FY2027 H-1B: The OPT Lottery Playbook Is Dead. Source Transfers Instead.

DHS finalized the wage-weighted H-1B selection rule without changes, effective February 27, 2026, and the $100,000 supplemental fee on qualifying overseas hires runs through at least September 21, 2026. FY2027 cap registrations open in March. If your sourcing plan still reads "recruit OPT juniors in the fall, gamble on the March lottery," you are running a 2015 playbook against 2026 rules.

The math changed in two places at once: who wins the lottery, and who can afford to be brought in from abroad. Both changes push the same direction, toward already-in-US, higher-paid candidates. That reshuffles which pool a startup should actually be sourcing.

What actually changed for FY2027

Two rules, stacked. USCIS now hands out lottery tickets by OEWS wage level instead of one-per-registration, and a $100,000 supplemental fee applies to most new H-1B hires who need consular processing from abroad.

Here is the weighting matrix that goes live for the March 2026 registration window:

• Level 1 wage offer: 1 lottery entry
• Level 2 wage offer: 2 entries
• Level 3 wage offer: 3 entries
• Level 4 wage offer: 4 entries
• Tiebreaker rule: if the same beneficiary is registered at multiple wage levels, USCIS uses the lowest one for the whole selection pool

DHS's own modeling assumed each wage level had roughly 29.59% selection probability under the old random model. The new model shifts probability meaningfully toward Levels II through IV. FY2027 registrations dropped to about 210,000, but the 85,000 cap still filled, so the funnel narrowed rather than the demand.

Then the fee. The $100,000 supplemental was imposed in September 2025 under a presidential proclamation and is scheduled to run through September 21, 2026. It primarily targets new hires outside the US who require consular processing. Change-of-status filings inside the US and cap-exempt petitions are not subject to it. Multiple lawsuits from hospitals, universities, and the U.S. Chamber of Commerce are pending, but as of May 2026 no court has issued an injunction and the fee is being collected on every qualifying petition.

$100,000

Why the OPT junior playbook stops working

The classic startup move - hire an F-1 grad on OPT, register in March, gamble on a 30% lottery hit - no longer pencils out at Series A budgets. Under the weighted lottery, a Level 1 offer to a new grad gets one ticket, and the highest-paid Meta or AWS registration for the same profile gets four.

The Consumer Technology Association warned on October 24, 2025 that the wage-weighted system would "disproportionately harm the F-1 pipeline" of entry-level graduates. The mechanism is compensation, not paperwork:

• OEWS Level 1 is roughly the 17th percentile of the local occupation wage. Most new-grad startup offers land here or at Level 2.
• OEWS Level 4 is roughly the 67th percentile and up. A Series A company paying a new grad at Level 4 is paying above senior-engineer market, which detonates the internal band.
• DHS explicitly endorsed paying above-role wages to improve odds. That is a legally clean fix that only well-capitalized employers can actually execute.

The downstream evidence is already visible in adjacent sectors. The American Hospital Association reported on April 1, 2026 that 64% of hospitals had "paused, deferred, or limited recruitment" because of the $100,000 fee. Tech will not publish a similar number, but the funnel arithmetic is the same.

A Series A startup paying a new grad at Level 4 gets four lottery tickets and one broken compensation band.

The multi-employer registration trick is now negative-EV

Tell candidates this out loud. If any registration for the same beneficiary is filed at Level 1, USCIS treats the beneficiary as Level 1 for the entire selection pool. A candidate who lets three employers register them at Levels 1, 2, and 3 gets one ticket, not six. A decade of "register with everyone" advice just inverted. Sourcing conversations in February and early March 2026 need to include an explicit ask: accept only one registration, at the highest wage-level offer.

The replacement playbook, in numbers

Source H-1B transfers and cap-exempt candidates who are already in the US, and treat overseas hires as either "pay the $100K" or "hire in-country and leave them there." Refolk's index gives the concrete pool sizes for each lane.

Two things stand out. The US-based senior pool is roughly a third the size of the offshore pool, which is why it feels scarce and priced accordingly. And the transfer lane is not lottery-subject at all: change-of-employer H-1B petitions are not cap-subject and not fee-subject when filed as change of status inside the US.

212,695

H-1B transfer sourcing is the alpha channel

Every already-approved H-1B holder at a big company is now dramatically more valuable to a startup than an equivalent OPT candidate. The mechanism is arithmetic:

1. No cap exposure. Change-of-employer petitions are not lottery-subject.
2. No $100K fee. Change of status inside the US is exempt.
3. Faster onboarding. Premium processing gets you an approval in 15 calendar days.
4. Better retention signal. These candidates have already survived one visa cycle; they know what they need from an employer.

