13 IMLS FY 2027 grant programs. $80.
72M in planned funding.

Thirteen Institute of Museum and Library Services funding opportunities account for exactly half of the 26 rows labelled new in WebTruffle’s September 2 federal-grants release. All 13 are posted FY 2027 notices, all carry a September 1 posting date, and all point to November 13, 2026 as the closing date. Their source-reported estimates sum to $80.72 million in planned program funding and 590 expected awards.

Those totals describe 13 separate notices. They are not awarded money, federal obligations, an available balance, or a promise that 590 awards will be made. Nor can $80.72M ÷ 590 be treated as an average grant: each program has its own range, selection process, applicant rules, cost-share conditions, and estimate.

This report audits the September 2, 2026 WebTruffle federal-grants release, checks the cohort against all 27 earlier published editions available at the cutoff, compares 117 normalized values with live Grants.gov detail responses on September 4, and uses the official IMLS FY 2027 funding-opportunity index as the application-facing authority.

What the September 2 IMLS grant cohort shows

Six findings survive the artifact, identity, history, and official-source checks:

1. IMLS supplied 13 of the release’s 26 new rows. Each has a distinct source_opportunity_id and a distinct opportunity number. All 13 are normalized as synopsis, posted, and funding instrument Grant.
2. The cohort arrived as one dated group. Every row reports posted_date=2026-09-01, close_date=2026-11-13, and first_seen_at=2026-09-02T14:15:16Z.
3. None appears in earlier WebTruffle history. Neither the 13 source IDs nor the 13 opportunity numbers occurs in any of the 27 earlier published editions from July 22 through September 1. “First observed” is therefore proven across the complete release history available at the cutoff; it is not a claim about when IMLS first conceived a program.
4. The source estimates sum to $80.72 million and 590 expected awards. Museum and museum-workforce titles account for a derived $54.77 million and 310 expected awards; library services and leadership titles account for $25.95 million and 280. The grouping is editorial and title-derived, not a Grants.gov field.
5. The portfolio has two very different scales. Museums for America carries the largest planned pool, $21.699 million, with 122 expected awards and a $5,000–$350,000 range. Native American Library Services Basic carries the largest expected award count, 170, but a $1.7 million pool and a $6,000–$10,000 range.
6. Screening fields are only the start of eligibility review. Eight rows carry a true cost-sharing flag and five false. Eleven use the coarse applicant code 25 (“Others”), while two use code 07 for federally recognized tribal governments. Neither field is rich enough to decide whether a particular organization or project may apply.

The useful headline is therefore not “IMLS awarded $80.72 million.” It is: 13 posted FY 2027 IMLS opportunities first observed in one verified release report $80.72 million in planned program funding, 590 expected awards, and a common November 13 closing date.

Scope, cutoff, and method

The measured dataset cutoff is the September 2 release generated at 2026-09-02T14:15:17Z. It contains 1,586 active normalized records: 1,035 posted and 551 forecasted. The official source extract named in the manifest is GrantsDBExtract20260902v2.zip.

The audit followed seven steps:

1. Download the September 2 manifest and current CSV and independently record each byte count and SHA-256.
2. Require the declared schema and headers, then assert that source_opportunity_id is unique before any aggregation.
3. Select rows where agency_code == "IMLS", change_type == "new", record_type == "synopsis", and opportunity_status == "posted".
4. Require 13 unique source IDs and 13 unique opportunity numbers; then verify the shared posting and closing dates.
5. Search both identifiers across every earlier published CSV available from July 22 through September 1.
6. Sum source-reported funding estimates and expected-award counts only after the uniqueness checks; derive the museum/library grouping explicitly from program titles.
7. On September 4, compare nine fields for each of the 13 source IDs with Grants.gov’s documented fetchOpportunity response: opportunity number, title, posting date, closing date, estimated funding, expected awards, floor, ceiling, and cost-share flag.

The last step produced 117 matches out of 117 comparisons. That is a checked-at source consistency result, not a guarantee that the notices will remain unchanged.

Why the broad updated count is not the story

The release manifest reports 26 new and 1,560 updated rows. A direct September 1/2 set comparison finds 1,559 shared source IDs, 27 IDs present only on September 2, and 13 present only on September 1. The extra incoming record is a reappearance carried as updated, so the exact membership equation is:

1,572 prior rows + 27 incoming IDs − 13 outgoing IDs = 1,586 current rows

Among the 1,559 shared records, only 17 changed any non-observation source field. Three of those changed only last_updated_date; 14 changed at least one other normalized source field; eight moved from forecasted to posted. The other 1,542 shared rows were observation or classification changes rather than source-content revisions.

