US Small Business Funding Statistics 2026

What the 736-program catalog counts, and what it deliberately leaves out

The GrantCompass US catalog is a curated register of 736 named, individually verified funding programs that a US small business can apply to — not a census of American business funding. A program earns a row only when it has an identifiable administering body, a public description of what it funds, and an application route a business can actually use. That definition excludes bank credit, venture capital, personal savings, revenue-based finance and every informal source, all of which move vastly more money than the 736 programs here.

Scale matters when reading any figure below. The SBA Office of Advocacy counts 36,207,130 small businesses in the United States (Frequently Asked Questions About Small Business, February 2026). Of those, 82.3% have no employees at all, a population almost no program in this catalog is built for. Against that denominator, 736 named programs is a small, specific universe: this page describes the shape of the discoverable public and philanthropic funding layer, not the shape of small business finance. Field coverage decides which denominator each figure below uses: 545 of the 736 programs (74.0%) publish a maximum award, 730 (99.2%) carry an application-effort estimate, 714 (97.0%) record whether a match is required, and 695 (94.4%) carry a competition score. Where an external source is quoted, it is named with its publisher, year and URL, and kept separate from GrantCompass figures.

410 of 736 US small business funding programs are grants — 326 are something else

Grant-type programs make up 55.7% of the catalog (410 of 736), which means 44.3% of what gets marketed as "small business funding" is a different instrument with different money, different timing and different obligations. Tax credits account for 108 programs (14.7%), loans for 105 (14.3%), non-cash support programs for 87 (11.8%), awards and contests for 23 (3.1%), and forgivable loans for 3 (0.4%). A business that searches only for the word "grant" is filtering out 326 tracked programs before it starts.

The instruments do not behave alike. Loans are the most persistently available — 95 of 105 loan programs (90.5%) were open on August 28, 2026, against only 174 of 410 grants (42.4%), because lending windows rarely close while grant cycles do. Grants also carry the lowest typical ceiling of the major instruments, at a $67,500 median, while loans median $500,000.

Median ceiling is computed only over the programs of that instrument which publish a maximum award (grants 374, loans 91, tax credits 28, support programs 27, awards 22, forgivable loans 3). Median hours uses each program's published application-effort estimate.

Grant, tax credit, loan and support program: what actually separates the four

A grant — 55.7% of this catalog — is money awarded against a competitive or eligibility test and is never repaid, though most grants reimburse spending after it happens rather than paying in advance. A tax credit reduces a tax bill rather than sending cash, so it is worth $0 in a year with no tax liability and everything to a profitable business — 108 of the 736 tracked programs (14.7%) work this way, including the federal R&D tax credit. A loan is repaid with interest; the catalog's 105 loan programs (14.3%) exist because subsidised credit, such as SBA 7(a) lending, is often the only instrument available at scale. A support program delivers advisory hours, training, equipment access or subsidised services instead of cash, and 87 of the 736 programs (11.8%) took this form on August 28, 2026.

The practical sequence differs too. A grant application is written before any money moves and judged once. A tax credit is claimed on a filed return, usually after the qualifying spend, and is checked in audit rather than in competition. A loan is underwritten on repayment capacity. Sorting programs by instrument before comparing dollar figures is the single step that stops a $500,000 loan ceiling from being read as a $500,000 grant.

US states administer 310 small business programs; the federal government administers 198

State governments are the largest single administrator of US small business funding programs by count: 310 of 736 programs (42.1%) are state-run, against 198 federal programs (26.9%). Government at those two levels accounts for 508 programs (69.0%). Private corporate programs contribute 147 (20.0%), municipal and county programs 44 (6.0%), and private foundations 37 (5.0%). Put together, as computed on August 28, 2026, 538 of the 736 programs (73.1%) are administered by someone other than the federal government, which is why a search confined to Grants.gov and federal small business grants misses roughly three programs in four. The structural differences are set out in our federal vs state grants guide.

Federal programs are the biggest and slowest; foundation programs are the smallest and fastest

Administering level predicts both award size and workload more sharply than any other field in the catalog. Federal programs carry a median published ceiling of $1,000,000 and a median application-effort estimate of 40 hours. State programs median $150,000 and 15 hours. Corporate programs median $25,000 and 5 hours, municipal programs $25,000 and 8 hours, and foundation programs $15,000 and 4 hours — a foundation application is typically about one-tenth of the work of a federal one, for about one sixty-sixth of the ceiling.

