USCIS does not publish an NIW approval rate. It publishes counts, and the one defensible figure from them is the approval share of decisions, meaning approvals divided by approvals plus denials in the same reporting period. That share has fallen every year of the series, from 95.7% in FY2022 to 45.5% in the first half of FY2026. It is not your odds.
USCIS publishes counts, not rates. Its Form I-140 data series is the only public source that separates national interest waivers from the rest of EB-2, and it begins with fiscal year 2022. From those counts one defensible figure can be computed — the approval share of decisions, meaning approvals divided by approvals plus denials in the same reporting period. That share has declined every year of the series: 95.7% in FY2022, then 79.6%, 71.0%, 55.2%, and 45.5% in the first half of FY2026. What it is not is your odds. It measures decisions issued in a period, most of them on petitions filed earlier.
You have seen an approval-rate figure quoted somewhere and want to know where it came from. That is what this page is about: the published data, what can be computed from it, and what it will not support. If your question is whether you qualify, the requirements pages cover the three-part test itself.
There is no official "NIW approval rate." USCIS publishes counts — received, approved, denied, pending — in a quarterly report titled "Form I-140 by Fiscal Year, Quarter and Case Status," produced by its Office of Performance and Quality. That report is the only USCIS publication breaking the national interest waiver out of EB-2 as its own line, E21-NIW.
The series begins with fiscal year 2022. The predecessor report folded these petitions into the general E21 line under a footnote reading "Includes National Interest Waiver petitions," so no USCIS-published NIW-specific count exists for FY2021 or earlier. Any figure you see for those years is an estimate or a different classification wearing the wrong label.
Everything below comes from that series, pinned to each fiscal year's fourth-quarter (full-year) edition — except FY2026, which is the second-quarter edition covering October 1, 2025 through March 31, 2026 only. All figures are as of August 2026.
"Approval share of decisions" below is our own computation and not a USCIS statistic: approvals divided by approvals plus denials, within the same reporting period. We state the formula because different sources divide different things and reach very different answers.
The FY2026 line is half a year. USCIS has published quarters one and two; there is no full-year figure yet, and doubling a half-year is not a forecast, it is arithmetic dressed up as one.
We will update this page when USCIS publishes the FY2026 fourth-quarter edition. Until then, treat that row as a partial-period observation and nothing more.
The report counts actions taken during the reporting period. The approvals and denials issued in a given year are largely decisions on petitions filed in earlier periods — USCIS's own notes caution that some applications approved, denied, or pending a decision "may have been received in previous reporting periods." The approvals in one column and the receipts in another are not the same petitions.
The same file makes the point vividly. On the labor-certification-based E21 line in FY2025, USCIS recorded 52,795 approvals against 49,921 receipts — more approvals than filings that year, because it was working down a backlog. A period ratio can exceed anything a filing cohort could produce. It is a throughput measurement, not an outcome probability.
Two further limits. The denominator includes only decisions: withdrawals, administrative closures, and transfers are absent, and USCIS says so — "Some actions taken on cases are not reflected in this report such as administrative closures and transfers between offices." And the counts are vintage-dependent, since "[c]ounts may differ from previous periods due to system updates and post adjudicative outcomes."
One more calculation to watch for. Articles reporting an NIW approval rate in the low forties for FY2024 are dividing approvals by receipts — 27,526 by 63,549 — treating every unadjudicated case in a growing backlog as though it had been denied. Our approval share of decisions for that same year is 71.0%. Neither is your odds, but if a page quotes a rate without saying what was divided by what, you cannot tell which of two calculations nearly thirty points apart you are reading.
Two things are unambiguous. Filings surged: receipts roughly doubled from 21,973 in FY2022 to 39,803 in FY2023, rose about sixty percent again to 63,549 in FY2024, and stayed elevated at 66,276 in FY2025. Over that span the national interest waiver went from about a quarter of all EB-2 I-140 filings to well over half — 21,973 of 84,472 EB-2 receipts in FY2022 against 66,276 of 116,197 in FY2025, shares we compute from the same series. And the backlog grew with it: pending petitions went from 18,599 on September 30, 2022 to 89,215 on March 31, 2026.