That is why this analysis filters directly to the 13 IMLS rows labelled new and independently proves their history. A pipeline change label is a review signal, not a count of material opportunity changes.

Pin the September 2 edition before counting

The exact artifacts used here are:

September 2 manifest
bytes     4,626
SHA-256   307750484e97ca8f09c7916a5935a4c226b3aea9ace8b89cd7aabc26d195101f

September 2 funding-opportunities.csv
bytes     1,874,798
SHA-256   1d2ebec4f49de103369327d91e494c29f5e5999cffc542694f330c52e167f5e8

Source XML extract declared by the manifest
file      GrantsDBExtract20260902v2.zip
SHA-256   ba6e2b28d98ac331e9ffac9d260fbb9915ed6970625bfbb920855b3068c1b68b

For the adjacent-edition comparison, the September 1 CSV is 1,861,664 bytes with SHA-256 23e931f66f2afec93760014448e75bd1592b4cbfec6d4cc85bfda6204c79a503; its manifest is 4,626 bytes with SHA-256 b8b04fa2810516ecb360d271c8ced11a662b8111d112cb9497d3c694deb2af36.

The release’s source validation recorded 1,586 XML-derived rows and 1,587 API-search rows: 1,035 versus 1,036 posted and 551 versus 551 forecasted. That one-row difference is inside the configured validation tolerance. It should remain visible; the API search count is an advisory cross-check, not a replacement for the XML-derived edition.

Grants.gov describes its XML extract as a once-daily database export and says the v2 format includes forecasts. (XML extract documentation) A matching digest proves the bytes analyzed. It does not prove that an agency value is correct or that an applicant qualifies.

Why these 13 notices qualify as first observed

The September 2 rows share first_seen_at=2026-09-02T14:15:16Z, but that field alone is not sufficient evidence. A collector can reset, an identity can change, or an older record can return after an absence.

The stronger test uses two independent identifiers:

target source_opportunity_id values   13
target opportunity_number values      13
earlier published editions scanned    27
earlier source-ID matches               0
earlier opportunity-number matches      0

Every accepted WebTruffle edition available before September 2—from the first July 22 release through September 1—was scanned. Because neither identifier appears earlier, the 13 notices are first observed across the complete published dataset history available at the cutoff.

The source’s own posting date provides separate evidence: all 13 say September 1. Still, the precise dataset claim remains “first observed in the September 2 release.” It does not imply that the programs themselves are new, that IMLS had never funded them before, or that September 2 was their official publication date.

The official IMLS FY 2027 index lists the same 13 programs and their current NOFO files. IMLS had also said FY 2027 opportunities were expected to open in September when it began announcing its 2026 library investments. (IMLS library announcement)

What $80.72M and 590 expected awards actually mean

The 13 values in estimated_total_program_funding_usd sum to $80,720,000. The 13 values in expected_number_of_awards sum to 590. These are arithmetic summaries of source-reported planning fields across distinct notices.

They do not mean:

• IMLS has already awarded, obligated, or paid $80.72 million;
• the whole amount is currently available to one applicant or program;
• exactly 590 applicants will be funded;
• every award will fall at the midpoint of its published range; or
• dividing the two totals produces a meaningful average award.

The correct unit of interpretation is the notice. A $1 million ceiling in one program and a $10,000 ceiling in another cannot be blended into one applicant-level promise. The source can also amend funding estimates, award counts, dates, attachments, or conditions after the dataset cutoff.

The title-derived museum group contains eight notices, $54.77 million in planned funding, and 310 expected awards. The library group contains five notices, $25.95 million, and 280 expected awards. The first four programs by planned funding account for $49.699 million, or 61.57% of the cohort total, but concentration of planned funding does not predict applicant competition or award probability.

Compare the 13 FY 2027 programs

The full source-ID map is:

363768  MFA-FY27                 Museums for America
363766  NLG-LIBRARIES-FY27       National Leadership Grants for Libraries
363765  LB21-FY27                Laura Bush 21st Century Librarian
363773  NLG-M-FY27               National Leadership Grants for Museums
363770  AAHC-FY27                Museum Grants for African American History and Culture
363772  ALHC-FY27                Museum Grants for American Latino History and Culture
363776  ME-FY27                  Museums Empowered
363775  NANH-FY27                Native American/Native Hawaiian Museum Services
363762  NAG-ENHANCEMENT-FY27     Native American Library Services Enhancement
363771  IGSM-FY27                Inspire Grants for Small Museums
363769  21MP-FY27                21st Century Museum Professionals
363767  NAG-BASIC-FY27           Native American Library Services Basic
363763  NAG-HAWAIIAN-FY27        Native Hawaiian Library Services

Each ID opens at https://www.grants.gov/search-results-detail/{ID}. For example, use the live Museums for America notice for source ID 363768 and the Native American Library Services Basic notice for 363767. The IMLS index is the cleaner starting point for the complete set.