The top current employers of US-based senior engineers with H-1B signals in Refolk's index include AWS, DocuSign, Fidelity, Wells Fargo, Capgemini, NICE Systems, and Automation Anywhere. Those are the companies you are poaching from in FY2027, not the campus career fairs you were at in FY2024. Describing that shape in one sentence is what Refolk was built for: keyword-first tools (LinkedIn Recruiter boolean, GitHub bio search) miss most of this pool because visa status is not in the profile text.

STEM OPT candidates whose employer can credibly pay Level II or higher

This is the narrow lane that still works for OPT-to-H-1B conversion. If your Series B or later company can back a Level 3 or Level 4 wage offer for a specialized skill (ML infra, security, compilers, distributed systems), DHS's endorsement of above-role wages gives you legal cover. The catch is compensation-band discipline: you cannot pay one new grad at Level 4 and the next at Level 2 for the same role without inviting a comp review.

Target STEM OPT candidates with:

• A specialty the market pays a premium for anyway, so Level 3-4 is not a stretch
• No prior H-1B registrations at Level 1, which would sink the whole selection
• 24 to 36 months of OPT runway remaining, so a missed lottery is not fatal

Hire-in-country as a first-class option, not a fallback

At typical $150-200K total comp, the $100,000 fee is 50 to 70% of a Year 1 US package. Unfundable at seed and Series A without an enterprise budget. The rational response is to hire in Canada or India and keep the person there. Refolk's index shows roughly 730,275 SWEs and developers based in India plus Canada, which is 3.4x the US senior pool. If the plan was always "bring them over in year two," rewrite the plan: assume they stay where they are, and build the tooling and management practice to make that work.

What to change in your sourcing stack this quarter

Rewrite three things before March 2026: the ICP, the outreach script, and the search queries. The rule change is not a marginal tweak; it moves the center of gravity of the pipeline.

• ICP change: from "F-1 grad, top CS program, will need cap sponsorship next spring" to "US-based, currently on H-1B or L-1, at a company where transfers are common."
• Outreach script change: lead with "I file change-of-status, premium processing, no lottery risk for you." That sentence outperforms equity talk with H-1B holders in year three or four of their stint.
• Search query change: stop optimizing boolean strings for "OPT" and start optimizing for tenure signals at the top H-1B employers. Two years at AWS with a Green Card start date more than 12 months out is the highest-signal profile in the FY2027 market.

The bottom line for FY2027

Demand is not falling. USCIS still filled the 85,000 cap in FY2027 despite the $100K fee, which means the effective per-employer odds actually improve for companies that can play the wage-weighting game. Startups that pull back cede slots to Google, Meta, Amazon, and AWS.

The winning move is not to opt out. It is to shift the pool. Fewer OPT juniors, more H-1B transfers, more specialized STEM OPT candidates at Level 3-4 wages, more in-country hires you never plan to relocate. The 212,695-person US-based senior pool is the spine of that plan, and it is sourceable today.

FAQ

Does the wage-weighted lottery apply to FY2026 registrations already filed?

No. The final rule takes effect February 27, 2026 and applies to the FY2027 H-1B cap lottery in March 2026. FY2026 registrations were selected under the old random-lottery process. If you have a candidate mid-cycle right now, they are on the old rules; every candidate you register in March 2026 or later is on the new ones.

Is the $100,000 fee going to survive the lawsuits?

Unknown, and you should not plan around it disappearing. The U.S. Chamber of Commerce, hospitals, universities, and nonprofits are all plaintiffs, but as of May 2026 no court has issued an injunction and the fee is being collected on every qualifying petition. The proclamation is scheduled to run through September 21, 2026 unless extended or modified. Budget as if it is real through at least the FY2027 cycle.

How do I identify H-1B holders on LinkedIn or GitHub without asking directly?

You cannot verify status from a profile, and you should not try to. What you can identify is likely status from employer history, tenure, and country signals: multi-year tenure at a top H-1B sponsor like AWS, Capgemini, or NICE Systems, a US location paired with a non-US education history, and role levels consistent with H-1B occupations. Refolk's semantic search reasons about those signals together instead of keyword-matching "H-1B" in a bio, which almost no one puts there.

If I only have budget for one change to my FY2027 sourcing plan, what should it be?

Move your top-of-funnel from campus and OPT channels to H-1B transfer sourcing at the seven or eight biggest sponsor employers. That single change captures the largest arithmetic advantage in the new rules (no lottery, no fee, no cap exposure) and does not require you to raise wage bands or rewrite your comp philosophy. Everything else - wage-level tuning, in-country hiring, STEM OPT targeting - is a second-order optimization on top of that.