"Do STEM for the visa runway" just inverted

Why the old advice worked

An F-1 student who graduates can work for 12 months on Optional Practical Training. A graduate of a designated STEM program can extend that by 24 months. That's up to 36 months of authorized work, which meant up to three attempts at the H-1B lottery before the clock ran out.

Under the old lottery every registration had the same odds. More attempts meant better cumulative odds, and STEM was the way to get more attempts. The advice followed directly. It wasn't a folk belief. It was arithmetic.

What changed

The Department of Homeland Security published a final rule on December 29, 2025, effective February 27, 2026, in time for the FY2027 registration window of March 4 to 19. The random lottery was replaced with weighted selection based on the Department of Labor's four prevailing wage levels for the occupation and location. A Level I registration gets one entry in the pool. Level II gets two, Level III three, Level IV four.

Wage level reflects seniority and complexity within an occupation in a specific area. It's a relative position, and a salary that looks high in absolute terms can still sit at Level I or II. A new graduate's first job is Level I or II almost by definition.

So under the new system a new graduate has one entry against a senior candidate's four. Still eligible. Structurally behind.

The $100,000 fee announced in September 2025 got far more coverage and matters less here. USCIS guidance on October 20, 2025 confirmed the fee doesn't apply to change-of-status petitions for people already in the United States, including F-1 students on OPT or STEM OPT. The fee lands on employers bringing workers in from abroad. The lottery weighting lands on your students.

Why STEM specifically inverted

STEM's advantage was attempts. The weighting makes attempts matter less than wage level. Three Level I attempts give three total entries over three years. One Level III registration gives three entries in a single year. The senior candidate gets there first, and the student burns their runway waiting.

At the same time, the entry-level market STEM graduates were counting on has thinned. Indeed Hiring Lab put software development postings about 27.5% below their pre-pandemic level as of early 2026, with senior roles making up 69.3% of what remained. A STEM graduate now faces a weighted lottery and a market where the junior roles that would have qualified them for it are scarcer.

The major that maximized lottery attempts now points students at the part of the market with the fewest Level I openings and the least chance per registration. That's the inversion.

What the advice becomes

Employer choice matters more than major choice. Universities, nonprofits affiliated with universities, and nonprofit research organizations are cap-exempt. They file H-1B petitions outside the lottery entirely. A student who wants to stay in the United States has better odds with a cap-exempt employer in any field than in a Level I lottery registration at a tech company.

Wage level is negotiable in a way a major isn't. A Level II offer doubles the entries. Students should know the OEWS wage levels for their occupation and metro before they accept anything, and should ask employers directly what level they'll register at. "Do you sponsor?" is no longer the right question.

Geography changes the level. The same salary can be Level II in one metro area and Level I in another, because the levels are set locally. Where a student takes the job affects their odds.

Fewer, better-targeted employers beat volume. The application-count strategy never worked well for international students, and it works worse when the outcome depends on a specific employer's wage decision.

Don't leave the country while a change-of-status petition is pending. Departure can cause USCIS to treat the petition as abandoned, convert it to consular processing, and trigger the fee. This is the kind of detail that ends up costing a student their path.

And the honest version: for some students the math now says the United States is no longer the likely destination, and a career center should be able to say so early enough for them to plan. That's a harder conversation than "do STEM." It's also more useful.

What would change this reading

Both the fee and the weighted rule face litigation. An injunction would restore the old odds. If registrations fall enough that the cap isn't reached, weighting stops mattering. If USCIS publishes selection rates by wage level and Level I registrations are selected at rates close to the old lottery, the disadvantage is smaller than the entry counts imply. None of those has happened.