The largest funding pool is not the largest award count

Two programs show why the portfolio totals cannot be converted into one “average grant.”

Museums for America reports the largest planned pool: $21.699 million, 122 expected awards, a $5,000 floor, and a $350,000 ceiling. It represents 26.88% of the cohort’s planned funding and 20.68% of its expected awards.

Native American Library Services Basic reports the largest expected award count: 170, with $1.7 million in planned funding and a $6,000–$10,000 range. It represents only 2.11% of planned funding but 28.81% of expected awards.

National Leadership Grants for Libraries and Laura Bush 21st Century Librarian each publish a $1 million ceiling, the highest in the cohort. That ceiling is not the expected award, and it does not mean every eligible project can request the maximum. The complete NOFO defines project categories, request limits, match rules, review criteria, and allowable costs.

Treat cost sharing as a review flag

The normalized field cost_sharing_required is true on eight notices and false on five. It is useful for routing:

cost_sharing_required = true     8 notices
cost_sharing_required = false    5 notices

It is not enough for budgeting. A true value does not reveal the match ratio, whether the rule applies to every project type, whether it begins above a request threshold, what counts as an allowable contribution, or whether an exception or waiver exists.

The official National Leadership Grants for Libraries program page, for example, presents project-category and request-specific instructions. That nuance cannot be compressed safely into one Boolean.

Use the flag to decide which NOFO to inspect first. Build the application budget from the current NOFO and package, never by translating true into one universal cost-share rule.

Applicant codes do not decide eligibility

Eleven cohort rows carry applicant code 25, decoded as “Others.” Two—Native American Library Services Basic and Enhancement—carry code 07, “Native American tribal governments (Federally recognized).”

Those labels are facets from the source record. They do not answer whether a museum, library, nonprofit, university, tribal entity, unit of government, consortium, or particular project is legally eligible. They also do not encode every program definition, geography rule, organizational characteristic, registration requirement, partnership condition, or excluded activity.

IMLS tells applicants to read the current NOFO and use the package on Grants.gov; its application guidance is the authority for beginning an application. A safe shortlist records the source ID and opportunity number, then attaches an eligibility decision made from the notice—not from the normalized applicant code alone.

All 13 notices point to November 13

Every row reports close_date=2026-11-13. From this article’s September 4 publication date, that is a 70-calendar-day planning window.

The shared date makes portfolio-level triage easier, but it also creates a resource constraint. A team considering several programs should schedule:

1. Source confirmation now: open the current IMLS page, Grants.gov detail, NOFO, and application package for each candidate.
2. Eligibility review before drafting: document the exact organization, project, request-size, registration, and cost-share rule that applies.
3. Internal selection: compare mission fit, allowable work, evidence requirements, staff capacity, partners, and budget—not the headline pool alone.
4. Submission buffer: set an internal deadline before the official one and assign ownership for registrations, attachments, approvals, and portal submission.
5. Amendment checks: re-open the official notice before final submission because dates and documents can change after the September 2 snapshot.

This article records the date as observed and verified on September 4. It does not replace the source’s exact time-zone, package, or amendment instructions.

Reproduce the IMLS cohort audit

Load the verified CSV with a real CSV parser, preserve identifiers as text, and assert identity before aggregation:

import csv
from collections import Counter
from decimal import Decimal
from pathlib import Path

release_csv = Path("funding-opportunities-2026-09-02.csv")

with release_csv.open(encoding="utf-8-sig", newline="") as handle:
    rows = list(csv.DictReader(handle))

source_ids = [row["source_opportunity_id"] for row in rows]
assert len(source_ids) == len(set(source_ids)) == 1_586

cohort = [
    row for row in rows
    if row["agency_code"] == "IMLS"
    and row["change_type"] == "new"
    and row["record_type"] == "synopsis"
    and row["opportunity_status"] == "posted"
]

