H1B Lottery 2027: Registrations, Odds and the Wage Rule

The H1B lottery is the selection USCIS runs when employers register more people than the annual cap can take: 65,000 regular visas plus 20,000 for holders of a US master's degree or higher. For fiscal year 2027 it stopped being a plain random draw. A rule that took effect on 27 February 2026 gives a registration one to four entries depending on the wage the employer offers, and 211,600 registrations went into that first weighted draw, according to preliminary USCIS figures reported by SHRM. USCIS has not published how many it selected, which is why every "odds" number you read for 2027 is an estimate.

This post uses USCIS, the Federal Register and the rule's own analysis. It is general information, not legal advice. For anything about your own status, talk to your school's designated school official or an immigration attorney.

How the lottery works now

You do not enter the lottery yourself. USCIS describes the registrant as the "prospective petitioner", the employer that wants to file an H-1B petition for you, and the employer pays a fee for each person it registers: $215 per registration for the FY2027 cap (USCIS).

The steps for a cap year:

1. Registration. For FY2027 the window opened at noon Eastern on 4 March 2026 and closed at 5:00 p.m. Eastern on 19 March 2026. An employer may have only one registration per person per fiscal year. If it submits two, USCIS invalidates all of that employer's registrations for that person after the window closes, and USCIS says "there is no way to correct this error".
2. Selection, by person. USCIS selects unique beneficiaries, not registrations. If you are selected, every employer that registered you gets a selection notice. Selection only happens at all "if a random selection process is necessary", meaning more people were registered than USCIS projects it needs.
3. Two rounds. The final rule describes the order: USCIS "first selects from a pool of all unique beneficiaries, including those eligible for the advanced degree exemption", then selects from the people left over to fill the 20,000 master's cap (Federal Register). A US master's degree gives you two chances in the same year.
4. Weighting. Within that process, each person is entered once, twice, three times or four times according to wage level, covered in detail below.
5. Petition. A selected employer can then file the actual H-1B petition.

The results do not come to you directly. They appear in the employer's USCIS account as a status for each person registered. USCIS lists the possible statuses, including "Submitted", "Selected", "Not Selected", "Denied" and "Invalidated - failed payment". The one that confuses people is "Submitted": a registration that was not picked in the first round stays "Submitted", and USCIS says registrants "will not be notified until after USCIS has determined that the H-1B cap for that fiscal year has been reached". If your employer says your status has not changed in April, that usually means not selected so far.

Registration numbers year by year

USCIS publishes the counts for each cap year on its registration page. The last column is our arithmetic: selected divided by eligible registrations.

The FY2027 row comes from SHRM, which reported that "according to USCIS, 211,600 registrations were submitted for FY 2027, down 38.5% from 343,981 registrations the previous year", and noted that USCIS had not yet released complete data. The full FY2027 row is still missing from the USCIS table, which was last updated on 21 September 2026.

Three things explain the shape of the series:

• FY2024 was inflated by duplicates. That year 408,891 eligible registrations belonged to people with more than one registration. USCIS then switched to selecting by person, and that count fell to 47,314 in FY2025 and 7,828 in FY2026. USCIS reports an average of 1.01 registrations per person for FY2026, against 1.06 for FY2025.
• Registrations are not people. USCIS counted about 442,000 unique beneficiaries in FY2025 and about 339,000 in FY2026.
• USCIS always selects more than 85,000. Not every selected employer goes on to file, so USCIS selects "the number of unique beneficiaries projected as needed to reach the H-1B numerical allocations". In FY2026 that was 120,141 selections for 85,000 visas, about 1.4 selections per visa.

What weighted selection changed

DHS published the final rule, "Weighted Selection Process for Registrants and Petitioners Seeking To File Cap-Subject H-1B Petitions", on 29 December 2025 (90 FR 60864), and it took effect on 27 February 2026, in time for the FY2027 season (Federal Register).

USCIS summarizes the mechanism on its cap season page: registrations assigned wage level IV "are entered into the selection pool 4 times", level III three times, level II twice and level I once. Each person still counts once toward the cap no matter how many entries they hold.

The details that matter to a candidate:

• The level is relative, not absolute. It is the highest Occupational Employment and Wage Statistics (OEWS) level that the offered wage "equals or exceeds" for the job's occupation code and work location. The rule notes that "OEWS prevailing wages and wage levels are set by DOL", and anyone can look them up in the Department of Labor's OFLC Wage Search. The same salary can be level II in one city and level III in another. The guide to H1B wage levels shows how to find the level for your own offer.
• Several work locations means the lowest level. If you will work in more than one place, the employer "must select the lowest corresponding OEWS wage level" among them.
• Several employers means the lowest level too. DHS chose to assign a person "to the lowest OEWS wage level among all of the registrations submitted on his or her behalf". Its own example: someone registered at level I by one employer and level IV by another "will be assigned to wage level I". A second registration can lower your chances rather than raise them.
• The wage cannot quietly drop later. The rule lets USCIS deny a new or amended petition if lowering the offered wage to a lower level looks like an attempt to "unfairly increase the odds of selection".

For a sense of scale, the rule cites the FY2024 national median salaries of H-1B workers in computer and mathematical occupations: $89,253 at level I, $106,000 at level II, $140,000 at level III and $163,257 at level IV. Those are medians of filed petitions, not the thresholds for your job, which depend on the occupation code and location.

