The 2026 Grant Map directory snapshot: 57,347 records associated with 2,225 cities

As of June 17, 2026, the directory held 57,347 records associated with 2,225 cities across all 50 states and Washington DC. The frozen analysis included every city-index entry: 2,203 live and 22 pending cities; the pending cities contributed 88 rows. This report measures that dataset. It does not establish that each record represented a unique, open, funded, locally available program. The public derivation artifact records the pinned commit and data-tree hashes, formulas, exclusions, and tied ranks used here.

Three findings stand out in this snapshot. The median city had 25 listings, 46.9 percent of rows stored the exact income-field sentinel any, and San Diego and Baltimore each had 17 listings with a positive stored numeric amount of $25,000 or more. Those are field-level directory observations, not proof of no income test, an open program, a ceiling, or an amount a homeowner could receive.

The headline numbers

The 57,347 figure counts city-associated listings. Federal and state records repeat across cities by design, sometimes as broad context rather than proof of local eligibility. Deduplicating only by stored program name produced 22,997 names; naming differences can split one program or merge different programs, so this is not a verified count of unique administering programs.

Of the 57,347 listings, 42,415 had a positive numeric maxAmount value. Despite the field name, labels show that these values do not all represent caps: the set also includes averages, ranges, estimates, tax or assessed-value figures, and possible extraction errors. From those 42,415 we set aside 33 mega-scope listings representing 27 normalized names, mostly multimillion-dollar small-business loan records. The full exclusion audit trail is recorded in the versioned frozen-analysis artifact. That leaves 42,382 included numeric-value records behind the distribution below.

• The median stored numeric value is $10,000. The mean is $22,575, pulled upward by a small number of very large tax-credit, disaster, financing, and other values.
• The 25th percentile is $3,000, the 75th is $25,350, and the 95th is $126,526. These are stored-value distribution points, not expected awards or current benefit limits.
• 13,264 included records, 31.3 percent, carried a stored numeric value of $25,000 or more. These include different benefit types and broader rehabilitation, tax, property-value, or financing records.
• Stored values are analyzed as a field distribution only. We do not total them because their semantics differ and replicated listings, loans, credits, and mutually exclusive programs make an aggregate dollar figure misleading.

Here is the full distribution of the included positive numeric values:

The stored-value distribution is concentrated between $10,000 and $50,000, but it mixes ceilings, averages, ranges, estimates, tax and property-value figures, grants, loans, credits, rebates, and broader rehabilitation scopes.

Cities with the most stored values of $25,000 or more

This table counts listings whose positive numeric maxAmount field was $25,000 or more. It does not establish that those values were ceilings, rank likely awards or current funding, or measure homeowner eligibility. Cities are dense-ranked by count; all ties shown here are alphabetical, not broken by dollar sums.

The next count below the displayed tie was 13 in five cities: Denver, Detroit, Greensboro, New York City, and Newark.

The table does not support a causal claim about housing age, city wealth, policy generosity, or funding. Replicated state and federal records, naming, coverage, and source quality can influence a city's count.

We intentionally do not publish city-level sums. A city with one large loan can appear more generous than a city with several smaller grants, and a homeowner may be ineligible for most rows. Compare each program's assistance type, source status, and rules instead.

Income fields: any does not prove no test

In the snapshot, 26,915 of 57,347 records, 46.9 percent, stored the exact incomeRequirement value any. That is a normalized-field count, not proof that the source imposed no income rule or that every income qualifies. The sentinel can reflect broad wording, missing specificity, or an extraction/classification error; confirm the current source rule.

Some programs use area median income while others use poverty guidelines, benefit enrollment, categorical eligibility, or different definitions. Do not self-reject based on an assumed limit, but do not infer eligibility from a blank field either. Confirm the current calculation with the administrator.

The snapshot also includes listings that name seniors or veterans in their rules. Those tags indicate resources to review, not confirmed eligibility or benefits that can be added. Our senior and veteran guides explain what to confirm.

Records tagged for renters

13,077 listings, 22.8 percent, were tagged as renter-inclusive. Of those, 9,152 had a positive numeric maxAmount; the median among those 9,152 values was $1,500. Tags and values can be incomplete, misclassified, or wrong and do not prove that a renter, property, landlord, or project qualifies. Use the renter records as a research filter and verify landlord-consent and tenancy rules.

