DHS Wants Universities to Pay $70,000 Before a Student Can Work. The Comment Window Closes in Four Weeks.

Most immigration fee proposals are a line item. This one is a staffing model.

On October 8, 2026, the Department of Homeland Security published a proposed rule that would charge $70,000 for an F-1 student's first Optional Practical Training recommendation — pre-completion or post-completion — and $30,000 for every subsequent recommendation, explicitly including the STEM OPT extension. A single student who completes standard post-completion OPT and then the 24-month STEM extension would generate $100,000 in fees.

The fee is not charged to the student. It is charged to the SEVP-certified school, and it must be paid before a designated school official can enter the OPT recommendation into SEVIS. No payment, no recommendation, no work authorization.

Comments are due November 9, 2026 under Docket No. ICEB-2026-0100 at regulations.gov. (Thirty days from October 8 lands on Saturday, November 7; some early analyses cite that date. Treat November 7 as your internal deadline and November 9 as the outside edge.) A separate set of Paperwork Reduction Act comments on the information collection is due December 7, 2026 — a second, less-contested channel that most commenters will miss.

Why this is a grants story

It is tempting to file this under immigration and move on. Do not, for three reasons.

First, OPT is research labor. A large share of post-doctoral and technical staff on federally funded projects at U.S. institutions pass through OPT or STEM OPT on their way to H-1B or permanent status. Those salaries are charged to grants. A fee that gates OPT gates the hiring pipeline for work already budgeted, already awarded, and already on a performance schedule.

Second, the fee is almost certainly unallowable as a direct charge. Nothing in the proposal establishes a cost principle for it, and a six-figure per-person immigration fee does not map cleanly onto any existing allowable cost category under 2 CFR Part 200. Which means it lands in institutional funds — the same pool already absorbing indirect-cost pressure. An institution that currently employs forty people on OPT across its research enterprise is looking at a potential annual exposure in the low millions that no sponsor will reimburse.

Third, DHS says the payer can shift. The agency explicitly acknowledges schools "may pass the financial obligation of this proposed fee onto F-1 nonimmigrant students, all students, or employers." That sentence is the entire policy design. DHS is not deciding who bears the cost; it is creating a $70,000 obligation at the institution and letting institutional politics allocate it. The three options are: absorb it, charge the student (which prices OPT out of reach for nearly everyone), or charge the employer (which, for grant-funded research, means charging the project — see the cost-principle problem above).

There is a fourth option DHS also acknowledges, and it is the one with the largest effect: schools may simply decline to pay, which ends OPT at those institutions by administrative choice rather than by rule.

The numbers DHS put on the record

The proposal's own economic analysis is unusually direct about scale:

• 2,478 SEVP-certified schools affected.
• $8.4 billion to $16.5 billion in annual fee revenue, with a central estimate around $12.4 billion annually after the first year.
• Fees deposited to the U.S. Treasury.
• Statutory hook: INA 214(a), with fraud-prevention and worker-protection rationales.
• Curricular Practical Training (CPT) is excluded. No other exemptions are detailed.

Those revenue projections are the most useful paragraph in the document for anyone drafting a comment, because they are internally revealing. A $12.4 billion annual estimate implies DHS expects participation to continue at close to current volumes. The same document also concedes that the fee could reduce OPT participation. Both cannot be true at once, and the gap between them is the analytical soft spot in the rule.

The grandfathering clause creates a window

The rule, as proposed, takes effect 60 days after publication of a final rule — not 60 days from now. And it does not reach backward:

• Students currently on approved OPT are not subject to the fee.
• Students approved before the effective date are not subject to the fee.
• Students with pending recommendations at the effective date are not subject to the fee.
• A student grandfathered on initial OPT would still owe $30,000 for a later STEM extension requested after the effective date.

That last line is the one to plan against. The grandfather protects the recommendation you have already issued; it does not protect the next one for the same student.

The practical consequence is a defined, closing window. The earliest a final rule could take effect is roughly 60 days after a final rule publishes, and final rules following a 30-day comment period rarely appear in under a few months. Between now and then, every OPT recommendation and every STEM extension recommendation that a DSO can legitimately enter into SEVIS is a recommendation that escapes the fee permanently.

If you are an institution: inventory every F-1 student who is eligible, or will become eligible, for an initial OPT recommendation or a STEM extension in the next two to three terms, and determine which of those recommendations can be lawfully issued now rather than later. This is not a loophole; it is the rule's own transition provision operating as written. The limiting factor is eligibility timing, not willingness.

If you are a grant-funded employer: identify which of your current staff will need a subsequent OPT recommendation during the project period, and flag those positions in your continuation planning. A $30,000 fee attached to a renewal you assumed was administrative is a budget event.

How to write a comment that counts

Mass-identical comments do not move rules; data does. The sections of this proposal most exposed to a well-sourced objection:

1. The payer-shifting assumption. DHS assumes cost pass-through is feasible. If your institution's own analysis shows the fee cannot be lawfully charged to a federal award and cannot be charged to students at any realistic tuition level, say so with the actual accounting, and state plainly that the operational outcome is ceasing OPT recommendations rather than paying.
2. The participation-versus-revenue contradiction. Quantify it from your own numbers: how many OPT recommendations your institution issued in the last three years, and how many you would issue at $70,000 each. If the honest answer is zero, that is a $12.4 billion revenue estimate failing at your institution specifically.
3. Federally funded research impact. Name the awards. "Three NIH R01s and one NSF CAREER award currently employ staff on STEM OPT" is a sentence that enters the administrative record with evidentiary weight. Agency-funded project disruption is a cognizable harm; abstract concern is not.
4. The December 7 PRA channel. Burden-estimate comments on the information collection are a separate, lower-traffic docket. If the fee-collection mechanism imposes tracking and reconciliation work the estimate does not capture, that belongs there.

Submit under ICEB-2026-0100. Institutional comments carry more weight than individual ones, and comments that include a specific dollar figure, a specific headcount, and a specific award number carry more weight than either.

Where this sits in the broader pattern

This is the third major fee action aimed at employment-based immigration in the research sector this year, and the trajectory is consistent.

The $100,000 H-1B payment proclamation was extended through September 2027 and, critically, did not carve out cap-exempt universities and nonprofit research organizations — we covered that in the H-1B proclamation extension. Separately, DHS proposed a $103,265 H-1B fee in August, analyzed in that proposed rule's treatment of cap-exempt institutions. Both the $100,000 payment and its successors have drawn litigation; two federal district courts — in Massachusetts and, as of October 1, California — have blocked the $100,000 fee, and DHS's proposed $103,265 fee reads in part as a response to those rulings.

The OPT proposal closes the remaining door. An institution that cannot afford the H-1B fee could previously keep a researcher working on STEM OPT for up to three years. Under this rule, that bridge costs $70,000 to build and $30,000 to extend.

Expect litigation here too — the statutory authority for a fee of this magnitude under INA 214(a), and the adequacy of a 30-day comment period for a rule with a $12.4 billion annual effect, are both contestable. But litigation is slow and the comment window is four weeks. Do the inventory, issue the recommendations you can issue now, and file a comment with numbers in it.