The $103,265 H-1B Fee Makes "Already in the US" Your Best Filter

On August 25, 2026, DHS proposed a $103,265 fee on every FY2028 cap-subject H-1B petition. The original $100,000 proclamation from a year ago is scheduled to sunset (or renew) on September 21, 2026, and DHS has warned the two could stack past $200K per hire. If you are still ranking candidates by school and stack, you are leaving six figures per offer on the table.

What actually changed on August 25

DHS filed formal rulemaking for a new $103,265 fee on every cap-subject H-1B petition, on top of every existing fee, with public comments open through September 24, 2026. It is the department's workaround after the First Circuit denied a stay on July 24, 2026, keeping alive the Massachusetts district court order in State of California v. Noem that had vacated the earlier $100,000 proclamation as an unlawful tax.

The mechanics that matter for sourcing:

• The fee is triggered at petition filing, not at registration. Employers can register, get selected, then walk away.
• It applies to consular notification for a visa abroad, and it is due at petition filing whether the employer wants consular routing or not.
• It does not apply to cap-exempt petitions (universities, nonprofit research orgs, government research orgs).
• It does not apply to extensions, amendments, or H-1B transfers between employers.
• The same carve-out logic that exempted F-1 change-of-status filings under the $100K proclamation extends here: individuals already present in the US, changing status, are outside the fee.
• Nothing changes for O-1, L-1, TN, E-3, H-1B1, J-1, or E-1/E-2.

Whichever regime survives, the operational reality for a recruiter is the same: a candidate already inside the US, in a status that supports change-of-status, is worth roughly $100K to $200K more than an identical candidate abroad.

Why "already in the US" is now your highest-leverage filter

Physical presence in the US in a change-of-status-eligible status (F-1 OPT, STEM OPT, L-1, TN, O-1, E-3) is now worth more per hire than seniority, school, or stack. That is not hyperbole. It is arithmetic: two identical résumés, one triggers up to $203,265 in stacked fees at petition filing, the other triggers $0.

The mechanism is the carve-out. A change-of-status petition converts a person already lawfully present (say, an F-1 STEM OPT graduate) into H-1B status without a consular trip. Both the vacated $100K proclamation and the proposed $103,265 rule explicitly exempt these. A cap-subject petition for a candidate abroad, whether the employer chooses consular notification or not, sits inside the fee.

The practical rewrite of your Boolean this fall: presence flag first, everything else second. That means promoting fields most ATS systems bury (visa status, work authorization free text, city vs. country of residence) to the top of your ranking.

This is the exact re-ranking problem Refolk is built for. You describe the person in plain English ("entry level backend engineers currently on F-1 OPT in the US, ideally with STEM OPT extension eligibility, open to change-of-status H-1B in April") and get a ranked shortlist back across GitHub, LinkedIn, and the open web, with the presence flag actually weighted instead of buried on page four.

The scarcity nobody has priced in

The pre-filtered "already in the US" pool is not large, and it is not what your candidate database will surface by default. In Refolk's index of professional profiles, the ratio of India-based entry-level software engineers to US-based entry-level engineers who explicitly self-tag "OPT" is roughly 7,063 to 1.

7,063x

That 79 is not the true universe of US-based OPT engineers, of course. It is the number who make it easy to find. The rest are inferrable through indirect signals: graduation date within the last 36 months from a US institution, US city on LinkedIn plus non-US citizenship signals in a GitHub bio, prior J-1 or F-1 language in a personal site. Most sourcers stop at the self-tag and miss the rest.

Here is what the segments look like with the fee stack applied:

The math tells you why the offshore-bulk playbook is collapsing on its own even before your pipeline changes. A Level I entry-level offer to a candidate abroad now runs the gauntlet of a ~15% weighted lottery, a $100K supplemental fee already targeting large overseas outsourcing firms on registrations from abroad, and a proposed $103,265 filing fee if selected. TCS, Infosys, Cognizant, Wipro and HCL have already visibly cut their registration volumes.

The weighted lottery kills the cheap-junior play

DHS replaced the random H-1B cap lottery with a wage-weighted lottery that favors candidates paid at higher Department of Labor OEWS wage levels, which pushes entry-level Level I offers to roughly 15% odds versus 46% at Level III. That single change punishes the "cheap junior from India" model twice: worse odds on the front end, six-figure fee on the back end.

For a boutique sourcer or a founder competing against staffing giants for the same juniors, this is actually good news. The people you want are:

1. Level II or Level III wage candidates already inside the US on F-1 OPT or STEM OPT (better odds, no fee).
2. Level III candidates on L-1 who can convert (better odds, no fee, and L-1 is untouched anyway).
3. Anyone on TN (Canadian/Mexican), E-3 (Australian), or H-1B1 (Chilean/Singaporean). These categories do not touch H-1B at all.

For the L-1 slice, a good workflow does not stop at F-1. It rolls straight into "senior engineers at Google Bangalore, Microsoft Hyderabad, Amazon India who moved to a US office in the last 18 months," because those are your L-1 candidates ready to transfer.

