPCs, developers, and their subcontractors across energy, data-center, and heavy-civil projects. It does not cover the adjacent but distinct AI compliance markets that have grown up around other regulatory regimes.
The IRS regulations don't expire. Those requirements persist for the credit's duration (up to ten years on some energy projects). That means the spreadsheet nightmare isn't a one-time scramble; it's a standing order. The contractors who automate now aren't just chasing a credit. They're building the compliance infrastructure that the next decade of federal energy spending will run on.
Working in frontier tech? Zero G Talent tracks the openings: see every open ASML role, browse frontier tech jobs, openings at Stripe, and the people building the field.