H-1B Sponsorship: Chasing It Broadly Is Why You're Not Getting Sponsored

H-1B sponsorship hasn't disappeared, it has concentrated: a small set of employers files at scale, in a handful of roles, and from FY2027 the lottery favors higher wage bands. So spraying 200 applications at "companies that sponsor" is the slowest way to get sponsored, because most of them never have. The candidate who builds a 30-company list of confirmed active sponsors in their exact occupation is playing with computable odds. Everyone else is playing a lottery they built by hand.

Here's where H-1B sponsorship still clusters, and what decides your odds in each lane. Every row is sourced in the sections below.

Lane	Who files	What decides your odds
Volume filers	Amazon, Cognizant, EY, Microsoft, Google (the top 10 filed 88,939 LCAs in FY2025)	Your exact job title and wage level
Computer roles	64% of FY2024 H-1B approvals	Wage level, under the new weighted lottery
Level III and IV wages	Any cap-subject employer	About 46-61% selection odds, vs about 15% at Level I
Lower-cost cities	Employers outside NYC and SF	A lower prevailing wage makes Level III easier to clear
Cap-exempt employers	Universities, affiliated nonprofits, research bodies	No lottery at all; they file year-round

The instinct most people have is the wrong one. You hear "the visa market is hard" and you respond with volume: more applications, more companies, more hours on LinkedIn Easy Apply. But the market didn't get hard evenly. It narrowed on specific axes, and volume aimed at the wrong axes is just faster failure, the same reason applying to hundreds of jobs is a symptom rather than a strategy. The work that pays off is narrowing your aim to match the market that still exists.

Where did all the visa sponsorship jobs go?

They mostly evaporated from the places you're looking. On Handshake, the early-career job board, the share of full-time listings offering visa sponsorship fell from 10.9% in 2023 to 2.6% in 2026, a 76% collapse in three years. Across all of Indeed, only 0.14% of US job postings offered visa or green card sponsorship as of May 2025. And the surviving share is lopsided: medical occupations make up roughly three-fourths of all sponsoring postings, with physicians at 3.2% and therapy at 1%.

Then policy made it worse for mid-market employers. A Presidential Proclamation in September 2025 imposed a $100,000 fee on new H-1B petitions for beneficiaries living outside the US. For a 5,000-person software company weighing one international hire, that fee ends the conversation. The companies that can absorb it are the ones already filing hundreds of petitions a year. The market you imagined, where any decent tech firm might sponsor a strong candidate, is the one that's gone.

Which companies offer H-1B sponsorship?

A short list, and they're public. In FY2025, the top 10 employers by LCA filings accounted for 88,939 filings: Amazon led with 15,524, then Cognizant (11,140), Ernst & Young (10,371), Microsoft (9,362), Google (8,933), TCS (7,879), Apple (7,037), Meta (6,627), Infosys (5,786), and Deloitte (4,278). This is not a secret. It's disclosed, downloadable, and almost nobody applying ever reads it.

Two things to hold at once. First, this concentration tells you where to aim. Second, even the giants are pulling back: Amazon's H-1B approvals dropped from over 11,000 in FY2023 to around 7,000 in FY2024, and Infosys fell from 7,300+ to about 5,900 over the same window. So the list is necessary but not sufficient. You want employers who are on it and trending up and hiring your role right now.

Weak: "I'm targeting FAANG and the big consulting firms, broadly." You're aiming at a category, half of which is shrinking, without checking a single filing number.

Strong: "I pulled the FY2025 LCA data, filtered to companies that filed 10+ petitions for my job title, cross-checked their current openings, and built a 40-name list I'm working through with direct outreach to hiring managers."

The strong version takes an afternoon. It replaces 150 blind applications with 40 aimed ones.

Does the UK's growing sponsor list mean it's getting easier there?

Not the way the headline number suggests. The UK register of licensed sponsors has grown to well over 120,000 organizations, roughly four times the early-2021 count. Read fast, that looks like a market opening up. Read carefully, it's a license count, not a hiring count. A license means a company can sponsor, not that it currently is. Only a fraction of those organizations are advertising sponsored roles at any given time, so the register vastly overstates how many doors are actually open this month.

So the register size is a trap if you treat it as a target list. The real list is the small, active subset. Here's the difference in practice.

Weak: Search "visa sponsorship jobs London," apply to anything with a license.

Strong: Download the Home Office register, filter to companies granted a license in the last 12 months, cross-reference against open postings, and find the fintech (a Wise or a Monzo) actually hiring your role right now.

The honest read: the UK is genuinely larger than the headline lottery panic implies. But "larger register" and "more places hiring you" are different claims. Target the active subset and the UK is workable. Treat the whole register as your funnel and you'll drown.

Why does the role you target matter less than the salary band?

Because the lottery math just changed under everyone's feet. Yes, the roles are concentrated: computer-related occupations were 64% of all H-1B approvals in FY2024, and Software Engineer was the single biggest title at 30,014 LCAs in FY2025. Get a software role and you're in the densest sponsorship lane there is. But the role is only half the story now.