assert len(cohort) == 13
assert len({row["source_opportunity_id"] for row in cohort}) == 13
assert len({row["opportunity_number"] for row in cohort}) == 13
assert Counter(row["posted_date"] for row in cohort) == {"2026-09-01": 13}
assert Counter(row["close_date"] for row in cohort) == {"2026-11-13": 13}

planned_funding = sum(
    Decimal(row["estimated_total_program_funding_usd"])
    for row in cohort
)
expected_awards = sum(
    int(row["expected_number_of_awards"])
    for row in cohort
)

assert planned_funding == Decimal("80720000")
assert expected_awards == 590

Then scan every earlier edition using both identifiers:

target_ids = {row["source_opportunity_id"] for row in cohort}
target_numbers = {row["opportunity_number"] for row in cohort}

matches = []
for path in sorted(Path("earlier-editions").glob("*.csv")):
    with path.open(encoding="utf-8-sig", newline="") as handle:
        for row in csv.DictReader(handle):
            if (
                row["source_opportunity_id"] in target_ids
                or row["opportunity_number"] in target_numbers
            ):
                matches.append((path.name, row["source_opportunity_id"]))

assert len(list(Path("earlier-editions").glob("*.csv"))) == 27
assert matches == []

Verify every earlier manifest and CSV before trusting this result. The directory name, release tags, artifact hashes, checked-at date, filter, code version, and output hash belong in the analysis receipt.

Verify the current notice before you apply

The classic Grants.gov detail endpoint is documented as:

POST https://api.grants.gov/v1/api/fetchOpportunity
Content-Type: application/json

{"opportunityId": 363762}

The audit repeated that request for all 13 numeric source IDs and compared nine fields per response. The 117 values matched the normalized rows on September 4. See the official fetchOpportunity documentation.

The API comparison is a consistency check. Application decisions should follow this order:

1. Open the current IMLS funding-opportunity index.
2. Open the specific program page and current NOFO.
3. Open the matching Grants.gov detail and application package.
4. Confirm amendments, eligibility, closing time, request range, cost sharing, registrations, and attachments.
5. Preserve the checked-at time and source URLs with the internal decision.

Grants.gov defines a forecast as a planned opportunity that is not an official posted funding opportunity announcement and may never become one. (Forecast guidance) These 13 rows are posted synopses, but the forecast distinction still matters when building a wider funding pipeline from the dataset.

Choose the dataset, official index, or live detail API

Use the WebTruffle US federal grants dataset when the job needs:

• one normalized active snapshot across agencies;
• posted and forecasted opportunity screening;
• stable source identifiers, arrays, dates, and numeric funding fields;
• CSV, JSON, or JSONL with a manifest and file hashes; or
• reproducible cross-release comparison.

Use the federal grants Python guide for reusable manifest verification, eligibility-label screening, funding-estimate coverage, deadline clocks, and snapshot differences.

Use the official IMLS index and NOFOs when the job is deciding whether and how to apply. Use Grants.gov detail when you need the current source record for a known numeric ID. Use the official daily XML when you need source bulk data or fields outside the normalized public schema.

This dated report owns the September 2/4 empirical findings. The dataset owns current downloads. The Python guide owns the evergreen analysis workflow. IMLS and Grants.gov own application requirements and current source state.

Limitations and interpretation boundaries

This result has explicit limits:

• It measures a September 2 active-opportunity snapshot and checks live detail on September 4; it is not a continuously live grants service.
• “First observed” means absent by source ID and opportunity number from the complete 27-edition WebTruffle history before the cutoff. It does not mean first-ever IMLS program, first agency planning date, or first public discussion.
• new is a dataset-history label. It is not an award, application, obligation, or guarantee that the source record was created that day.
• The $80.72 million sum combines 13 source-reported program estimates. It is neither federal spend nor applicant-level available funding.
• The 590 total combines expected award counts. Expected is not promised, and programs use different ranges and selection rules.
• The museum/library split is a WebTruffle classification derived from program titles, not a source field.
• Cost-share and applicant-type fields are screening aids. They cannot replace the current NOFO’s detailed rules.
• All 13 rows reported November 13 when checked. The official notice can be amended after the dataset cutoff.
• The 117/117 API result compares nine selected fields, not every narrative, attachment, contact, or application-package value.
• A matching file hash proves artifact identity, not agency accuracy, universe completeness, legal eligibility, or program fitness.
• The API search and XML-derived snapshot differed by one posted row at validation time. The release preserves that source check inside its tolerance instead of presenting the two systems as identical.
• Closed and archived Grants.gov records are outside the active daily dataset.

Frequently asked questions