The approval share of decisions declined in every published year. What the data does not do is explain why. It has no field for RFE issuance, no adjudicator notes, and no way to separate a change in what officers require from a change in who is filing during a period when filings tripled.
One framing to avoid: FY2022's 95.7% is not a historical norm that later years "fell from." It is the first year of the breakout series, with only 468 denials recorded and a docket that predates the surge. It is where the series begins, and nothing more than that.
On January 15, 2025, USCIS issued Policy Alert PA-2025-03, "Second Preference Eligibility for National Interest Waiver Petitions," revising its guidance at 6 USCIS-PM F.5, effective immediately and applicable to pending as well as future filings.
Factually, what it says: a petitioner must establish the underlying EB-2 classification before the waiver is considered; a claim of exceptional ability must relate to the proposed endeavor; whether an occupation is a profession, and whether post-baccalaureate experience is "in the specialty," are assessed case by case. It adds guidance on national importance, support letters, and business plans, and states that not every entrepreneur qualifies and that broad assertions regarding general benefits to the economy and potential to create jobs will not establish eligibility.
We describe that separately from the numbers on purpose. The decline in approval share began in FY2024, a year before the alert issued. Anyone telling you the January 2025 guidance caused the numbers to move is asserting something the published data cannot support, in either direction. Read it because it tells you what officers are instructed to look for — not as an explanation of a trend line.
You will not find a Locke Immigration Law approval percentage anywhere on this site. That is deliberate, and we would rather explain it than let the absence look like modesty.
A firm-published approval rate is a number with no auditor. The firm chooses the denominator: cases it declined to take are not in it, and cases withdrawn, refiled after a denial, or still pending after two years can be counted or not at its discretion. Every one of those choices moves the number, none is visible to you, and no regulator checks any of it.
It also measures the wrong thing. A firm that accepts only petitions it considers near-certain will report a higher rate than one that takes harder cases and tells clients exactly what they are looking at — and the second firm may be doing better work. Selection drives that statistic far more than skill does.
And a percentage from someone else's cases tells you nothing about yours. What is useful before you spend money is a real read on your own record: which prong is weak, what would strengthen it, and whether filing now is the right call.
Schedule a consultation with Loren Locke to see if this visa is the right fit.
Schedule a ConsultationThe published data describes an environment. It tells us the queue is long and that denials are a far larger share of decisions than in 2022 — which changes how much we insist on before filing, and tells us nothing about how a specific record will be received.
Before filing we test the record against the guidance an officer is actually working from, including the January 2025 Policy Manual revisions. Where an argument can be answered with a sentence from that guidance, it gets rewritten or dropped.
If a case is not ready, you hear it while there is still time to fix it — well before a Request for Evidence arrives. Petition work is a flat fee, with any RFE response quoted separately.
The EB-2 threshold and the three-part waiver test — one page per question.
ExploreHow physicians, engineers, researchers, founders, and others build an EB-2 NIW case.
ExploreThe two physician waivers — the three-prong test and the statutory five-year shortage-area route — plus the J-1 layer.
ExploreReal situations: an RFE on a self-filed petition, a denial, a pending PERM, a later EB-1A.
ExploreEB-2 NIW measured against EB-1A, employer-sponsored EB-2, EB-1B, and the O-1A.
ExploreAggregate statistics describe conditions. Where your own petition is strong, where it is thin, and whether now is the right time to file are questions only your record answers — and reading that record is what a consultation is for.
Immigration counsel to Fortune 500 employers at a national firm · Adjudicated 12,000+ visas at the U.S. Consulate, Mexico · Working in U.S. immigration since 2008
Featured in Newsweek, Condé Nast Traveler, Daily Mail