A Department of Homeland Security proposed rule published in the Federal Register on 8 October 2026 would charge international university graduates US$70,000 for their first period of Optional Practical Training — the work authorisation that lets F-1 student visa holders take a job in their field after graduating — and $30,000 for any subsequent period. The rule is a proposal, not law: comments close 9 November 2026, and no fees take effect before a final rule.
DHS frames the fees as fraud control: “to combat fraud, strengthen the integrity of the immigration system, and protect U.S. workers,” per the proposal’s abstract (91 FR 64566). Schools, not students, would be billed for each participating graduate.
🔍 THE BOTTOM LINE
This is a proposal, and its opponents are already lawyered up. But for anyone tracking the AI-era tech workforce, the interesting part is the collision it sets up: the same week that firms like Anthropic put $100 million into training 10,000 engineers and Fortune 500 boards report AI builder roles doubling as a share of tech demand, a new cost barrier sits directly on one of the main pipelines feeding those roles — international graduates, who make up roughly 75% of full-time graduate enrolment in AI-related fields, by one count.
The rule, and the timing
Under the proposal, an “initial OPT” authorisation costs $70,000 per F-1 student; a second or later authorisation costs $30,000. Comments must be received by 9 November 2026, and the regulatory process alone gives the industry and universities months to respond.
The timing is pointed. On 1 October, a judge struck down the latest version of the $100,000 H-1B filing fee — a separate instrument. Publishing the OPT rule a week later while that case is pending reads to immigration analysts like policy sequencing: if a fee at the border falls in court, place the next one at the campus gate.
The announcement landed the day before the Microsoft and Adobe PERM suspension — a two-part message to skilled-migration-dependent industries in the same news cycle. (The AP report on that announcement noted the OPT proposal alongside it.)
Why the tech job market is exposed
OPT is the bridge from student visa to employer sponsorship. Most F-1 graduates get roughly 12 months of work authorisation, extended for STEM graduates — the population that populates AI-adjacent master’s programmes. Making the first job after graduation cost an employer or university $70,000 compresses exactly the entry-level rung that research has already flagged as thin: the quiet crisis in entry-level hiring, the 227,000 entry roles a year now not opening.
If the rule takes effect as written, the arithmetic is unforgiving. A $70,000 charge against an entry-level salary in the $70,000–$120,000 range is a doubling of the effective first-year cost of an international graduate hire. Some employers will pay it; more will reroute to remote hires in other countries, or to domestic graduates — which is partly the policy’s stated aim (“protect U.S. workers”).
The complication, cited by opponents including universities and immigration researchers, is that “protecting” domestic hiring by taxing degree-holders who train in US universities assumes those graduates are substitutes for domestic ones. In AI-adjacent fields, employers keep saying the opposite: demand outruns domestic supply, which is why the visa exists at all.
The counter-argument, stated plainly
It’s possible the rule works as designed. A large fee on a low-value use of OPT could deter diploma-mill enrolments and sham applications that exploit the pathway, tightening the channel for genuine skills matches — precisely the “fraud” the proposal names. Whether a $70,000 price discriminates between fraud and genuine hiring, or just prices everyone the same, is the empirical question the comment period will test. Universities and the tech industry are expected to argue the latter in submissions.
NZ angle
New Zealand runs a similar bridge: the post-study work visa lets international graduates stay and work for up to three years, no fee beyond the standard application cost, aimed squarely at the same “retain the talent we trained” logic. NZ job ads asking for AI skills have more than doubled in a year, and NZ tech employers are competing for the same globally mobile graduates — many of whom would have shortlisted US programmes first.
If the US narrows its graduate pipeline, two scenarios matter here:
- Diversion: graduates who might have taken an OPT job in San Jose consider Auckland, Sydney, Toronto or London instead — and NZ’s post-study pathway has no $70,000 toll gate. NZ recruiters told RNZ the flood of AI-written applications already makes sorting hard; a diverted cohort of trained graduates is the upside case.
- Third-country effect: nothing stops an AI-capable graduate who can’t afford US fees from taking a remote role for a US employer from anywhere, or from India’s expanding AI deployment market, which added hundreds of thousands of AI-era roles this year.
Either way, the pricing of a talent pathway in the world’s largest tech labour market is now an input into NZ’s own hiring plans.
FAQ
What is OPT? Optional Practical Training is US work authorisation for F-1 international students, letting them work in their field of study for up to 12 months after graduating (longer for STEM graduates), usually en route to employer-sponsored visas.
Is the $70,000 fee in effect? No. It is a proposed rule (Federal Register 2026-20660, published 8 October 2026). Comments close 9 November 2026; a final rule would follow, and legal challenges are considered likely by immigration analysts.
Does this affect AI jobs specifically? Not by name — the fee applies to all F-1 graduates. But international students make up a majority of graduate enrolment in AI-related US fields by most counts, so AI-adjacent employers are among the most exposed if it proceeds.
Does NZ have an equivalent fee? No. NZ’s post-study work visa carries standard application fees only (currently measured in hundreds, not tens of thousands, of dollars).
— CJ Murden, editor of Singularity.Kiwi. Former digital technologies teacher, author of AI-focused books. Writing with a New Zealand focus.
