EB-1A Approval Rate Hits 32.2%: FY2026 Q3 USCIS Data Explained

USCIS published the FY2026 Q3 Form I-140 data file in early September. The EB-1A line is the worst it has been in the seven quarters we track. Fewer than one in three decided extraordinary ability petitions was approved.

This analysis uses only the primary spreadsheets from the USCIS immigration and citizenship data page, the FY2025 Q4 file for the earlier quarters and the FY2026 Q3 file for the recent ones. Every rate below is approved divided by approved plus denied, computed per quarter. Where we draw an inference, we say so.

One caution before any of these figures goes into a client letter or a firm newsletter.

EB-1A Approvals Fell Six Quarters in a Row, to 32.2%

The decline did not come in one step. It came in six.

In FY2025 Q1 (October to December 2024), USCIS approved 3,227 EB-1A petitions and denied 1,091. By FY2026 Q3, approvals had dropped to 1,348 and denials had risen to 2,832. Approvals fell 58%. Denials rose 160%. The total number of decisions barely changed: 4,318 in the first quarter of the series and 4,180 in the last.

That last point matters for reading the trend. USCIS is deciding roughly the same number of EB-1A petitions per quarter as it did in late 2024. The mix of outcomes flipped.

EB-1A (extraordinary ability)EB-2 NIWEB-1B (outstanding researcher)

Approval ratePending cases

The chart shows three different stories. EB-1B (outstanding professor or researcher, employer-filed) barely moves, drifting from 98.6% to 93.0%. NIW drops hard into FY2025 Q4, bottoms at 35.7%, and then recovers for three straight quarters. EB-1A never recovers.

Here are the EB-1A numbers as text, so they can be checked against the spreadsheets:

Quarter	Months	Approved	Denied	Approval rate	Pending at quarter end
FY2025 Q1	Oct to Dec 2024	3,227	1,091	74.7%	13,526
FY2025 Q2	Jan to Mar 2025	3,402	1,276	72.7%	16,093
FY2025 Q3	Apr to Jun 2025	3,508	1,765	66.5%	18,081
FY2025 Q4	Jul to Sep 2025	2,331	2,033	53.4%	21,157
FY2026 Q1	Oct to Dec 2025	2,164	2,394	47.5%	24,785
FY2026 Q2	Jan to Mar 2026	1,720	2,394	41.8%	27,126
FY2026 Q3	Apr to Jun 2026	1,348	2,832	32.2%	28,874

For the fiscal year to date (October 2025 through June 2026), USCIS approved 5,232 EB-1A petitions and denied 7,620. That is a 40.7% approval rate, against 66.9% for all of FY2025. We covered the first leg of this decline, from 74.7% to 53.4%, in our FY2025 approval rate analysis. The FY2026 numbers show it did not level off.

Denials Rose 2.6x While Filings Fell

Receipts tell you about demand. EB-1A filings did not surge. They fell.

USCIS received 7,370 EB-1A petitions in FY2025 Q1 and 6,036 in FY2026 Q3, an 18% drop. FY2026 receipts by quarter were 7,814, 6,646 and 6,036, for 20,496 year to date. Declining filings and a falling approval rate in the same period suggest that at least some practitioners and self-petitioners are already steering away from the category, although the data cannot confirm motive.

The queue still grew. Every quarter, USCIS took in more EB-1A petitions than it decided. In FY2026 Q3 it received 6,036 and decided 4,180. The pending count went from 13,526 at the end of December 2024 to 28,874 at the end of June 2026, more than double.

EB-1A (extraordinary ability)EB-2 NIW

Approval ratePending cases

NIW's backlog dwarfs EB-1A's. At the end of FY2026 Q3, 96,297 NIW petitions were pending, up from 56,611 seven quarters earlier. NIW receipts fell even faster than EB-1A receipts, from 20,141 to 13,831 per quarter (down 31%), and the queue grew anyway.

A rough way to read these queues: divide the pending count by the number of decisions in the latest quarter. For EB-1A that is 28,874 divided by 4,180, or about 6.9 quarters of work (roughly 21 months) at the Q3 pace. For NIW it is 96,297 divided by 6,938, about 13.9 quarters (roughly 42 months). This is our arithmetic and has nothing to do with official USCIS processing times. Premium cases jump the line, and USCIS warns that pending counts shift with administrative closures and transfers. For comparison, Manifest Law's I-140 processing time tracker puts regular processing at about 32.5 months for EB-1A and 30 months for NIW.

NIW Now Beats EB-1A, While EB-1B and O-1 Hold Steady

The comparison across categories is where the Q3 file says the most.

