Suing USCIS Over a Denial: APA Challenges in Federal Court
A federal court can set aside a USCIS denial that is arbitrary, capricious, or contrary to law — but review is on the record USCIS already had, and the usual remedy is a remand, not an order to approve.
When this comes up
USCIS denied your I-140 (EB-1A, EB-1B, or NIW) or your I-129 (O-1), and the denial reads as legally wrong rather than factually close — the officer applied a standard that is not in the regulation, ignored evidence you actually submitted, or held that you failed a test the law does not impose. A denial like that is not just a bad outcome; it may be reviewable error.
Under the Administrative Procedure Act, a federal district court can review final agency action and set it aside where the agency got the law or the reasoning wrong. For a denied employment-based petition, that means a district judge — not another USCIS officer — takes a fresh look at whether the denial holds up as a matter of law. This is a different forum, a different decision-maker, and a different standard from the motion and AAO-appeal paths that stay inside the agency.
It is not the right tool for every denial, and we say so plainly. A denial that turned on a genuinely thin record is usually better answered with more evidence — through a motion, an appeal, or a refiled petition — than with a lawsuit. Litigation is for the denial that is wrong on the law the agency already had in front of it.
What the law provides
The operative standard is 5 U.S.C. § 706(2)(A): a reviewing court shall set aside agency action found to be "arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law," under the APA's grant of review to a person "adversely affected or aggrieved by agency action" (5 U.S.C. § 702). A USCIS denial is "final agency action" reviewable under these provisions. The court does not ask whether it would have approved the petition; it asks whether USCIS's stated reasoning survives review. The controlling formulation comes from Motor Vehicle Manufacturers Association v. State Farm, 463 U.S. 29 (1983): an agency must "examine the relevant data and articulate a satisfactory explanation for its action, including a rational connection between the facts found and the choice made." A denial that skips a step in that chain — ignoring evidence, applying a standard the regulation does not contain, or asserting a conclusion without connecting it to the record — is the target.
For extraordinary-ability and outstanding-researcher cases, the governing framework is the two-step analysis of Kazarian v. USCIS, 596 F.3d 1115 (9th Cir. 2010): first, whether the petitioner meets the regulatory criteria; second, a "final merits determination" of whether the record as a whole establishes the required level of acclaim. The most litigable EB-1A denials are the ones that blur those steps — importing final-merits skepticism into the count of criteria, or denying at final merits without a reasoned explanation. Federal courts have pushed back hard on exactly that failure mode. In Mukherji v. Miller, No. 4:24CV3170 (D. Neb. Jan. 28, 2026), USCIS conceded the petitioner met five criteria and then denied at final merits on the theory that her acclaim had tapered after 2015; the court vacated the denial and remanded with instructions to approve the petition, holding both that the agency never articulated what its standard required and that nothing in the statute demands an applicant "stay indefinitely at the top of their field." This is not an EB-1A-only remedy: courts have set aside employment-based denials across categories — the D.C. Circuit reversed a USCIS L-1B specialized-knowledge denial in Fogo de Chao (Holdings) Inc. v. DHS, 769 F.3d 1127 (D.C. Cir. 2014), and federal courts have likewise vacated H-1B specialty-occupation denials as arbitrary and capricious.
A federal APA suit generally does not require exhausting the AAO first. The Supreme Court held in Darby v. Cisneros, 509 U.S. 137 (1993), that a court cannot impose an exhaustion requirement unless a statute or an agency rule expressly requires the administrative appeal and makes the decision inoperative while it is pending. The USCIS appeal regulation does neither — 8 C.F.R. § 103.3(a)(1)(ii) says an unfavorable decision "may be appealed" — so for most employment-based petitions there is no rule forcing you through an AAO appeal before a court will hear the case. The important exception is EB-5: by statute, an immigrant investor or regional center must appeal an I-526 or regional-center denial to the AAO to exhaust before any court has jurisdiction. That carve-out does not apply to EB-1A, EB-1B, NIW, or O-1.
There are limits worth naming up front. Under Patel v. Garland, 596 U.S. 328 (2022), and 8 U.S.C. § 1252(a)(2)(B), courts cannot review certain discretionary judgment calls and the factual findings underlying them, and some purely discretionary denials — the Ninth Circuit treated a discretionary NIW denial this way in Poursina v. USCIS, 936 F.3d 868 (9th Cir. 2019) — are largely insulated from review. The strongest APA challenges therefore target legal error and unreasoned decision-making, not a disagreement about how the agency weighed the equities.
Two features shape everything about how these cases are litigated. First, review is on the administrative record — the court decides on the evidence USCIS actually had, and absent bad faith that record generally cannot be supplemented with new material. Second, the statute of limitations is long: 28 U.S.C. § 2401(a) gives a six-year window from when the claim accrues, in sharp contrast to the 30-day clock on a motion or an AAO appeal. That means litigation is rarely foreclosed by timing, but it also means the case will be won or lost on the record you already built.
When this remedy is the right tool
The denial conflates Kazarian Step 1 and Step 2
USCIS counted you out of a criterion using reasoning that belongs to the final-merits stage — e.g., discounting a qualifying award or membership because it was not "impressive enough," rather than because it failed the regulatory definition.
This is the classic reviewable error in extraordinary-ability adjudication. The regulations set objective criteria at Step 1; the officer's holistic judgment belongs at Step 2. A denial that smuggles Step-2 skepticism into the Step-1 count applies the wrong legal standard, which is precisely what § 706(2)(A) review is for. We map the denial against the Kazarian framework line by line to show where the agency's reasoning departed from its own regulations.
The denial imposes a requirement the regulation does not contain
The officer denied because you lacked something the rule never asks for — a specific number of citations, a particular kind of award, a novel sub-test invented in the decision itself.
