Body Shop

Illustrative. Factory placement marks the industrialised tigers; India’s question mark is the point. The hard numbers are below.

The honest limit. None of this is automatic or unlimited. The far branches take a green card, then naturalization, then a wait measured in decades, and many never materialize. The aggregate multiplier is debated. But the mechanism holds: the program admits one worker and, over time, a lineage.

Registrations for the FY2024 lottery reached 781,000, and 409,000 of them, more than half, were for people entered multiple times by multiple firms. USCIS moved to one entry per person and added fraud investigations. The next year the multiply-registered pile fell to 47,000; the year after, to under 8,000. It isn’t a clean experiment, two things changed at once, but the collapse of the duplicate pile is hard to read as anything but gaming.

Here is the trap: you can change H-1B employers, but your place in that line is tied to the job you leave, so switching sends you back to the end of it. So nobody leaves. A worker who cannot leave is a worker who can be carried, benched, underpaid, and silent.

On strings, walking after the card

Source · Cato Institute / David Bier: employment green-card waits of India ~134 years, China ~17 years, most of the rest of the world roughly current; the 7% per-country cap is at INA §202(a)(2). Belt speeds are illustrative; the waits are the sourced figures

The gate that lets people in ignores nationality, so it fills with one nationality. The gate that lets them stay counts nothing but nationality, so that same group never gets through. The bench is not a side effect of the indenture. It is the product of it.

Computer occupations are about 64% of approved petitions; every other kind of engineering combined, mechanical, electrical, civil, aerospace, is a distant ~10%. Systems analysis and programming alone are more than half of all beneficiaries. Why software and not aerospace? The staffing-firm model places IT contractors by the thousand; software is portable and project-based in a way a jet engine is not; and export-control rules and security clearances lock foreign nationals out of much of aerospace and defense before the question is even asked.

It earns roughly 79% of its revenue from exports ($224B abroad against $58B domestic), and the United States alone is 60 to 62% of it. At the firm level the tilt is starker still: Infosys books about 60% of revenue in North America and 2.5% in India. The work the H-1B feeds is the work that lives on the American market.

The counterpoint, stated plainly.This is not “India can’t build companies.” India built Jio, UPI, and a deep consumer-internet scene at home. It is the export IT-services industry, the one that supplies the bodies, that lives or dies on access to the United States.

India’s IT and business-services exports earn about $224 billion a year. Its diaspora wires home about $129 billion, the largest remittance inflow of any country on earth, roughly 3.3% of GDP. Against those two streams, the investment flowing in has nearly stopped: gross foreign direct investment still topped $80 billion in FY2024–25, but after Indian firms’ own outflows and repatriated profits, net FDI collapsed to about $0.35 billion, essentially zero. This is an economy powered by exporting services and people, not by attracting capital to build at home.

The honest limit.Not all of India’s services output is the offshore-IT stereotype, much is domestic and non-tradable. The tradable export engine, the part the H-1B actually feeds, is the $224B line.

India did once challenge the United States over the H-1B. In March 2016 it filed WTO dispute DS503, arguing that steep US visa fees and numerical limits discriminated against Indian services suppliers. Then it stopped. The case has sat at consultations, the first and lowest rung of the dispute process, for a decade; India never even requested a panel to rule on it. The reading is not subtle. India collects roughly 71% of all H-1B approvals, and a WTO victory that forced the program open to everyone could put that share at risk. The lobbying that matters here is not to change the pipeline. It is to leave it exactly as it is.

At the top sit crowdsourced self-help blogs, RedBus2US, immihelp, where people swap 221(g) timelines and port-of-entry stories. That is not fraud; it is a community. Below it is a training-and-placement layer: consultancies that teach the staffing vocabulary and advertise “500+ hiring partners.” Legal on its face. Only at the bottom does it turn criminal: bench-and-switch schemes, fabricated client contracts, fake end-clients, worker-paid fees. That bottom layer is real and prosecuted, but it is a subset, and how large a subset is not something anyone can honestly measure.

What the middle tier looks like in the wild

The sting.To catch the fraud tier, ICE built a fake school, the “University of Farmington,” and let it advertise. Of the first 130 people arrested, 129 were Indian nationals; eight recruiters were convicted. It is the sharpest single picture of who the bottom layer serves.

These are documented cases, tied to specific firms and the staffing model, not a claim about every H-1B employer.

The one time the government pulled a random sample and audited it, a 2008 USCIS compliance assessment of 246 petitions, it found 13.4% involved fraud and another 7.3% had technical violations; the rest were compliant. That is a small, dated sample, and it does not say most H-1B petitions are fraudulent. But it does not say the problem is imaginary either. Two things sharpen it: enforcement since has concentrated on exactly the IT-staffing firms this dossier keeps returning to; and when USCIS finally required one registration per person in 2024, duplicate lottery entries collapsed from 408,891 to 7,828, in the agency’s own words “a reduction in misuse.”

In 2015, employees at Disney and Southern California Edison were laid off and, on the way out, made to train the H-1B contractors who replaced them. Those cases are vivid but rare; the systematic version is quieter and lives in the pay. In litigation, HCL’s own internal documents showed the company calculating that it paid its visa workers 47 to 64% below the market rate for the role. The best causal estimate, from an economics study using the H-1B lottery itself as a natural experiment, finds each additional H-1B hire is associated with about 1.5 fewer other workers at the firm, with some downward pressure on earnings and higher firm profits. This is not the whole labor market, and honest economists disagree about the economy-wide wage effect. But at the firm level, in the staffing-heavy corner of the program, the discount is documented, and it is the point.

Two more numbers give it a shape. In the same years the industry shed hundreds of thousands of workers, it kept hiring new visa holders by the hundred thousand. And the wage floor the law calls “prevailing” is set, by design, below what the job actually pays.