Transit agency data projects, explained: the funding line decides the deal

Nearly every dollar a U.S. transit agency spends on capital work arrives through a Federal Transit Administration program with a statutory number, a fixed federal share, a written list of eligible activities, and a date after which the money disappears. A data project that fits one of those lines gets bought on a predictable schedule. A data project that fits none of them competes against brake jobs and bus tires for scarce local dollars, and loses. Firms that lose in transit usually built the right system against the wrong funding line.

Where a transit agency's money comes from

Four formula programs carry most of the weight. Section 5307 sends money to urbanized areas. Section 5337 pays for state of good repair on fixed guideway and high-intensity motorbus. Section 5339 buys buses and bus facilities. Section 5311 funds rural areas. The scale is public: FTA's FY2025 apportionment notice records $20,937,068,868 appropriated, of which $7,394,716,046 was apportioned under Section 5307, $4,349,878,139 was available under Section 5337, $2,250,495,863 across the three Section 5339 components, and $956,643,454 for rural areas.

The federal share is the first number to internalize. Under 49 U.S.C. 5307, a capital grant "shall be for 80 percent of the net project cost," and a grant for operating expenses "may not exceed 50 percent of the net project cost." Vehicle-related equipment and facilities required by the Americans with Disabilities Act or the Clean Air Act carry a 90 percent share for the portion attributable to that compliance, under 49 U.S.C. 5323. The remainder is local match, and local match is what agencies actually run out of. A vendor who cannot say where the 20 percent comes from is asking the agency to solve the hard half.

Operating assistance is deliberately hard to get in a large system. Under 5307, operating costs are broadly eligible in urbanized areas below 200,000 population. Above that line, the statute restricts operating assistance by fleet size: a system operating 75 or fewer buses in peak service may apply up to 75 percent of its apportioned share to operating, and a system with a minimum of 76 and a maximum of 100 buses may apply up to 50 percent. ADA complementary paratransit vehicles are excluded from that count. For a mid-size or large agency, a service contract classified as an operating expense has almost no federal money behind it.

Those availability windows are the most underused fact in transit business development. Section 5307 money apportioned in a given year can be obligated across a six-year window, so an agency that says "we have no budget this year" may be sitting on unobligated prior-year apportionments. A scope that phases across two fiscal years is easier for a grants manager to approve than one needing a single large obligation.

Capital or operating: the distinction that decides everything

The single most useful thing a data vendor can learn about transit is the statutory definition of "capital project" at 49 U.S.C. 5302. It is broader than most software firms assume. It expressly includes preventive maintenance; associated capital maintenance; leasing equipment or facilities for public transportation use; "introduction of new technology through innovative products"; "transit-related intelligent transportation systems"; and mobility management, described as short-range planning and coordination activities among transportation service providers.

Read those clauses again with a software contract in mind. A dispatch and scheduling system that qualifies as a transit-related intelligent transportation system is capital-eligible at an 80 percent federal share. A condition-monitoring pipeline feeding preventive maintenance argues from a category the statute names by hand. A shared-eligibility and trip-brokering layer across providers looks like mobility management. Leasing is capital-eligible, which is the doorway hosted software walks through when structured as a leased system rather than an indefinite subscription.

This is not a labeling trick and should not be sold as one. The classification has to be honest and survive an FTA review. What it means in practice is that the shape of the deal is a design decision, made early — during scoping, with the agency's grants staff in the room, not after a procurement is drafted around a monthly fee.

The FY2026 edge, and what a vendor should do about it

The current authorization has a hard stop. 49 U.S.C. 5338 authorizes contract authority from the Mass Transit Account of the Highway Trust Fund for fiscal years 2022 through 2026: $13,355,000,000 for FY2022 rising to $14,642,000,000 for FY2026. FY2026 is the last fiscal year the statute covers. What follows is a matter for Congress, and anyone who tells you they know the shape of the next surface transportation bill is guessing.