Agencies do not get to add elements to their own regulations by adjudication. When a denial rests on a made-up standard, it is "not in accordance with law." The litigation theory is straightforward: identify the invented requirement, set it against the actual regulatory text, and ask the court to hold that the agency exceeded what the rule allows.
The denial ignores material evidence in the record
You submitted expert letters, a documented record of contributions, or other probative evidence, and the decision simply does not engage with it.
An agency must consider the evidence before it and explain why it was insufficient; a decision that passes over material evidence in silence is vulnerable as arbitrary and capricious. Courts have treated formulaic, boilerplate denials that fail to give an individualized analysis of the record as an abuse of discretion. We build the challenge around the specific evidence the decision failed to address.
The denial rests on the Kazarian "final merits" framework itself
You want to challenge not just how the officer applied the two-step test, but whether the second-step "final merits determination" is a lawful requirement at all.
There is a newer, more structural theory — and a federal court has now accepted it. The two-step "final merits" analysis was not written by Congress and does not appear in the 1991 rule that created the EB-1A criteria; USCIS adopted it by policy memorandum in 2010, after the Ninth Circuit floated it in dicta, without ever going through the notice-and-comment rulemaking the APA requires for a rule of that force. In Mukherji v. Miller, No. 4:24CV3170 (D. Neb. Jan. 28, 2026), the court held that memorandum "ultra vires" for precisely that reason — the agency had once classified the change as substantive, yet adopted it with only a 14-day informal comment period — and vacated the denial. Two 2024 Supreme Court decisions strengthen the attack: Corner Post v. Board of Governors, 144 S. Ct. 2440 (2024), holds that the clock to challenge a rule runs from when it injures you rather than from when it was issued, and Loper Bright v. Raimondo, 144 S. Ct. 2244 (2024), strips the deference agencies once received in defending their interpretations. Two honest caveats: a district court's ruling binds USCIS only as to the parties, so the agency continues to apply the final-merits framework everywhere else — and the question has not yet reached a court of appeals, where a ruling would bind USCIS across the Circuit. USCIS filed a notice of appeal in Mukherji and then withdrew it, which left the petitioner's win intact but kept the issue off the appellate docket. This is a live and gaining argument, not settled law; what will matter most is whether a Circuit adopts it.
You want a neutral forum instead of a second trip through the agency
The legal error is clear, and asking the same service center — or the AAO — to reverse itself is unlikely to be faster or fairer than a federal judge.
An AAO appeal keeps the decision inside the agency and can take a year or more; a motion goes back to the same office that denied you. Where the defect is a pure legal error and the record is already complete, a federal district court can be the more credible and, in some cases, the more efficient forum. Whether court or the AAO is the better first move depends on the specific defect, the record, and your timing — it is the threshold question we work through with you.
What to expect when
- You generally have up to six years from the denial to file an APA suit (28 U.S.C. § 2401(a)) — but there is rarely a reason to wait, and a stale record does not improve with age.
- A motion to reopen or reconsider, or an AAO appeal, must be filed within 30 days of the decision (33 if it was mailed). Those administrative deadlines are not extended by the fact that litigation remains available later — if you want to preserve the in-agency options, they move first.
- After a complaint is filed, the government answers and lodges the administrative record. Most petition-denial cases are then resolved on cross-motions for summary judgment briefed on that record, rather than a trial.
- The typical win is a remand — the court vacates the denial and sends the petition back to USCIS to decide again under the correct standard — not a direct order to approve. A remand is a genuine second look on a corrected legal footing, but it is not a guaranteed approval.
The honest downside
- Review is on the record USCIS already had. Absent a showing of bad faith, you generally cannot add new evidence in court. If the real problem was a thin evidentiary record, a motion, appeal, or refiling — not a lawsuit — is usually the better path.
- The usual remedy is a remand, not an approval. A court that agrees the denial was flawed typically sends the case back to the agency to redo under the right standard. USCIS can still deny on remand, though it must this time do so lawfully.
- It is expensive and slower than it sounds. Federal litigation carries real cost and can take many months to over a year. For a denial that is close on the facts, that investment may be better spent strengthening the petition.
- Case law is developing and not uniform. A decision like Mukherji v. Miller (D. Neb. 2026) is a genuine win against an aggressive final-merits denial, but it is one district court's ruling — it binds USCIS only as to the parties, is merely persuasive elsewhere, and has not been tested in a court of appeals (USCIS filed and then withdrew its appeal). USCIS still applies the final-merits framework nationwide. Outcomes vary by court and by record. We give an honest read of how a specific denial is likely to fare, not a promise.
- Not every denial is reviewable. After Patel v. Garland, courts cannot second-guess certain discretionary determinations and their underlying fact-finding, and some discretionary denials (as in Poursina) are largely unreviewable. A challenge has to be built on legal error or unreasoned decision-making, not a bare disagreement with the agency's judgment.
What we tell clients
Attorney Ryan Locke has spent nearly his entire career litigating against the government — as a public defender, in appellate practice, and now in federal immigration litigation. When USCIS denies a petition on grounds that do not hold up, suing is not a threat we make lightly or a service we upsell; it is one tool among several, and often not the right one. We will tell you candidly whether your denial is a legal-error case that belongs in federal court or an evidence case that belongs back before the agency. The goal is the client's outcome, not the lawsuit.
This page describes how APA challenges to petition denials are generally analyzed. It is general information, not legal advice about your specific case and not a prediction of any particular result. Whether litigation is appropriate, in which forum, and on what theory depends on the exact language of your denial, the administrative record, and your circumstances — and it is a decision we scope, and are upfront about the cost of, separately before any work begins.
Frequently Asked Questions
Ready to Get Started?
Tell us about your immigration needs and we'll be in touch to discuss how we can help.
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