Federal tax-credit records require a current-law check

Directory records can retain prior-year tax provisions, dates, percentages, and caps after legislation or agency guidance changes. Before including any energy or rehabilitation credit in a 2026 project budget, check current IRS material for the relevant tax year and placed-in-service date, then consult a qualified tax professional for your circumstances.

This report does not determine whether a tax provision currently applies. Percentage-based credits also should not be mixed with grants, loans, and local program ceilings in an aggregate total.

Tax credits, rebates, grants, and loans work differently. Do not infer that another benefit remains available or is "better" because a tax provision changed; verify current law, application timing, upfront costs, and tax treatment separately.

Selected large stored-value examples, with loans labeled as loans

The dataset mixes grants, tax credits, property-value figures, and loans. The following pinned records are editorial examples chosen to show why type and context matter; the structured fields do not support an automated ranking of the largest homeowner grants.

• $12,000,000: New Jersey Historic Property Reinvestment Program. A tax credit against rehabilitation costs on historic properties, not a check, and mostly relevant to large projects.
• $5,000,000: Ohio Historic Preservation Tax Credit. The snapshot record described a credit equal to 25 percent of certain rehabilitation costs. Verify current law and program terms; it is not cash assistance.
• $500,000: likely misclassified Texas disaster-recovery value. The pinned rows stored this number, but the current official GLO pages describe homeowner-rehabilitation caps of $90,000 and $175,000 while the Affordable Rental Program uses $500,000 as the lower end of project awards for developments with at least eight units. We therefore treat the stored value as likely non-homeowner or otherwise misclassified, not as homeowner assistance.
• $126,526: VA Specially Adapted Housing grant. The pinned record described a grant ceiling at this value and appeared across many covered cities, subject to the VA's detailed disability and project rules.

These examples show the narrower defensible point: a widely replicated $126,526 SAH value appeared in the pinned snapshot, while the $500,000 Texas value exposes a probable classification defect. Neither proves an automated ranking, current awards, or household totals.

Below those two tax credits, the next tier of big numbers in the directory is loans, and we label them as loans. These are historical values preserved in the frozen snapshot, not current limits. HUD's 2026 FHA announcement and FHFA's 2026 conforming-limit announcement supersede older annual figures where applicable.

Loan caps vary by year and county, and several products appear under more than one stored name. Read the table only as evidence of what the frozen directory stored, not as a ranking or current quote. Repayable financing is included for comparison, but it is not a grant or a universal fallback. Our grants versus loans guide explains the legal distinctions and questions to ask; no application order works for every household.

What the data says to do

Review your city's cataloged records, treat missing fields as unknown, and confirm status directly with each administrator. The current Grant Map homeowner fee is $0; confirm all administrator, inspection, financing, contractor, and project costs.

Methodology and how to cite this report

All figures were computed on June 17, 2026 from 57,347 directory rows associated with 2,225 cities. Dollar distributions use the positive numeric maxAmount field; the versioned frozen-analysis artifact records the exact 27-name, 33-row exclusion trail. The field does not have one consistent economic meaning. Eligibility shares use structured fields that can be stale or extracted incorrectly. Arithmetic consistency does not validate the underlying claims. See the methodology and audit limitations.

Journalists, researchers, and bloggers may cite figures with attribution, a link, the snapshot date, and the stated limitations. For custom cuts, contact us; no response time is guaranteed.

Common questions

Does 57,347 records mean 57,347 different programs?

No. It means 57,347 city-associated records, including replicated federal and state entries. Deduplicating by stored name produced 22,997 names, not a verified count of unique programs.

Is all of this free money?

No. The directory holds grants, rebates, forgivable loans, tax credits, property-value figures, and repayable loans. In this dated snapshot, the median included positive numeric value was $10,000; that statistic is not a median award or even necessarily a median ceiling.

Does this table identify the most generous city?

No. It identifies cities with more stored numeric values of $25,000 or more in the June 2026 snapshot. It does not establish that each value is a ceiling or measure current funding, grant-only dollars, approval rates, or what one household can receive.

Can I republish these numbers?

Yes. Cite The Grant Map and link to this page. The figures reflect a June 17, 2026 snapshot, so for anything time-sensitive, check the date and ask us for a refresh if you need one.

Counts and figures reflect our directory data as of June 17, 2026. Confirm current terms, income limits, funding, and application route directly with the administrator.