The right-hand column below divides each level's median ceiling by its median hours. It is a workload ratio, not an expected value: it says nothing about the chance of winning, only about how much headline money sits behind an hour of paperwork. Read alongside the odds — which we treat separately in small business grant approval rates — it explains why federal applications remain rational for R&D-heavy firms and irrational for a two-person retailer.

416 of the 736 tracked programs were open to applications on August 28, 2026

Slightly more than half the catalog is live at any moment: 416 programs (56.5%) were open on August 28, 2026. A program counts as open when its recorded status is active and it either accepts applications year-round or carries a published deadline that has not passed. The remaining 320 are not all dead — 160 (21.7% of the catalog) are between intakes and expected to reopen and 15 (2.0%) are announced but not yet accepting, while only 40 programs (5.4%) are recorded as discontinued, closed, expired, paused or winding down. Distinguishing "shut for now" from "gone" is the whole difference between a list worth watching and a list worth deleting.

Openness is not evenly spread. Only 174 of 410 grant programs (42.4%) were open, against 95 of 105 loan programs (90.5%) and 68 of 87 support programs (78.2%). By level, 200 of the 310 state programs were open versus 104 of 198 federal ones. The current live list is maintained at small business grants open now.

378 programs accept applications year-round; only 52 open programs have a dated deadline

378 of the 736 tracked programs (51.4%) describe a rolling, year-round, continuous or ongoing intake with no fixed cut-off date. That single fact reframes deadline anxiety: of the 416 programs open on August 28, 2026, 364 (87.5%) were rolling and only 52 (12.5%) had a specific published closing date. Of those 52, 20 closed within 30 days (4.8% of open programs), 37 within 60 days (8.9%) and 42 within 90 days (10.1%) — so at any moment only 42 programs in the entire catalog, 5.7% of it, are genuinely time-pressured.

Rolling intakes cluster in lending and in state training funds, where money is appropriated and disbursed until it runs out rather than judged in a cycle. That has a practical consequence a deadline calendar hides: a first-come rolling program can stop accepting applications with no announcement when its allocation is exhausted, so "rolling" means no fixed date, not permanently available. The rolling subset is listed at grants with no deadline.

The median US small business funding program caps its award at $150,000

Among the 545 programs that publish a maximum award, the median ceiling is $150,000. That is the single most useful number on this page, and it is very different from the mean, which is $6,346,642. The published ceilings run from $500 — the Etsy Seller Relief Fund, the Giving Joy Microgrants for Women Entrepreneurs and Colorado's Green Business Network equipment grant all cap there — to $500,000,000 at the Department of Energy's Grid Resilience and Innovation Partnerships program and the Treasury CDFI Bond Guarantee Program. That is a span of one million to one between the largest published ceiling and the smallest.

Three shares define the practical middle. 98 of the 545 programs (18.0%) cap at $10,000 or less, 52 (9.5%) cap at $5,000 or less, and 127 (23.3%) cap at $1,000,000 or more. The most winnable end is documented at microgrants under $10,000 and the top end at the biggest grants a small business can win. For the average rather than the ceiling, see average small business grant amount.

The full award-ceiling decile table, for all instruments and for grants alone

Deciles describe a skewed distribution far better than any average. The table below cuts the 545 published ceilings into ten equal groups, and repeats the exercise for the 374 grant-type programs that publish a ceiling. Half of all published ceilings fall between $15,000 and $500,000; half of grant ceilings fall between $10,000 and $300,000. The grant column is the one most readers actually want, and it sits roughly one band lower than the all-instrument column at every point.

Deciles computed with Python's default exclusive quantile method over the sorted list of published maximum awards. Reproduce with statistics.quantiles(ceilings, n=10).

Why the mean award ceiling is 42 times the median

The mean of the 545 published award ceilings is $6,346,642 — 42 times the $150,000 median. That gap is not a rounding artefact; it is the defining feature of this distribution, and it is why almost every "average small business grant" figure circulating online is misleading. Only 36 of the 545 published ceilings (6.6%) are at or above the mean. An average that 93.4% of programs fall below is not describing a typical program.

