The H-1B lottery used to treat a $60,000 registration and a $300,000 one exactly the same: a coin flip, just under 30% either way, by the Department of Homeland Security's own account. A rule that took effect February 27, 2026 changed that, weighting the pool so a registration tied to the highest wage tier now clears above 61%. For a funded engineering team paying senior-market wages in the Bay Area, that's a real improvement over a flat lottery. It still isn't a hiring plan, and DHS's own numbers say why.
What the weighted selection process actually does
DHS finalized "Weighted Selection Process for Registrants and Petitioners Seeking To File Cap-Subject H-1B Petitions" in the Federal Register on December 29, 2025, effective February 27, 2026, ahead of the fiscal year 2027 cap season. When registrations exceed what USCIS projects it needs to fill the 65,000 regular cap plus the 20,000 U.S. advanced-degree exemption, and random selection is required, USCIS assigns each unique beneficiary the highest Occupational Employment and Wage Statistics (OEWS) wage level their offered wage equals or exceeds for the relevant occupation code and metro area, then enters that beneficiary into the selection pool more than once depending on the tier. A Level IV wage gets four entries, Level III gets three, Level II gets two, and Level I gets one. Each beneficiary still only counts once toward the cap if selected, so the extra entries change the odds, not the arithmetic of how many people get in.
Why a Bay Area engineering hire rarely lands at wage level I
OEWS levels are set relative to the wage distribution for that occupation in that specific metro area, not against a national baseline. A software developer role priced against Bay Area comp, where BLS puts the average San Jose developer salary at $221,710 against a $148,100 national mean (see what a Bay Area software engineer actually costs), is being measured against a local wage floor that's already elevated. That's an inference from how OEWS levels are constructed, not a figure DHS states directly, but it's the mechanical reason a staff or senior engineering offer in Mountain View or San Francisco tends to clear into Level III or IV territory almost by default, while the same title and years of experience in a lower-cost metro can land a tier lower against its own local wage table.
What DHS's own numbers say about who loses
DHS didn't present this as a neutral shuffle. Its own regulatory impact estimate, published in the same December 2025 final rule, projects the number of selected Level I wage petitions will fall by 10,099 a year, and quantifies the flip side as a wage transfer toward higher-paid workers: $858 million in fiscal year 2026, rising each year to $4.292 billion annually from fiscal year 2030 through fiscal year 2035. That's DHS describing its own rule's intended effect, not an unintended side effect discovered later. The department's stated reasoning is to stop the H-1B program filling relatively low-wage roles at the expense of higher-skilled ones, which is a defensible policy goal and also, mechanically, a direct subsidy to exactly the profile a funded, well-paying engineering team is trying to hire.
What happened when the first cap season closed
The prior full cycle gives a baseline: for fiscal year 2026, USCIS's own registration data shows 358,737 total registrations, 343,981 of them eligible, and 120,141 selected in the initial round. The fiscal year 2027 season, the first run entirely under the weighted rule, opened its registration window March 4 through March 19, 2026, USCIS completed the initial selection process March 31, and on July 17, 2026 the agency announced it had received enough properly filed petitions to reach both the regular cap and the advanced-degree exemption. What USCIS had not done, as of this writing, is publish a wage-level breakdown of who actually got selected under the new system. The agency running the lottery is the same agency that quantified this rule's effect down to the dollar before it took effect, and it still hasn't handed back the one number a workforce planner would want most: how the pool actually split by tier once real registrations, not projections, went through it.
A worked example
Picture a 60-person, Series B computer-vision company in Mountain View trying to bring on two staff-level ML engineers on H-1B status, each offered a package that clears comfortably into Level IV for their occupation code and metro area. Under the old flat lottery, at just under 30% a registration, the odds that both clear are roughly 9%. Under the new weighted system, at DHS's own projected 61%-plus for Level IV, that rises to somewhere north of 37%. That's a genuine improvement, more than four times better, and it's still worse than a coin flip on getting both hires through in one season. A roadmap that assumes both land is planning against the minority outcome, not the majority one.
What a better lottery still doesn't fix
If a hiring plan depends on a single H-1B registration clearing the cap, weighted or not, that's a bet, not a commitment, whatever the wage level attached to it. A 61% chance is a good bet. It is not a start date. Staff augmentation through a bench that doesn't sit inside this lottery at all is a genuine hedge against that specific risk, and I'd defend that on a call. But it's a hedge against capacity continuity, not a substitute for a specific named hire: if the person you need is a particular engineer who already has to be that exact individual, on H-1B status, no subcontracted bench changes their odds in March. It only changes what the roadmap does while the lottery, or the visa process behind it, plays out.
None of this applies uniformly, either. A company whose open roles genuinely price at Level I or Level II, often the case at very early-stage or non-Bay-Area teams, is now worse off in this specific lottery than it was a year ago, and no amount of reading the rule changes that structural fact; for that team, a subcontracted or disclosed hybrid bench is a stronger argument than it is for a team already clearing Level IV. And none of this touches a role that isn't visa-dependent at all, where the entire discussion is moot. The weighted rule reallocates who wins a scarce lottery. It doesn't create more engineers, and it doesn't replace a plan for the months a registration, even a well-weighted one, hasn't cleared yet.
None of this changes when the work actually needs to start: now, not whenever a fiscal year 2028 cap season resolves. Tell us which roles you're trying to fill and we'll help work out where a disclosed, US-led bench covers the gap while an H-1B registration clears, or doesn't. The wage data behind why a Bay Area offer tends to land where it does is in what a Bay Area software engineer actually costs, and the broader squeeze this lottery sits inside is in the senior AI engineer hiring market. If a registration does clear, the first week with an embedded engineer is worth reading before they start. Our Silicon Valley page has more on how we work with funded Bay Area teams in the meantime.
Sources
- U.S. Department of Homeland Security, Federal Register: Weighted Selection Process for Registrants and Petitioners Seeking To File Cap-Subject H-1B Petitions (published Dec. 29, 2025, effective Feb. 27, 2026)
- USCIS: H-1B Electronic Registration Process (fiscal year 2026 registration and selection figures)
- USCIS: USCIS Reaches Fiscal Year 2027 H-1B Cap (July 17, 2026)
- USCIS: FY 2027 H-1B Cap Initial Registration Period Opens on March 4