Cap-exempt is the arbitrage nobody talks about

Cap-exempt H-1B recruiting (universities, nonprofit research institutes, government research organizations) is fee-free under both regimes and lets a private employer subsequently run a concurrent H-1B without paying the cap fee. This is the play biotech has used quietly for years and that tech recruiters, outside a handful of AI labs, still ignore.

Refolk's index carries roughly 23,593 US-based Research Scientist and Postdoctoral profiles, concentrated at employers like Google, Columbia University, MIT Media Lab, NIH, Stanford, Fred Hutch, Broad Institute, and Scripps Research. Every one of those university and nonprofit institutions is a cap-exempt H-1B sponsor. The mechanic:

1. A US research institution files a cap-exempt H-1B for the person. No cap, no lottery, no $103,265 fee.
2. A private-sector employer files a concurrent H-1B for the same person for a second, unrelated role.
3. Because the underlying H-1B status came through a cap-exempt petition, the concurrent private-sector petition is also outside the cap and outside the fee.

You need real coordination with an immigration firm (Fragomen, Ogletree Deakins, and Murthy Law Firm have all published on the new rule and understand the concurrent structure), but the population is 23,593 people who never appear in an ATS search for "software engineer."

Two identical résumés. One triggers $203,265 at filing. The other triggers zero. That is now the delta.

The "selected but never filed" cohort DHS is about to create

Because the proposed fee is due at petition filing rather than at registration, FY2028 will almost certainly produce a large cohort of candidates who were selected in the lottery and then dropped when their sponsoring employer refused to write the check. That cohort is a sourcing goldmine.

The play, once March/April 2027 numbers are out:

• Track lottery selection announcements at your top offshore-heavy targets.
• Cross-reference against actual H-1B petitions filed with USCIS (public LCA data lags but surfaces in about 60-90 days).
• Every name selected but not filed is a candidate who has just lost their sponsor and needs a new one before their existing status runs out.

This is the exact "list of people the market just dropped" shape that separates real work from spray-and-pray outreach. The window is short, the candidates are pre-qualified (they were selected in a weighted lottery favoring higher wages), and most of them are already in the US on F-1 OPT.

The change-of-status vs consular decision, in one table

The change-of-status vs consular choice is the operational lever that decides whether a hire costs you $0 or $203,265 in fees. Both routes yield the same H-1B status. Only one triggers the stack.

What to do in the next 30 days

Rank your open reqs by whether the person you need can realistically come from the OPT, cap-exempt, or alternate-visa pool, and rebuild the sourcing list under each. Concretely:

1. Pull your last 12 months of H-1B petitions. Segment by change-of-status vs consular. The consular column is your new savings target.
2. For every open req that historically went H-1B-abroad, source three parallel lists: F-1 OPT/STEM OPT in the US, L-1 candidates at your target companies' non-US offices, and TN/E-3/H-1B1 candidates.
3. Add university and nonprofit research employers to your target list for any research-adjacent role. The 23,593 US Research Scientist and Postdoc profiles in Refolk's index are the pool nobody in tech is working.
4. Set a March 2027 calendar reminder to work the "selected but never filed" cohort at TCS, Infosys, Cognizant, Wipro, and HCL.
5. File public comments on the DHS rule by September 24, 2026 if the change-of-status carve-out matters to your business. It is currently in the proposed text; comments determine whether it stays.

The offers you make this fall will land in a market where an identical candidate abroad costs you six figures more. Re-rank now.

FAQ

Does the $103,265 fee apply if my candidate is already in the US on OPT?

No. Both the vacated $100,000 proclamation and DHS's proposed $103,265 rule carve out change-of-status petitions for individuals already present in the US, and the same carve-out logic that exempted F-1 change-of-status filings under the proclamation extends to the new rule. An F-1 OPT or STEM OPT candidate converting to H-1B inside the US is outside the fee. A consular filing for the same person abroad is not.

When does the new fee actually take effect?

Not yet. DHS filed a proposed rule on August 25, 2026, with comments open through September 24, 2026. DHS must review comments and issue a final rule before the fee applies. Separately, the original $100,000 proclamation is scheduled to expire on September 21, 2026 unless extended, and it is currently paused by the First Circuit anyway. The uncertainty is precisely why "already in the US" is the safest filter to optimize on right now.

What about O-1, L-1, TN, and E-3 candidates?

Untouched by both regimes. The $100K proclamation and the proposed $103,265 rule apply only to H-1B. A Canadian senior engineer on TN, an Australian on E-3, a Chilean or Singaporean on H-1B1, an intracompany transfer on L-1, and an extraordinary-ability hire on O-1 are all outside the fee entirely. Post-September 21, these categories are not just "alternative visas," they are structurally cheaper hires than a cap-subject H-1B from abroad.

How does the weighted lottery interact with the fee?

It compounds the effect. DHS replaced the random lottery with a wage-weighted lottery favoring higher OEWS wage levels, so Level I entry-level offers now have roughly 15% selection odds versus roughly 46% at Level III. A Level I offer to a candidate abroad now runs bad odds, a $100K supplemental fee already targeting large overseas outsourcing firms on registrations from abroad, and a proposed $103,265 filing fee. The offshore-bulk playbook TCS, Infosys, and Cognizant ran for a decade is being priced out of existence.