The concentration is extreme at the very top. The 545 published ceilings sum to $3.46 billion of headline capacity, and the five largest programs alone account for 62.4% of that total; the largest ten account for 74.9% and the largest 27 for 88.5%. Four programs cap at $500,000,000 each: three Department of Energy lines — Grid Resilience and Innovation Partnerships, the Industrial Demonstrations Program, and the clean energy demonstrations line we now record as discontinued — plus the Treasury CDFI Bond Guarantee Program. All four fund infrastructure-scale consortia rather than the businesses reading this page. Removing just those four drops the arithmetic mean from $6,346,642 to $2,696,709.

The $3.46 billion figure is a sum of published maximum awards, not a budget: a ceiling is the most one recipient could receive, not money committed or available.

Grant ceilings are much lower than the all-instrument figure: a $67,500 median

Filtering to grant-type programs alone moves the median ceiling from $150,000 down to $67,500 across the 374 grants that publish a maximum. Loans pull the all-instrument figure upward, with a $500,000 median ceiling across 91 loan programs, as do tax credits at $387,500 across the 28 that publish a cap. The honest headline for a founder searching the word "grant" is therefore closer to $67,500 than to $150,000, and closer still to $50,000 once availability is imposed: the 143 currently open grants that publish a ceiling have a $50,000 median.

The gap between those three numbers — $150,000 all-instrument, $67,500 grants, $50,000 open grants — is a good test of any funding statistic you read elsewhere. A page quoting a six-figure "average grant" without stating its instrument mix, its denominator and whether closed programs are included is quoting an artefact. Ours are stated: 736 tracked programs, 545 with published ceilings, computed August 28, 2026.

A typical US small business funding application takes 12 hours of work

The median published application-effort estimate across the 730 programs that carry one is 12 hours. The mean is 30.4 hours, skewed by federal solicitations; the range runs from 15 minutes to 800 hours, and 730 of the 736 programs (99.2%) carry an estimate. This is an original statistic — no government directory publishes application-burden estimates program by program — and it is deliberately conservative: it counts drafting, assembling and submitting, not the time spent finding the program or the post-award reporting.

The distribution is bimodal in practice. 282 of the 730 programs (38.6%) estimate 8 hours or less — a single working day — while 132 (18.1%) estimate 4 hours or less, 117 (16.0%) estimate more than 40 hours — a week of full-time work or a consultant engagement — and 56 (7.7%) estimate more than 80. Grant-type programs median 12 hours; support programs are the lightest at 8; loans and tax credits both median 15.

Application-effort deciles across 730 programs

The effort table below cuts the 730 published estimates — 99.2% of the catalog as of August 28, 2026 — into ten equal groups, alongside the 408 grant-type programs that carry an estimate. The bottom 30% of programs estimate 8 hours or less; the top 10% estimate 75 hours or more. The middle 60%, from the 20th to the 80th percentile, spans 5 to 36 hours, and the 10th percentile of 3 hours is 4.0% of the 90th percentile's 75. The grant column is lighter at the bottom and heavier at the top: its 90th percentile of 100 hours is 33% above the 75-hour all-program figure, because grants include both the microgrant forms that take an afternoon and the federal research solicitations that take weeks.

Application burden is the one cost nobody publishes, and the federal government is the largest source of it

Federal programs are where the hours go. As of August 28, 2026 the 196 federal programs carrying an estimate median 40 hours, against 4 hours for foundation programs — 10.0% of the federal figure — and 5 for corporate ones. The concentration at the top is near-total: 68 of the 73 programs estimating 75 hours or more (93.2%) are federal, though federal programs are only 26.9% of the catalog. A founder who applies to eight corporate microgrants at 5 hours each spends about the same 40 hours as one federal solicitation, and gets eight independent decisions instead of one. That is a real strategic choice the ceiling figures alone conceal, and it is the reason the easiest grants to get are worth ranking separately from the largest.

The federal government measures this burden in aggregate and says so plainly. The SBA Office of Advocacy reports: "Federal paperwork collections cost small businesses over $81 billion in 2025. Over 80% of the paperwork burden for small businesses comes from the Internal Revenue Service alone" (Frequently Asked Questions About Small Business, SBA Office of Advocacy, February 2026). That estimate covers all federal compliance paperwork rather than funding applications specifically, but it establishes the order of magnitude: application time is a genuine cost, and it is not published program by program by anyone except us.