• EB-2 NIW: 55.3% (3,835 approved, 3,103 denied). Up from 35.7% in FY2025 Q4. NIW has out-approved EB-1A since FY2026 Q2.
• EB-1B: 93.0% (1,408 approved, 106 denied). Outstanding professor or researcher petitions require a U.S. employer and an offer of permanent research or teaching work.
• EB-1C multinational managers and executives: 97.4% (3,658 approved, 96 denied).
• O petitions: 91.0% for October 2025 through June 2026 (21,321 approved, 2,096 denied), from the FY2026 Q3 I-129 RFE data file. This file combines O-1 and O-2. 24.2% of completed O petitions received an RFE.
• EB-1A: 32.2%. The only first-preference category open to self-petitions, and the only one that collapsed.

Two readings fit that pattern. One is about the pool: EB-1B and EB-1C petitions arrive through employers who screen candidates first, so weak cases get filtered before USCIS sees them. Manifest Law's EB-1 approval rate page quotes an attorney making that argument. The other is about the standard: EB-1A carries the "small percentage who have risen to the very top of the field" definition in 8 CFR 204.5(h)(2), and USCIS reviews every EB-1A petition against that definition in a separate final merits step.

Neither reading explains why EB-1A fell and NIW recovered within the same period. That divergence is the most interesting fact in the release, and the data alone cannot resolve it.

What the FY2026 Q3 Data Does Not Show

Before quoting 32.2% to a client or in a blog post, know what it cannot support.

1. It is not a filing cohort. A petition denied in May 2026 may have been filed in 2024. A shift in the approval rate can reflect which old cases USCIS chose to clear, as much as how it treats new ones.
2. It does not separate premium and regular processing, so it cannot answer whether premium cases fare better or worse.
3. It does not separate denials after an RFE from denials without one, or criteria-stage denials from final merits denials.
4. It does not show withdrawals or administrative closures. USCIS says those actions are not reflected in the report, and they move cases out of the pending count without showing up as approvals or denials.
5. It does not count people. USCIS notes that one beneficiary can have several petitions in the file, so a client who refiled after a denial can appear twice.
6. It does not cover field of endeavor, nationality outcomes, or petitioner type. The state receipt tables cover filings, not outcomes.
7. It cannot say why. Every causal story you read about these numbers, including ours, is an inference layered on top of counts.

None of this makes the trend less real. Six consecutive quarterly declines with stable decision volume is not noise. It does mean that any article claiming to know the cause from this data alone is overreaching.

Documented Changes vs Plausible Explanations

Some things are on the record. Here is what is documented, kept apart from what is inferred.

Documented: USCIS still applies final merits. The USCIS Policy Manual, Volume 6, Part F, Chapter 2 describes the two-part analysis from Kazarian v. USCIS, 596 F.3d 1115 (9th Cir. 2010): count the criteria under 8 CFR 204.5(h)(3), then decide on the totality of the record whether the beneficiary has sustained national or international acclaim and sits among the small percentage at the top of the field.

Documented: Mukherji did not change national practice. In Mukherji v. Miller, No. 4:24-cv-03170 (D. Neb. Jan. 28, 2026), the court held that USCIS adopted the final merits step without notice-and-comment rulemaking. The government appealed to the Eighth Circuit (No. 26-1578), then moved to dismiss its own appeal on June 10, 2026, and the Eighth Circuit docket shows the dismissal granted the same day. No circuit precedent came out of it. Sedaghat Law's September 11 review notes the Policy Manual has not been amended.

Documented: the AAO keeps applying final merits. In an August 27, 2026 non-precedent decision involving a classical composer who met three criteria, the AAO acknowledged Mukherji, said a Nebraska district court ruling does not bind a case arising out of New York, and dismissed the appeal at final merits. It faulted the record for failing to contextualize awards and compare the beneficiary with "other renowned composers in the field," citing Nakos (N.D. Tex. July 15, 2026) and Trajano (D. Mass. July 24, 2026). Our breakdown of why meeting criteria is no longer enough walks through the pattern.

Inferred: stricter final merits review drives the EB-1A drop. Decision volume stayed flat while the outcome mix flipped, EB-1A (with its top-of-field standard) fell while employer-screened categories did not, and AAO decisions show final merits as the dispositive step. That is consistent with tougher final merits review. It is not proven by the counts.

Not a cause of Q3: PA-2026-05. The policy alert came out on August 5, 2026. FY2026 Q3 ended June 30. Whatever PA-2026-05 does to outcomes will start to show in FY2026 Q4 data at the earliest.

What PA-2026-05 Changes for Filing Strategy

PA-2026-05, the August 5, 2026 USCIS policy alert on evidence, RFEs and NOIDs, took effect immediately for pending and new filings. Four points from the text matter for I-140 practice:

• USCIS "may deny the benefit request for failure to establish eligibility based on lack of required initial evidence without first issuing an RFE or NOID."
• Officers are no longer told to give the full 12-week response period. Under 8 CFR 103.2(b)(8)(iv), twelve weeks is only the ceiling. An officer can set less.
• The extra 14 days for notices mailed abroad is gone. Only the 3 mailing days in 8 CFR 103.8(b) remain.
• Any response, even a partial one, is treated as a request for a decision.

