The FY2024 H-1B lottery, run in March 2023, set a record nobody celebrated: roughly 780,000 registrations chasing the same 85,000 visas Congress capped decades ago. Selection odds for any single registration fell to roughly one in seven at the initial draw. But the headline number concealed the real story. When USCIS published the data, it showed that a huge share of those registrations, roughly 408,000, belonged to people entered multiple times by different companies. The lottery was not just oversubscribed. It was being gamed, and the FY2024 numbers proved it beyond argument.
The numbers, plainly stated
Consider the trajectory. FY2021, the first registration-era lottery, drew roughly 275,000 entries. FY2022: roughly 308,000. FY2023: roughly 484,000. FY2024: roughly 781,000, nearly tripling in three years while the underlying U.S. tech job market was actually cooling into layoffs. Genuine demand does not triple during a hiring downturn. What tripled was gaming: the number of unique workers grew far more slowly than registrations, and the average selected-worker's entry count made clear that multiple registrations dramatically improved odds. Roughly 96,000 individuals accounted for about 408,000 entries, more than half the pool.
Why the design invited abuse
The 2020 shift to electronic registration was meant to save employers the cost of preparing full petitions for a lottery they would probably lose. Reasonable goal, fatal flaw: entry cost only $10, and the duplicate ban applied per employer, not per worker. So loosely affiliated staffing companies could each register the same candidate, multiplying that person's odds at trivial cost. Honest single-employer candidates, a new graduate with one real offer from one real company, were crowded out by design. It was a textbook case of an administrative shortcut creating an integrity hole, and it took three cycles of worsening data before the government acted.
What the broken lottery cost
The damage went beyond unfair odds:
- Legitimate workers lost seats. Every duplicate that won displaced someone playing by the rules; some selected registrations never even converted to petitions, wasting cap numbers.
- Employers lost planning ability. Companies genuinely needing engineers faced worse odds each year and could not tell whether demand or fraud was driving it.
- The program lost credibility. A visa category already under political fire handed its critics proof of systemic abuse.
We will say the quiet part: defenders of high-skilled immigration should have been the loudest voices demanding this cleanup. Nothing discredits merit-based legal immigration faster than letting it be gamed. Integrity enforcement is not hostility to immigrants; it is protection for the honest ones.
The fix that followed
USCIS answered with the beneficiary-centric rule, finalized in early 2024 for the FY2025 cycle: each unique worker gets exactly one lottery entry regardless of how many employers register them, and a selected worker can then pursue any legitimate offer. Registrations promptly fell to roughly 479,000, and the duplicate problem largely evaporated. Registration fees were also raised substantially soon after, ending the $10 entry era. The FY2024 lottery thus became the high-water mark of the gamed system, the dataset that forced reform.
What this means for you
If you are pursuing an H-1B now, the odds are still hard, the cap is still 85,000, but the game is fairer: one person, one entry. Practical moves follow. Only allow registration by employers with genuine job offers, since false attestations now carry real consequences for you, not just the company. If you are cap-subject and unselected, look at cap-exempt employers such as universities and affiliated nonprofits, or alternatives like O-1, L-1, TN, or E-3 depending on your profile. And if your long-term goal is a green card, ask employers about starting PERM early; the lottery is a gate, not the destination.
Not sure which visa fits your situation? Take the free 2-minute Open visa quiz and get your match instantly.
Join the conversation