Early signs of the effect come from the same SHRM report: "71.5% of selected beneficiaries hold U.S. advanced degrees, compared with 57% in FY 2026", and "only 17.7% of selected registrations fell within the lowest prevailing wage category".

Your odds, and why 40% is a guess

DHS published its own projection in the rule. Under the old random process, it put any one person's chance at 29.59 percent across all wage levels. Under weighted selection it estimated:

Read those as the shape of the rule, not as FY2027 results. DHS built them on a five-year average of 320,711 unique beneficiaries and on the wage mix of recent cap petitions, which it put at 28 percent level I, 55 percent level II, 12 percent level III and 5 percent level IV. FY2027 drew 211,600 registrations, a much smaller pool, and its wage mix is not public. The real per-level odds for FY2027 will only be known when USCIS publishes the selected count and the breakdown.

Then there is the number that circulates as "your odds": about 40 percent. It comes from dividing the 85,000 cap by 211,600 registrations, which gives 40.2 percent. One H-1B data blog's own table gives a range rather than a figure, listing an estimated FY2027 selection rate of roughly 34 to 42 percent (H1B Data Hub). Neither is a measurement, for three reasons:

1. 85,000 is the number of visas, not selections. USCIS has selected well over 85,000 every year in the table above. Put FY2026's 120,141 selections over 211,600 instead and you get 56.8 percent. That is also a guess, which is the point: the answer moves by more than 16 points depending on a number USCIS has not released.
2. Registrations are not people, and 211,600 is a preliminary figure for registrations submitted, not the eligible unique beneficiaries USCIS selects from.
3. There is no single rate any more. By DHS's projection, a level IV registration had four times the chance of a level I one. An average across both describes nobody.

The honest answer for 2027 is that your chance depended mostly on your wage level, somewhat on whether you hold a US master's degree, and on a selection count that has not been published.

If you are not selected

For FY2027 the draw is over. SHRM reported that USCIS confirmed on 17 July 2026 that it had received enough petitions to fill both caps, that "USCIS said it will not conduct a second H-1B lottery for FY 2027", and that "the next opportunity to sponsor an individual under the H-1B cap will be in March 2027 for the FY 2028 cap". USCIS itself says registration for cap-subject jobs "is typically in March" (USCIS).

The options people usually weigh, each to confirm with your school or an attorney:

• Cap-exempt employers. USCIS states that H-1B workers "petitioned for or employed at an institution of higher education or its affiliated or related nonprofit entities, a nonprofit research organization, or a government research organization, are not subject to this numerical cap". Those employers do not go through the lottery. USCIS also notes that moving later from cap-exempt to cap-subject employment means the new employer has to register. The post on H1B cap exempt employers covers how to find them.
• More time on OPT. Study in the States says eligible F-1 students with STEM degrees from SEVP-certified schools "may apply for a 24-month STEM OPT extension" (DHS Study in the States). More time in status is more time in which an employer can register you again.
• Another registration next March. An employer that registered you this year can do so again for FY2028. Under the weighted rule, the wage level on the offer matters as much as the registration itself.

Some engineers, researchers and founders with a strong record also look at a visa outside the lottery; the guide to O1 visa criteria and evidence sets out what it takes.

One more rule sits next to the lottery. A presidential proclamation of 19 September 2025 requires an additional $100,000 payment for certain H-1B petitions, mainly for people outside the United States without a valid H-1B visa; USCIS says it does not apply to a petition for someone inside the United States who is granted a change of status, amendment or extension. As of USCIS's page last updated on 21 September 2026, a federal court vacated the agency's guidance on 8 June 2026, the First Circuit denied the government's request for a stay on 24 July 2026, and DHS says it "will comply with the court's order while DHS considers next steps" (USCIS). The H1B $100K fee post follows that separately.

Keeping your search moving

The lottery is the part you cannot control. The offer is the part you can, and a few rules above change how to go after it.

Timing. A cap-subject employer can only register you during the March window. An offer that lands in April from a cap-subject employer is an offer for the following year's draw. Cap-exempt employers are not tied to that calendar.

Who sponsors. Start from the employers that have actually filed. USCIS's H-1B Employer Data Hub lists approvals by employer, and the guide to companies that sponsor H1B shows how to turn it into a target list.

Wage level. A role that pays at a higher level for its occupation and location now carries more entries. That is an argument for aiming at roles that fit your real experience level, not for asking an employer to inflate a wage, which the rule is written to catch. And because the lowest level among all your registrations is the one used, it is worth knowing what level each employer plans to register you at.

Volume. You are competing for offers from a limited set of employers in a narrow window, and international candidates are not a small group: IIE's Open Doors count put OPT participation at 294,253 in 2024/25, up 21 percent (IIE). Applying broadly and early to employers that sponsor is the lever you have. Our job search strategy post covers how to spread applications without losing quality.

ApplyEze can take on the applying. It watches the career boards of more than 100,000 employers directly, writes a resume for each posting, adds a cover letter when the form asks, and completes applications on 12 platforms in a browser on your own machine. Hard facts come from your profile, and saved answers for work authorization are reused. When a required question is not covered by your saved answers, ApplyEze still fills it rather than stopping the run, so that answer may not come from you. Fill in your saved answers, use dry run first, and read your receipts. It does not tell you whether an employer will sponsor you, so read each posting's sponsorship language yourself and answer those questions truthfully. For what to check before using any such tool on a visa, see auto-apply tools for international students.