141 of 714 US small business funding programs require matching funds

Matching requirements are the barrier founders most often assume is universal, and it is not. 141 of the 714 programs that record the field require the applicant to contribute matching funds — 19.7%, as of August 28, 2026. The other 573 (80.3%) do not. A match means the program pays a defined share of a project and the business supplies the rest in cash or, in some programs, in documented in-kind contribution, so a 50% match on a $100,000 project means $50,000 from the program and $50,000 from the business.

Matching clusters where you would expect: state manufacturing and workforce funds and federal cost-shared research, where the requirement exists to prove commitment rather than to exclude. It is rare in corporate and foundation microgrants. Because the field is recorded for 714 of the 736 programs, the 19.7% figure should be read as "of programs where we know", not "of all programs" — 22 programs have no recorded value either way.

441 of 711 programs are explicitly open to first-time applicants

441 of the 711 programs that record the field — 62.0% — are marked first-time-applicant friendly, meaning the program neither requires a prior award, an existing relationship with the agency, nor a track record of federal grant management as a condition of applying. The remaining 270 (38.0%) carry at least one such expectation, usually in the form of prior-award history, an incumbent registration, or a demonstrated management capacity that a first application cannot show.

Two ordinary requirements are separate from this and apply widely: an entity registration such as SAM.gov for federal awards, and a formed legal entity with a business bank account for almost everything else. Neither is an insider gate; both are administrative steps a first-time applicant can complete before applying. Programs combining all three of no match, a first-time-friendly posture and a rolling intake number 212 of 736 (28.8%), and 207 of those (28.1%) were open on August 28, 2026 — 48 of them grant-type. That intersection is the practical starting set, ranked at easiest small business grants.

The catalog covers 53 states and territories, and 322 programs are national

Every US state is represented, plus the District of Columbia, Puerto Rico and Guam — 53 jurisdictions in total. Separately, 322 of the 736 programs (43.8%) are national in scope, open to a qualifying business anywhere in the United States. Those two facts combine in the way that matters: a business in California sees the 33 California-specific programs — 4.5% of the catalog — plus the 322 national ones, for a working universe of 355; a business in Wyoming sees 5 plus 322, for 327. The state-specific layer runs from 33 programs down to 1; the national layer of 322 does not vary at all, so the thinnest jurisdiction still reaches 323 programs — 91.0% of California's 355.

Median state-specific count across the 53 jurisdictions is 12 programs, or 1.6% of the catalog. The thinnest coverage is Guam (1 program, 0.1%) and Puerto Rico (2, 0.3%); the thinnest states are Alaska, Idaho, Kansas, New Hampshire and Wyoming at 5 each (0.7%). Thin coverage reflects both smaller state programs and our own research depth — see "Where this count could be wrong" below.

Manufacturing has the widest program coverage: 373 of 736 programs

Programs usually serve several industries at once, so these counts show how many of the 736 tracked programs a business in each sector could be eligible for on industry grounds alone. As of August 28, 2026, manufacturing leads at 373 programs (50.7%), followed by technology at 346 (47.0%) and agriculture at 250 (34.0%). A further 231 programs (31.4%) are industry-agnostic, accepting any sector, while retail sits at 140 programs (19.0%), or 37.5% of manufacturing's 373. The ranking reflects where policy money is directed rather than where US small businesses actually are: state industrial and workforce funds concentrate on manufacturing, which is why it heads a list of programs rather than a list of firms.

Most US small business funding programs score low-to-moderate on competition

GrantCompass scores each program 1 (least competitive) to 5 (most competitive) on how contested it is, based on applicant volume signals, award counts and eligibility breadth. Of the 695 programs carrying a clean 1–5 score on August 28, 2026, 484 (69.6%) sit at 1–3, and only 78 (11.2%) sit at the most-competitive 5. The famous, heavily-marketed programs cluster at 4 and 5 and dominate what founders see, which is why competition is the figure most consistently overestimated.

This score is a structural estimate of contestedness, not a published approval rate, and the two are different things: a program can be uncontested and still reject most applicants for eligibility failures. Published approval rates — where they exist at all — are collected separately at small business grant approval rates. The 41 catalog rows (5.6%) carrying a non-conforming competition value are excluded from this table rather than coerced into the scale.

