Cato's analysis of USCIS internal data shows a 94% year-over-year collapse in Q1 2026 green card approvals (7,000 vs. 280,000 in Q1 2025) across every category. They argue this happened without any statutory change — DHS unilaterally added discretionary review layers, redirected adjudicators to enforcement vetting, and stopped processing pending I-485 cases.
By submitting the Cato piece to Hacker News, malshe surfaces the argument that a foundational immigration pipeline has been shut down by executive discretion rather than legislation. The 114-point score and 107 comments suggest the framing resonated with the developer community.
The editorial argues that two decades of tech labor strategy — hire on H-1B, file PERM, get I-140 approved, convert to permanent residence around year 5-7 — has been quietly broken at the adjustment-of-status step. I-140 approvals remain normal, but the final DHS-controlled step has collapsed, leaving the entire pipeline functionally inoperable.
The editorial emphasizes that the real problem isn't merely longer waits — it's that H-1B extensions past the six-year cap depend on I-140 approval and a functioning adjustment process. When that process becomes opaque and discretionary, employers and workers can no longer plan careers, relocations, or hiring around any reliable timeline.
The Cato Institute, citing USCIS internal data, reports that the Department of Homeland Security has effectively stopped approving green cards. In Q1 2026, USCIS issued roughly 7,000 lawful permanent resident approvals — compared to approximately 280,000 in Q1 2025, a 94% year-over-year collapse. The drop spans every category: family-based, employment-based (EB-1 through EB-3), diversity lottery, and humanitarian.
This isn't the result of any new statute. Congress hasn't changed the underlying immigration law, and the statutory annual caps (140,000 employment-based, 226,000 family-based) remain on the books. What changed is administrative posture. According to Cato, USCIS added new layers of discretionary review to adjustment-of-status filings, redirected adjudicator time toward enforcement-adjacent vetting, and effectively stopped processing the bulk of pending I-485 cases. The backlog isn't growing because demand spiked — it's growing because the agency stopped clearing it.
The Cato piece notes that even employment-based categories explicitly designed for high-skilled workers — EB-1A (extraordinary ability), EB-2 NIW (national interest waiver), and EB-3 (skilled workers, including most H-1B-to-green-card conversions) — have ground to near-zero approvals. Approval rates for I-140 petitions (the employer-sponsored predicate to a green card) remain normal; the chokepoint is the final adjustment-of-status step, which DHS controls unilaterally.
The tech industry's labor model has quietly depended on a specific pipeline for two decades: hire on H-1B, file PERM in year 2-3, get I-140 approved, then convert to permanent residence somewhere around year 5-7. That pipeline is now broken at the final stage, and it broke without a single vote in Congress.
The strategic problem isn't the slowdown — it's the unpredictability. H-1B holders can extend in three-year increments past the normal six-year cap only if their I-140 is approved or their adjustment is pending. So workers currently in extension territory aren't getting deported tomorrow. But the implicit deal — "stay loyal, we'll get you your green card" — is now a deal the employer cannot honor on any predictable timeline. Workers know this. Recruiters in Toronto, London, and Berlin know this even better.
Compare the numbers. Canada's Express Entry processed roughly 110,000 economic-class permanent residents in 2025 with a median adjudication time under six months. The UK's Global Talent visa moves in weeks for endorsed applicants. Even Germany's Blue Card, historically bureaucratic, now clears in 90 days under the 2024 Skilled Immigration Act reforms. A senior engineer who started the US green card process in 2020 may now wait longer than they would have spent earning permanent residence in three other G7 countries combined.
Community reaction on Hacker News (114 points, several hundred comments) split roughly into three camps. The largest cohort — software engineers currently in the pipeline — described stalled paperwork, frozen relocations, and quiet job searches in Canadian subsidiaries of their employers. A second group framed it as a deliberate policy outcome regardless of statute: "the Overton window for immigration enforcement now includes administrative shutdown." A third, smaller faction argued the numbers will rebound once a backlog-clearance push happens — though Cato's piece argues there's no evidence such a push is planned.
The second-order effect is on hiring. Sponsoring a new H-1B in 2026 used to be a 5-7 year commitment with a permanent-residence payoff; now it's an open-ended commitment with no defined endpoint. That math changes which roles companies are willing to sponsor at all. Expect to see more "US citizens or permanent residents only" listings, more remote-Canada postings from US companies, and more aggressive use of L-1 intracompany transfers (which DHS hasn't restricted at the same rate) as a workaround.
If you manage engineers on visas, three things change immediately. First, retention math. The non-cash component of compensation for visa-pipeline employees just lost most of its present value. If you were relying on "we're sponsoring your green card" as a retention tool — particularly for senior ICs and tech leads — you need to replace that with cash or equity, and you need to do it before your competitors do. Plan for 15-25% comp adjustments on affected employees or accept the attrition.
Second, hiring geography. If you're a US company and you've been resisting opening a Canadian or UK entity, the calculus just flipped. The cost of a Toronto subsidiary is amortized fast when you can hire the same engineers there in 90 days versus never. Companies like Stripe, Shopify, and Snowflake already operate this playbook; mid-size companies that haven't will increasingly need to. Watch for a spike in PEO services (Deel, Remote, Oyster) handling "US company, Canadian engineer" arrangements through 2026.
Third, succession planning. Any senior engineer, staff engineer, or principal whose long-term US status depends on adjustment of status is now a flight risk on a timeline you don't control. If your architecture roadmap assumes a specific senior IC will be around in 18 months, build a backup. Document tribal knowledge. Pair junior engineers harder. Treat visa-pipeline seniors the way you'd treat anyone with a competing offer in their inbox — because effectively, every employer in Canada now has one.
The political question is whether this is policy or process — whether a future administration could reverse it quickly, or whether the institutional decisions (staffing reallocations, procedural rule changes) take years to undo. Cato's reporting suggests the latter: "administrative shutdown" doesn't restart with a memo. For engineering leaders, planning for the realistic scenario — that US permanent residence remains effectively unavailable for new applicants through at least 2027 — is the more defensible bet. The companies that adjust their hiring geography and retention strategy this quarter will look prescient by year's end. The ones that wait for clarity from Washington will lose their senior engineers to Toronto first.
Top 10 dev stories every morning at 8am UTC. AI-curated. Retro terminal HTML email.