Practitioners are already seeing short windows. RN Law Group's September 2 explainer reports RFEs with about 30 days to respond, against the roughly 84 days practitioners were used to. Petitioners on Reddit describe the same thing, anecdotally: an NIW petitioner with a 30-day RFE on all three prongs and an EB-1A premium filer whose thread includes a commenter with a 30-day August RFE on original contributions that was later approved.

The policy is under challenge. On October 5, 2026, the American Association of University Professors and other plaintiffs sued DHS over the August evidence policy and the May 2026 adjustment-of-status policy, with counsel from Democracy Forward, the American Immigration Council, AILA and Colombo & Hurd, per the American Immigration Council's announcement. Until a court says otherwise, plan around the policy as written.

What to Change in EB-1A Practice Now

USCIS still approved 1,348 EB-1A petitions in a single quarter, so the category is open. The petitions that win are built for the second step from the first page.

Build the petition around final merits. Draft the totality argument first, then pick the criteria that feed it. The AAO composer decision shows the failure mode: three criteria met, no comparison to peers, no context for the awards, appeal dismissed. Each exhibit should answer one question the officer will ask at step two: how does this person compare with the people at the top of this field, and who outside their circle says so? Our RFE prevention playbook has exhibit-level patterns.

File as if no RFE is coming. Under PA-2026-05 the RFE is a courtesy, and when it comes it may come with 30 days on the clock. Anything you would have saved for an RFE response (the independent letter you could not get in time, the citation benchmarking, the salary survey) belongs in the initial filing. If it cannot be ready, that is an argument for waiting.

Treat initial evidence as a hard gate. Missing required initial evidence is now an independent basis for denial without an RFE. Run a regulation-by-regulation completeness check against 8 CFR 204.5(h)(3) and the form instructions before filing.

Plan the RFE response before filing. If 30 days is the window, the response team, the recommenders who can turn around a letter in two weeks, and the benchmark data need to be lined up in advance. Response structure has not changed much. The calendar has.

EB-1A or NIW First: Approval Odds vs Visa Availability

On Q3 numbers, a decided NIW petition was 23 points more likely to be approved than a decided EB-1A petition. If approval odds were the only variable, NIW would win. They are not.

The October 2026 Visa Bulletin lists EB-1 for All Chargeability Areas as current. EB-2 for the same group now carries a final action date of January 1, 2025, and the bulletin states that some Rest of World employment categories retrogressed to stay within FY2027 limits. USCIS's adjustment of status filing charts page tells employment-based applicants to use the Dates for Filing chart in October, which puts EB-2 Rest of World at March 15, 2026 for filing adjustment applications.

So the choice for a Rest of World client looks like this:

• EB-1A: lower approval odds (32.2% in Q3), a smaller queue (28,874 pending), premium action in 15 business days, and a visa number available on approval.
• NIW: higher approval odds (55.3% in Q3), a queue of 96,297, premium action in 45 business days, and a priority-date wait after approval that did not exist a few months ago.

NIW also has a review problem that EB-1A does not. In the Ninth Circuit, Poursina v. USCIS, 936 F.3d 868 (9th Cir. 2019), held that 8 U.S.C. 1252(a)(2)(B)(ii) strips federal courts of jurisdiction to review USCIS's discretionary denial of a national interest waiver. The court leaned on Poursina again in Barman v. USA, No. 24-464 (9th Cir. Sept. 14, 2026), a published opinion on NM-1 status that collects similar holdings from other circuits. EB-1A denials have been litigated on the merits in district court, Mukherji among them. Barman's list of cases shows that route closed to NIW petitioners in at least the Ninth and Eleventh Circuits.

India and China change the math again. EB-1 India has a February 1, 2023 final action date and EB-1 China July 1, 2023 in the October bulletin, while EB-2 India sits at November 1, 2013. For those clients the visa-availability gap between EB-1 and EB-2 is measured in years, which can justify an EB-1A attempt even at a 32% approval rate.

How to Set Client Expectations at 32%

Clients find these numbers. Some will have read the same Manifest Law page you did. Get ahead of it.

Say the number plainly, and then say what it measures. A 32.2% rate is the share of EB-1A decisions in one quarter that were approvals, across every petition decided, strong and weak, recent and old, attorney-filed and self-filed. It is not a forecast for their file. What changes their odds is the record: independent evidence, context for each achievement, and a final merits argument that survives an officer looking for a reason to say the beneficiary is "progressing toward the top" rather than at it, the phrase the AAO used about the composer.

Then give them the timeline under PA-2026-05. If an RFE arrives, they may have about a month, and they will be the one chasing recommenders. Clients who know that at intake make better decisions about when to file.

Last, put the backup plan in writing before the EB-1A goes in. An NIW or O-1 path agreed in advance is a strategy. One improvised after a denial feels like a failure, even when it was always the sensible second option.

Immigration Copilot maps each exhibit to the EB-1A criteria and to the final merits argument, then checks every claim in the draft against the evidence before an attorney files it. If you want to see how that works on a petition built for the current standard, book a demo.