The other half is your wage level. Historically, 83% of cap petitions sat at Level I or Level II (28% and 55% respectively). That was fine when the lottery treated everyone equally. It no longer does. The new weighted lottery effective FY2027 gives Level IV candidates about 61% selection odds versus 15% for Level I, through a system that hands Level IV four entries against Level I's one. Under the old uniform draw, everyone sat around 29.6%.

Wage level	Old odds (uniform)	New odds (weighted)
Level I	~29.6%	~15%
Level II	~29.6%	~31%
Level III	~29.6%	~46%
Level IV	~29.6%	~61%

A candidate at a firm offering Level I wages is now roughly 4x less likely to be selected than one at Level IV. Which surfaces the staffing-firm trap: the big IT consultancies are top-volume filers, so they look like prime targets, but they concentrate petitions at the low wage bands. High volume, worst odds. The title gets you into the pool. The wage band decides whether you get drawn.

How do you target wage band instead of just begging for any sponsor?

By treating the band as something you can engineer, partly through geography. The same Level III or IV classification is easier to clear at a mid-tier-city employer than at a NYC or SF mega-firm, because the prevailing wage you have to beat is lower where the cost of living is lower. This is the visa-lottery version of why the same job pays wildly differently depending on where you sit, except here the location math moves your odds of getting drawn instead of your paycheck.

Weak: A data scientist targets the biggest-name employers, lands a Level I/II band, and sits at 15-31% lottery odds without realizing it.

Strong: The same data scientist finds a comparable role at a tech company in Austin or Raleigh where the local prevailing wage is lower, clears the Level III threshold on the same base salary, and moves their odds from 31% to 46%.

Same person, same skills, same salary number. Different city, different wage-level classification, materially different odds. That's not a trick. It's reading the rule and positioning inside it. Most candidates never learn the rule exists, so they can't aim for it.

What about jobs that skip the lottery entirely?

They exist, and for the right candidate they're the best option on the board. Cap-exempt employers, universities, affiliated nonprofits, nonprofit research organizations, and government research bodies, can file H-1B petitions year-round with no annual cap and no lottery. No 85,000 ceiling. No March registration scramble. No weighted draw.

If you're in an academic, clinical, or R&D track, this changes everything about your strategy.

A bioinformatics PhD targeting pharma broadly misses that a university-affiliated research hospital can file their petition year-round, lottery-free. One aimed application there is structurally worth more than fifty to cap-subject employers, where you're still rolling dice.

This is the part of the market that's quietly more open, not less. It rewards a completely different list than the FY2025 LCA leaderboard. If your background fits, build that list instead.

What are the honest trade-offs?

Three things this post owes you, said plainly.

The tightening isn't uniformly bad. The weighted lottery genuinely helps senior, well-paid candidates: 46-61% odds at Levels III and IV beat the old 29.6% for everyone. It got harder at the bottom and more winnable at the top. If you're early-career at a Level I wage, that's the honest bad news. If you're senior, it's tailwind.

The bottleneck is the lottery and the employer's willingness to file, not USCIS being stingy after the fact. Filed petitions still get approved at very high rates. So precision targeting is about getting to the filing stage with an employer who'll actually file for your band. It doesn't change what happens once you're filed.

And the real cost of this approach: a narrow list means a smaller pipeline. Forty aimed applications feels thinner than 200 sprayed ones, and that's uncomfortable when the clock on your work authorization is running. The trade is fewer shots, each converting far better, against more shots that mostly hit employers who were never going to sponsor anyone. You're paying in volume to buy odds. For most people on a visa clock, that's the right trade. Name it so you choose it on purpose.

What to do now

This week, not someday:

1. Pull the filing data for your exact role. Use the public LCA disclosures and the USCIS Employer Data Hub. Filter to employers who filed 10+ petitions for your job title recently. That's your raw list.
2. Score your wage band. Find where your salary lands you (Level I through IV) and where it could land you in a lower-prevailing-wage city. Aim for III or IV if it's reachable.
3. Separate the cap-exempt track. If you're academic, clinical, or research-adjacent, build a parallel list of universities and affiliated research orgs. Different game, better odds.
4. For the UK, download the register and filter to fresh, active sponsors. License count is noise. Recent grants plus open postings is signal.
5. Reach out directly. Hiring managers, not just the ATS. A 40-name list you actually contact beats a 200-name list you spray.

The candidate who does this isn't hoping. They can see where the sponsorship still clusters and they're aiming at it, the same discipline that separates the people working a real list from the ones treating thousands of hiring cities as one undifferentiated funnel. Consistency on the right list beats intensity on the wrong one, every time.

Want me to build that list for you? Tell me your role and target country on WhatsApp, and I'll pull which employers actually sponsor it, at what wage band, with what recent filing history. That's the whole research step, done before you send a single application.