Two consequences follow, and neither requires a prediction. Agencies feel authorization edges before they arrive: capital committees turn conservative and discretionary pilots are deferred first, so a scope attached to an obligation the agency already carries outlasts one depending on new appetite. And the multi-year availability windows above are a real hedge — work funded from an existing apportionment with years left on the clock does not care what the next authorization says.

Four data workloads federal law already requires

The most reliable transit data work is what an agency is legally obligated to produce. These are not preferences a new executive can cancel. They are statutory duties with named deliverables, and the labor behind them already sits in the agency's budget as staff time.

National Transit Database reporting. 49 U.S.C. 5335 directs the Secretary to maintain a reporting system "using uniform categories to accumulate public transportation financial, operating, geographic service area coverage, and asset condition information," along with data on assaults on transit workers and fatalities resulting from bus impacts. It is not optional: the statute conditions grants under 5307 and 5311 on the applicant being subject to the reporting and uniform systems. FTA's most recent information-collection notice puts the collection at 2,914 annual respondents and 456,179 total annual burden hours. That is roughly 156 hours per reporter per year of pure data assembly and validation, and it is the clearest recurring pain in the sector.

The NTD is also moving. FTA's current approval covers July 2025 changes that consolidated station and facility reporting, clarified security reporting, added trip planning and geospatial data, and tightened state of good repair items. FTA separately rescinded the NTD Weekly Reference (WE-20) requirement in a final notice applicable January 15, 2026. A product that hard-codes last year's NTD forms is a liability.

Transit asset management. 49 U.S.C. 5326 requires a national transit asset management system, and 49 CFR Part 625 implements it. Recipients maintain capital asset inventories and condition assessments, use decision support tools and investment prioritization, set targets against FTA's state of good repair measures, and submit an annual report on progress plus next year's targets. Part 625 splits providers into two tiers: a Tier I provider owns, operates, or manages 101 or more vehicles in peak revenue service across all fixed route modes or in any one non-fixed-route mode, or operates rail transit; a Tier II provider has 100 or fewer such vehicles, or is a Section 5311 subrecipient, or is an American Indian tribe. Tier II providers may join a group plan, which is why state DOTs are often the real buyer for rural asset data work.

Agency safety plans. 49 U.S.C. 5329 and 49 CFR Part 673 require a Public Transportation Agency Safety Plan built on a Safety Management System, signed by a named Accountable Executive who also owns the TAM plan. It carries annual safety performance targets tied to the National Public Transportation Safety Plan measures and a documented annual review. A "large urbanized area provider" — defined in 49 CFR 673.5 as a 5307 recipient or subrecipient serving an urban area of 200,000 or more — carries extra obligations: a joint labor-management Safety Committee and a safety risk reduction program covering vehicular and pedestrian events and assaults on transit workers. Risk reduction targets are required only for those providers.

ADA complementary paratransit. 49 CFR 37.131 sets service criteria that are pure scheduling and records work: service within a corridor three-quarters of a mile on each side of each fixed route and within a three-quarter-mile radius of each rail station; next-day scheduling at any requested time, with negotiation limited to one hour before or after the rider's desired departure; a fare no more than twice the comparable full fixed-route fare; no restrictions or priorities based on trip purpose; the same hours and days as fixed route; and no capacity constraints, including no trip caps, no waiting lists, and no "operational pattern or practice that significantly limits the availability of service." Each is a measurable claim the agency must defend with its own trip records.

Read the bars as a purchasing argument, not a market forecast. Work near the top has a citation behind it, so a grants manager defends the spend by pointing at Part 625 or 49 U.S.C. 5335. Work near the bottom is often what an agency wants most, but it must be justified on merit — a longer sale, and the first casualty in a tight year.

GTFS is the cheapest credibility available

The General Transit Feed Specification is an open standard, maintained by MobilityData and published under the Apache 2.0 license. GTFS Schedule is a set of text files in a ZIP archive — agency, routes, trips, stops, stop_times, calendar, and calendar_dates are the core, with more than fifteen optional files covering fares, transfers, pathways, and geometry. GTFS Realtime is a separate feed using Protocol Buffers for trip updates, service alerts, and vehicle positions.