18.9% of US small business funding programs are demographic set-asides

Of the 665-program GrantCompass catalog as it stood on July 5, 2026, 126 programs — 18.9% — restrict or explicitly prioritize eligibility by owner demographic. Women-owned businesses have the largest set-aside share: 74 programs, or 11.1% of that catalog, specifically target women founders. Minority-owned businesses are targeted by 68 programs (10.2%), and veteran-owned businesses by 20 programs (3.0%). The remaining 81.1% of programs are open to any qualifying business regardless of who owns it. This is the only figure on the page computed against the older 665-program catalog rather than the current 736, which is 71 programs (10.7%) larger; it has not been recomputed for the August 2026 catalog, and it is dated accordingly wherever it appears.

Set-aside rates across all eight demographic categories

Two caveats make these GrantCompass Set-Aside Rates conservative. First, categories overlap: a program for "women of color" counts in both the women-owned and minority-owned rows, so the eight rows sum to more than 126. Second, the counts include only programs where the demographic is a genuine eligibility gate or named priority — programs that merely mention a group in marketing copy are excluded. No government directory publishes this breakdown; these figures are computed directly from the verified GrantCompass catalog of 665 US programs, as of July 5, 2026.

Transcript of the Set-Aside Rate video

How these figures sit against SBA, Census and Federal Reserve data

GrantCompass figures describe programs; federal statistics describe businesses. Kept apart, they answer different halves of the question. The SBA Office of Advocacy counts 36,207,130 US small businesses, of which 29,811,495 (82.3%) have no employees and 6,395,635 (17.7%) have paid staff; small businesses are 99.9% of all US firms, employ 62.3 million people (45.9% of private-sector workers), pay 38.7% of private-sector payroll ($3.5 trillion), earn 35.0% of private-sector receipts ($17.8 trillion) and produce 43.5% of GDP (Frequently Asked Questions About Small Business, February 2026, advocacy.sba.gov, read August 27, 2026).

On the demand side, the Federal Reserve's Small Business Credit Survey reports that "Sixty percent of firms applied for financing in the 12 months leading up to the survey" and that 86% use financing regularly, most often credit cards and loans (2026 Report on Employer Firms: Findings from the 2025 Small Business Credit Survey, fielded September 3 to November 14, 2025, 6,525 responses, fedsmallbusiness.org, read August 27, 2026). Set against 736 named programs, the arithmetic is stark: demand for external finance is measured in millions of firms, and the named-program layer is measured in hundreds.

What the federal government publishes about funding programs, and what it does not

Federal sources publish program terms well and program-level burden badly. SBIR is the clearest case of the former: "As of April 2026, agencies may issue a Phase I award (including modifications) up to $323,090 and a Phase II award (including modifications) up to $2,153,927 without seeking SBA approval" (sbir.gov/about, read August 27, 2026). The same source records that SBIR/STTR is coordinated by the SBA and funded through 11 participating federal agencies. Those published ceilings sit close to our own federal median of $1,000,000. The program mechanics are covered at SBIR grants for startups.

What no federal source publishes is a cross-agency register of small business funding programs with comparable fields — instrument, ceiling, status, effort — spanning federal, state, municipal, corporate and foundation money in one place. Grants.gov lists federal opportunities only — the layer that is 26.9% of this catalog — and includes many that no small business can receive. That absence is the reason this page exists, and it is also the reason its figures cannot be checked against an official equivalent: there is not one. The correct posture toward our numbers is therefore to read the method, not to assume the authority.

What this distribution means if you are looking for funding

The shape of the 736-program distribution, rather than any single headline, is what should change a search. It is wide at the small end and thin at the large one: 98 of 545 published ceilings (18.0%) cap at $10,000 or less while 127 (23.3%) cap at $1,000,000 or more, and almost nothing in the second group is reachable by a business without a research consortium behind it. The grant-only median ceiling of $67,500, and $50,000 among grants open today, is the number a founder should plan against. It is administered mostly outside Washington, with 538 of 736 programs (73.1%) run by states, cities, corporations and foundations. And it is far less deadline-bound than a calendar of closing dates implies, with 378 programs (51.4%) on rolling intake.

