Think Immigration: For Founders, the O-1A and EB-1A Question Is No Longer “Which One?”

Sooner or later, most founder clients ask the same thing across the table: “Can we skip the O-1A and file the EB-1A directly?” For years the answer turned largely on the strength of the record the day we met. In 2026 it turns on something else, and the question itself has changed shape. It is less which classification than in what order, and whether one body of evidence can be built to serve both.

What changed is where these cases are being lost. An EB-1A petitioner must satisfy at least three of ten regulatory criteria or show one major internationally recognized award. Since 2010, agency guidance has layered a second step on top: even after the criteria are met, an officer weighs the whole record to decide whether the person has sustained acclaim and sits at the very top of the field. For years that step confirmed approvals. Beginning in 2025 it became the place petitions were denied, often in language faulting achievements as too recent or acclaim as not sustained, in cases where the criteria had already been conceded.

Mukherji v. Miller, decided in January 2026 in the District of Nebraska, put that second step under a spotlight. USCIS agreed the petitioner met five of the ten criteria, nearly double what the regulation requires, then denied at the final merits stage for lack of sustained acclaim. The court rejected that approach on process grounds: the agency had adopted a mandatory two-step framework through internal guidance rather than notice-and-comment rulemaking, had never acknowledged it was changing long-standing policy, and had applied a recency expectation that appears nowhere in the statute. The court ordered the petition approved outright.

What happened next matters as much as the ruling. USCIS appealed to the Eighth Circuit, which would have been the first appeals court to address whether the framework required formal rulemaking, then in June 2026 asked that court to dismiss its own appeal. The petitioner kept her approval, but no appellate precedent was set. Days later the Administrative Appeals Office issued a separate non-precedent decision taking on the Mukherji reasoning and rejecting it. Read together: the agency stepped back from a test it might have lost on appeal, and it is not abandoning the framework.

For practitioners advising founders, three considerations stand out. First, founder evidence does not map cleanly onto criteria drafted with scientists and artists in mind. A valuation, a funding round, an accelerator's selectivity, none of these is a regulatory criterion. Each has to be translated. A large round is not by itself an original contribution of major significance; it has to be presented as independent validation of an innovation others in the field have adopted or built on. Accelerator selectivity can support an awards or critical-role argument when the acceptance rate is documented rather than asserted. The facts alone rarely carry a founder case. The framing, and the corroboration behind it, decides whether they register.

Second, the final merits step penalizes the ordinary shape of a founder's career. Intense achievement compressed into a few years, followed by an exit or a pivot, reads on a cold record as acclaim that was not sustained. A founder whose strongest press clusters around a single moment two years ago is precisely the profile that draws the objection the court found unlawful in Mukherji.

Third, sequencing is where most of the value now sits. For the founder with a genuinely durable record, meaning recognition spanning years, independent corroboration, and evidence that translates without strain, EB-1A remains the most efficient route, filed with the final merits objection answered before it is raised. For many others the O-1A is the better entry point, not as a consolation but because its criteria overlap substantially with EB-1A's. An O-1A record can be assembled so every exhibit does double duty, and the years in that status convert a snapshot into the continuous timeline a final merits reviewer looks for. Because the filing of an immigrant petition does not, by itself, preclude approval of an O-1A petition or extension, practitioners may be able to pursue the two strategies concurrently.

Underneath all three sits a point about preservation. Inside the agency the two-step test still governs, and the Administrative Appeals Office has signaled that it will defend the framework and will not entertain a challenge raised for the first time on appeal. When a denial imports requirements the regulation does not contain, recent achievements, unbroken acclaim, an undefined pinnacle, the objection has to be made early and on the record at the service center. Made late, it is lost before the case ever reaches a federal judge.

None of these developments forecloses EB-1A classification for founders. What has changed is that the planning window now matters more than the filing. For those of us advising founders, the most valuable conversation happens before the first filing, not after the first denial: whether the record is durable enough to go straight to EB-1A, and if it is not, what the next two years under an O-1A need to produce.