Feed quality is a real and unglamorous problem: stale calendars, shape geometry that does not match the road, stop coordinates on the wrong side of a divided highway, Realtime feeds that silently stop updating. And FTA's July 2025 NTD changes added trip planning and geospatial data to the collection, pulling feed hygiene closer to a reporting obligation. A firm that can show a before-and-after validation report on a real agency feed has demonstrated something concrete in a market where most pitches are decks.

How the procurement actually runs

Transit procurements run on the federal grant procurement standards in 2 CFR Part 200 as adopted for DOT, layered with FTA's own third-party contracting guidance. That guidance changed recently and many vendor playbooks have not caught up. FTA issued Circular C 4220.1G on January 16, 2025, superseding C 4220.1F, which had been in place since March 2013. The applicability date was originally February 18, 2025 and was pushed to March 20, 2025 by a subsequent notice. If your capture materials still cite 4220.1F, that is a visible tell.

Two things in 4220.1G matter directly to software and data firms. It added guidance on the protection and types of data rights while explicitly not creating new mandatory contract clauses — FTA maintains the table of required clauses separately from the circular. And when a transit authority asked FTA to expand guidance on software-as-a-service agreements covering data ownership, cybersecurity responsibility, and data portability, FTA declined, saying that detail exceeds the requirements in the federal procurement standards. The translation: nobody is going to hand the agency good SaaS terms. Whoever writes them first sets the baseline, and it should be you, in a form the agency's counsel can accept.

Dollar thresholds moved as well, and they move what kind of competition applies. Under 2 CFR 200.1 the micro-purchase and simplified acquisition thresholds are defined by reference to the Federal Acquisition Regulation. A FAR inflation adjustment effective October 1, 2025 raised the micro-purchase threshold from $10,000 to $15,000 and the simplified acquisition threshold from $250,000 to $350,000. Under 2 CFR 200.320, a micro-purchase can be awarded without competitive quotations if the recipient documents price reasonableness, and recipients may self-certify a higher micro-purchase threshold up to $50,000 annually with appropriate justification. Between the micro-purchase threshold and the simplified acquisition threshold, the agency needs price or rate quotations from an adequate number of qualified sources. Above it, sealed bids or proposals.

That is a wide informal band. A scoped diagnostic, a feed audit, or a bounded pilot under the simplified acquisition threshold can be bought in weeks; the same work at $600,000 is a full competitive solicitation with a months-long clock. Sizing a first engagement to the informal band is not a discount tactic. It is how you reach a working relationship before the large procurement is written.

The same section lists situations considered restrictive of competition: unreasonable qualification requirements, unnecessary experience and excessive bonding, noncompetitive pricing among affiliated firms, noncompetitive retainer contracts, organizational conflicts of interest, brand-name specification without an "or equivalent" allowance, and arbitrary action in the procurement process. Each is a protest ground. If an agency's draft specification names your product, that is not a win — it is an invitation for a competitor to challenge the award.

Sole source is available but narrow. 2 CFR 200.320 permits noncompetitive procurement only where the amount is at or below the micro-purchase threshold, the item comes from a single source, a public emergency will not permit delay, the federal agency approves in writing, or competition is inadequate after solicitation. FTA's guidance in 4220.1G notes that the mere existence of patent and restricted data rights does not by itself justify a noncompetitive award, and that recipients must conduct market research first. Proprietary formats are not a moat here.

The clauses that surprise software vendors

Buy America. 49 U.S.C. 5323(j) conditions FTA assistance on steel, iron, and manufactured goods being produced in the United States, and 49 CFR 661.5 requires that all iron, steel, and manufactured products used in the project be domestically produced. Rolling stock has its own regime: components produced in the United States must exceed 70 percent of the cost of all components for fiscal year 2020 and after, with final assembly in the United States. Waivers exist for public interest, non-availability, and cost increases above 25 percent. The regulation does not speak to software or professional services. Where a data project procures hardware — sensors, on-board units, servers, cameras — Buy America becomes a live question about that hardware, to be priced and sourced up front rather than discovered at delivery.

