US Tech Layoffs Parallel High H-1B Issuance to India
A set of numbers mentioned in policy debates states that since 2022, over 700,000 tech jobs have been cut in the US, while approximately 670,000 H-1B visas have been issued to Indian nationals during the same period. An executive order from the White House on September 18 used different metrics: tech employers are projected to lay off between 800,000 to 1.3 million Americans from 2022 to 2026, while applying for hundreds of thousands of H-1B visas. These two sets of figures are not published by the same institution under the same definitions; layoffs include multiple rounds of restructuring and contract workers, while the visa figures largely consist of renewals, employer changes, and amended petitions, which do not equate to an equal number of new entrants replacing them.
The verifiable official data is in percentages rather than those integers. USCIS shows that for fiscal year 2025, approximately 406,300 H-1B petitions were approved, with about 283,800 beneficiaries from India, accounting for about 70%, and around 49,200 from China, about 12%; computer-related occupations accounted for about 62%. Indian nationals have received about 70% of approvals in recent years, with consular issuance reporting 72.3% from 2022 to 2023. In the first nine months of fiscal year 2026, approximately 273,000 approvals for continued employment have already been granted, nearing last year's total record, indicating that channels outside the cap are still operational. Tracking sites like layoffs.fyi show that over 100,000 tech jobs have been cut in 2026, but it is not possible to directly translate this into the number of H-1B holders.
New applications from Indian IT service providers have significantly contracted: the applications from 18 listed service providers plus Cognizant dropped from about 36,500 in 2024 to about 8,160 in 2025; initial employment approvals for seven major Indian companies have drastically declined from their peak in 2015, with initial approvals for TCS, Cognizant, and others falling to three or low four digits in fiscal year 2025. US tech companies still lead in initial approvals, with Amazon, Meta, Microsoft, Google, and others continuing to submit new applications. Xpheno statistics indicate that about 15,100 Indian tech workers returned to India from the US in 2025, with about 7,300 in the first half of 2026, facing a 28-month low in Indian tech recruitment upon return.
Mechanically, what is being compared are two simultaneous flows: the contraction of US tech positions and the dominance of Indian nationals in the H-1B stock. The event-driven aspect is the executive order incorporating layoffs into the review process, and public opinion framing these two large numbers as a causal relationship. Capital is shifting from US on-site positions to Indian global capability centers and offshore delivery, with new application costs inflated by $100,000 in litigation fees and weighted salaries; benefiting are those already in the country with the ability to extend their visas, while newcomers from abroad and service providers relying on on-site presence are under pressure. The 700,000 versus 670,000 cannot be treated as a one-to-one replacement, but it is enough to turn "supplement or replace" into a political excuse for the Department of Labor to trace old applications.
Source: Public Information
ABAB AI Insight
Putting the total number of layoffs and the total number of Indian visas side by side is political arithmetic, not labor statistics. H-1B petitions can appear multiple times for one individual over the years, with renewals accounting for about 70%; the 670,000 figure is closer to cumulative paper approvals rather than 670,000 new immigrants filling 700,000 vacancies. The White House's 800,000 to 1.3 million also sums employer layoffs, including market, sales, and contract workers whose skills do not match H-1B positions. The observable structure is: new applications from Indian outsourcing are declining, while initial approvals from large US firms remain, and the stock of extensions is at a record high, with some on-site work returning to capability centers in Bangalore and Hyderabad.
The capital path is "US layoffs—India expanding centers—visas shifting from new applications to extensions." Even if the $100,000 offshore fee is banned, the weighted salaries have made low-wage on-site models more expensive, with service providers keeping people on existing H-1Bs to shift projects. The return wave coinciding with a low point in local Indian recruitment indicates that job movement is a global restructuring of positions, not an automatic opening of gaps in India due to US shortages. The executive order requires a review of layoffs in "similar positions," and the next phase of debate will focus on whether job levels are similar, rather than nationality percentages.
This is analogous to the H-1B debates following the internet bubble in the 2000s, and the UK's scrutiny of skilled worker visas alongside local unemployment: both use two sets of totals to imply replacement. The difference is that this round of AI restructuring coincides with visa fee and lottery reforms in the same year, with outsourcing firms first being cut from new applications while product companies maintain core teams through extensions. The industry phase has shifted from "Indian pipelines supplying talent to Silicon Valley" to "Silicon Valley's contraction, with pipelines redirecting to domestic stock." Whoever can prove that the laid-off positions and the applied positions are not from the same skill group will still be able to apply for labor condition applications; whoever replaces with the same job level will enter the wage and hour division's review list.
Structural judgments belong to the overlap of regulatory changes and industry chain restructuring. Pricing power shifts from the illusion of nationality quotas to employer labor records and wage percentiles. The mechanism is that the visa stock can continue to be extended during the layoff cycle, but public opinion demands a cutoff of new applications based on total flow; thus, policy uses layoff tables as a dam, rather than setting caps based on country of birth. The juxtaposition of numbers changes the review temperature, not the arithmetic identity.