USCIS reached the 85,000 H-1B cap on July 17, 2026 with no second selection round, after unique-beneficiary registrations fell by more than a third in the first season run under wage-weighted selection. The H-1B did not become harder to win — it became more expensive to want, which is why attention is moving toward categories with no cap and no lottery.
On July 17, 2026, USCIS announced it had received enough H-1B petitions to reach the congressionally mandated 85,000 cap for fiscal year 2027. There will be no second selection round: a registration that was not selected in the March 2026 draw cannot be used to file a cap petition this year.
FY2027 drew about 211,600 unique-beneficiary registrations — down more than a third from the prior year. That is the headline number, and it is the one that explains everything else on this page. Demand did not get filtered out at adjudication; more than a third of it simply never entered.
This was the first cap season run under weighted selection. The final rule (FR Doc. 2025-23853, published December 29, 2025, effective February 27, 2026) allocates entries by the OES wage level of the offered position: Level 4 gets four entries, Level 3 gets three, Level 2 gets two, Level 1 gets one. Selection remains possible at every level — the rule changes the weighting, not the eligibility.
The honest arithmetic runs in two directions at once. The cap did not shrink, so a smaller pool means better aggregate odds than the prior year. But a Level 1 registrant’s relative position got worse, because they now hold one entry against a Level 4 registrant’s four. Both statements are true, and anyone telling you only one of them is selling something.
Two separate legal threads are often collapsed into one. The weighted selection rule is in force: as of late July 2026 there is no injunction and no vacatur, and — notably — no business or immigration organisation has filed suit against it, though many practitioners consider it vulnerable. The $100,000 fee is the opposite: it was struck down. We cover that fight in full in a separate entry.
In one line, the $100,000 fee (Proclamation 10973, signed September 19, 2025) only ever reached petitions for beneficiaries outside the United States without a valid H-1B, or petitions requesting consular, port-of-entry, or preflight notification. It never reached change-of-status, extension, or amendment petitions for people already here — which is most first-time cap petitions, filed for F-1 students already in the country. A federal court vacated it on June 8, 2026 and the First Circuit refused to revive it in late July 2026, though a D.C. court had upheld it in December 2025 and that appeal is live. Full detail: /insights/california-v-mullin-h1b-fee.
The fee was nonetheless in force during the FY2027 registration window, and it is part of why registrations fell. A one-time six-figure charge that most cap filers were never going to pay still changed behaviour, because the sponsoring decision gets made months before anyone reads the scope carve-outs.
The strategic consequence is a shift in attention toward categories with no annual cap and no lottery — O-1A first, with EB-1A behind it. In those categories the binding constraint is not luck. It is evidence: a documented record measured against published regulatory criteria.
The O-1A adjudication data points the same way from the other side. Trackers citing USCIS figures put O-category approvals at roughly 91.0% in Q1 FY2026 (about 6,436 approvals against 639 denials) and about 93.9% across FY2025 (roughly 29,733 of 31,681). Approvals holding above 90% while officers ask more questions is the signature of a bar that rewards better-documented files — not of officers rejecting more people.
This entry is informational only and is not legal advice. It does not and cannot tell you whether you qualify for anything or what your chances are. Aggregate statistics describe populations; they do not predict individual outcomes. Confirm your own situation with a licensed immigration attorney.
For years the H-1B story was about a crush of registrations and a random draw. FY2027 is a different story. About 211,600 unique-beneficiary registrations came in — more than a third fewer than the year before — and the 85,000 cap still filled from a single March selection round, so there is no second bite this year.
Two things changed at once, and it matters that they are separate. First, the government stopped drawing names at random. Under the weighted selection rule that took effect on February 27, 2026, your entries in the draw depend on the wage level of the job you were offered: four entries at Level 4, three at Level 3, two at Level 2, one at Level 1. Everyone is still in the pool. The pool is just no longer flat.
Second, a $100,000 fee was hanging over the whole season. It was much narrower than most people believed — it never applied to F-1 students changing status inside the United States, who are the largest single group of cap beneficiaries at just under half — but it was live during the registration window, and a six-figure headline number does not need to apply to you in order to change what your employer decides in January. It was later struck down; that fight has its own page.
So the useful way to read FY2027 is this: the H-1B did not get harder to win. It got more expensive to want. Sponsorship became a bigger corporate decision, and more than a third of the demand quietly declined to make it.
Now the arithmetic, stated honestly in both directions. The cap did not shrink. A smaller pool competing for the same 85,000 slots means the aggregate odds were better than the prior year. At the same time, if the role you were offered sits at Level 1, your position relative to other registrants got worse, because a Level 4 registrant is holding four entries to your one. Neither of those sentences cancels the other, and neither of them says anything about any specific person.
That combination — an aggregate improvement that is unevenly distributed, plus an unpredictable political overlay — is why a lot of attention has moved to the categories that have no cap and no lottery at all. O-1A is the near-term one; EB-1A is the permanent-residence analogue behind it. You do not enter a draw for either. You file when your record is ready, and you can file at any point in the year.
The trade is real, though, and it is not a shortcut. In the H-1B the constraint is luck plus a wage level you mostly do not control. In O-1A the constraint is evidence: eight regulatory criteria at 8 CFR 214.2(o)(3)(iii)(B)(1)–(8), of which a petition must satisfy at least three — or show a major, internationally recognised award. That is a documentation problem, and documentation problems are solvable in a way that a lottery is not. They are also expensive and slow: typical all-in O-1 costs run roughly $10,000 to $25,000 including counsel and premium processing, and the premium processing fee alone is $2,965.
One more reading that gets mangled constantly. O-1 approval rates are still high — trackers citing USCIS data put them around 91% in the most recent quarter and about 94% across FY2025 — while RFEs have been climbing and processing times have lengthened by roughly four and a half months between December 2025 and May 2026. People read the rising RFEs as a crackdown. The approval rate says otherwise. When approvals stay above 90% and officers ask more questions, and when trackers report that roughly 70.9% of FY2025 petitions that drew an RFE were approved anyway, what is being demanded is a better-documented file, not a different person.
This is general information, not legal advice, and it describes a landscape that includes active litigation. Nothing here is a determination that you are eligible for anything, and nothing here is a prediction about your case. Those are questions for a licensed immigration attorney who has looked at your actual record.