Covered telecommunications. 2 CFR 200.216 bars a recipient or subrecipient from using federal award funds to procure or obtain, or to extend or renew a contract for, covered telecommunications equipment or services, with the definition drawn from section 889 of Public Law 115-232. It reaches named manufacturers and systems that use covered equipment as a substantial or essential component. If a platform ingests video from an agency's installed camera fleet, or ships with a bundled radio or gateway, the bill of materials must answer this before signature.

Data rights. The agency is spending public money and expects to keep what it paid for. Because 4220.1G raised the profile of data rights without prescribing clauses, the terms are negotiated per contract. Settle four things in writing: who owns the operational data, who owns derived models and features, the export format and cadence at termination, and what happens to the agency's data if the contract lapses. An agency that cannot get its own history back will not renew, and will say so to its peers.

Two areas genuinely in motion

Some of this ground is shifting, and a vendor who pretends otherwise gets caught.

Title VI analysis. FTA recipients have long submitted triennial Title VI programs including service and fare equity analyses and facility siting equity analyses, which the Department estimated at roughly 4,500 annual staff hours of burden. On June 11, 2026, DOT issued a final rule removing several disparate-impact provisions from its Title VI regulations at 49 CFR 21.5, including 21.5(b)(2) in its entirety and the "or effect" language in 21.5(b)(3), implementing Executive Order 14281's stated policy of eliminating disparate-impact liability. The rule anticipates that recipient analysis requirements may be reduced. The direction is clear; where FTA's circular-level guidance lands is not. Do not build a product whose only value is a disparate-impact analysis, and do not tell an agency the obligation has vanished — the statutory prohibition on intentional discrimination is untouched, and agencies still need defensible service and fare data.

Surface cybersecurity. TSA published a proposed rule, "Enhancing Surface Cyber Risk Management," on November 7, 2024, which would impose cyber risk management requirements on certain pipeline and rail owner-operators and a more limited requirement on certain over-the-road bus owner-operators. As of this writing it remains proposed, not final. Agencies are preparing for it and asking vendors about it — a reason to have credible answers on your own security posture, not a reason to cite a requirement that has not been finalized.

What a fundable scope looks like

Put the two halves together — a named grant line and a named obligation — and the proposal writes itself. The strongest transit data scopes share a short set of properties.

• Names the statutory obligation it serves — 49 U.S.C. 5335, 49 CFR 625, 673, or 37, cited in the scope.
• Names the funding line and eligible-activity category — including why the work is a capital project under 49 U.S.C. 5302.
• Says where the local match comes from, or is sized so the match is already budgeted.
• Fits the informal procurement band, or is phased so that phase one does.
• States the data-rights terms up front rather than deferring a negotiation the agency is not staffed to run.
• Separates advisory work from bid positioning so 2 CFR 200.319 does not disqualify you later.
• Delivers something the agency can operate without you — runbook, schema, validation report, export.

Pricing, and the honest version of the sale

Transit agencies are careful buyers with public boards, elected oversight, and a press corps that notices failed technology projects. They have watched fare systems, CAD/AVL replacements, and scheduling platforms arrive late. The reflex against a confident pitch is earned, and the counter to it is not enthusiasm — it is a small, bounded, verifiable first deliverable priced so the agency can approve it without a competitive solicitation, followed by a larger scope the agency writes itself because it now knows what it wants.

Price against the burden the agency already carries. When a reporting duty runs on the order of 156 staff hours per reporter per year, the value of removing half of it is arithmetic the finance office does in its head, against a line it already funds. That beats a percentage improvement in a model.

Bottom line

Transit is a real market for data work, and a more legible one than most of the public sector because the obligations are written down. The failure mode is rarely technical. It is a firm that builds a good system, prices it as a subscription, sells it as an operating expense to an agency with almost no federal operating money, and cannot say which grant line pays for it. Learn the funding structure first, attach the scope to an obligation the agency already owes, size the first engagement so it can be bought quickly, and put the data rights in writing.

Frequently asked questions