How every figure on this page was computed

The universe is the GrantCompass US catalog: 736 named US small business funding programs, each researched and recorded by hand with a source URL, an administering body, an instrument type, a status, and where the program publishes them, a maximum award and an estimated application effort. All figures on this page were computed from that catalog on August 28, 2026, except the demographic set-aside rates, which are dated July 5, 2026 against the then-665-program catalog and are labelled as such.

Specific rules, so any figure can be reproduced or challenged. Counts and shares use 736 as the denominator unless a narrower one is named in the sentence or the table caption, so 416 open programs is 56.5% and 378 rolling programs is 51.4% of that same base. "Open" means the program's recorded status is active and it either describes a rolling, year-round, continuous or ongoing intake or carries a published deadline dated on or after the computation date; this is a status flag, not a guarantee that funds remain. Award ceilings use each program's published maximum award; the 191 programs (26.0%) that publish no maximum are excluded from every ceiling statistic rather than imputed, which is why ceiling figures carry n=545 (74.0% of the catalog). Deciles use Python's default exclusive quantile method over the sorted values. Application effort uses each program's published hour estimate, covering 730 of 736 programs (99.2%). Medians, not means, are used for every headline because the distribution is severely right-skewed.

What is excluded from every figure on this page

Five things are outside the universe by design, and no figure above accounts for them. Bank and non-bank commercial credit is excluded except where a specific subsidised program exists, so ordinary term loans, lines of credit and credit cards are absent even though the Federal Reserve finds 86% of small employer firms use financing regularly and names those products the most common. Venture capital and angel investment are excluded entirely: they are equity, not funding programs. Personal and informal capital — savings, family, friends, retained earnings — is excluded. Procurement is excluded: a federal contract is payment for goods delivered, not a funding program, although it is the largest small business money flow of all — small businesses took 28.8% of federal contracting dollars in fiscal year 2024, against a 23% government-wide goal (SBA Office of Advocacy, Frequently Asked Questions About Small Business, February 2026, citing the SBA FY 2024 Small Business Reporting Scorecard). And one-off local awards with no published terms are excluded, because a program with no describable eligibility cannot be recorded consistently.

Two further exclusions are worth stating because they change the numbers reported for August 28, 2026. The seven-field paid assessment layer GrantCompass sells per program is never used in any aggregate on this page; every statistic here derives from public catalog fields only. And the 40 programs (5.4%) recorded as discontinued, closed, expired, paused or winding down remain in the 736 count — excluded from the 416 open figure but not from the catalog, because a dead program that still ranks in search results is worth recording as dead.

Where this count could be wrong

Four honest weaknesses. Coverage is curated, not exhaustive: 736 programs as of August 28, 2026 is what we have verified, not what exists, and thin state rows — Guam at 1 program, Puerto Rico at 2, five states at 5 — probably reflect our research depth as much as those jurisdictions' actual program counts. Ceiling coverage is 74.0%: 191 programs (26.0%) publish no maximum, and if they skew small, the true median is below $150,000. Application-effort estimates are ours, not the agencies' — no program publishes its own burden figure, so the 12-hour median covering 99.2% of the catalog is a researched estimate with real error bars, conservative by construction because it excludes discovery and reporting time.

Status drifts daily. The 416 open figure was 417 on August 27 and would be 414 on August 31 on identical data, purely because dated deadlines pass. Treat every open/closed figure on this page as a snapshot with a date attached, never as a standing fact. Separately, 41 of the 736 rows (5.6%) carry a competition value outside the intended 1–5 integer scale and are excluded from the competition table rather than coerced; that is a known data-quality defect in our own catalog, and it is being corrected. If you find a figure here that contradicts a program's own published terms, write to hello@grantcompass.co and we will check it and correct the catalog.

Cite or reuse these statistics

Every figure above may be quoted with attribution. An open extract of 484 of the 736 programs (65.8%) is published under CC BY 4.0 as CSV and JSON, carrying the instrument, level, state, industry, status, competition, ceiling and application-hours fields used here, so the method can be re-run independently. The extract is a subset and its figures differ from the full catalog in a predictable direction: its median ceiling is $250,000 against $150,000 for all 736, because the 252 catalog programs the extract does not include are smaller on average, with a $50,000 median ceiling.