The Visa Sponsorship Target List, From Disclosure Data to a Ranked Shortlist

You need work-visa sponsorship, which means you cannot apply the way everyone else does. A posting is only worth an application if the employer can file for your specific occupation, in your metro, at a wage level that gives you real odds. This guide runs the full job end to end: from the public disclosure files, through the per-role sponsorship math, to a ranked shortlist with a reason code on every row. It is written for a sponsorship-dependent job seeker who is tired of applying blind and wants to reject the dead ends before spending an application on them.

The pages that rank for this job stop at a binary: has this company ever sponsored. That question is now out of date. The rules changed, and the new question is whether a specific posting is sponsorable for you.

Why "has the employer ever sponsored" is the wrong question now

The old filter asked whether a company ever filed an H-1B petition. That filter no longer decides your odds, because the registration itself now carries role detail. From FY2027, each H-1B registration must include the position's SOC occupational code, area of employment, and corresponding OFLC wage level from I to IV. The wage level then weights your entry into the lottery.

So the gate moved. A company with a decade of filings that would only file your role at Level I is a weaker target than a company with a short history that files it at Level IV. The decisive attribute is no longer the employer's track record in the abstract. It is the per-role math.

There is a second reason to build a targeted list rather than spray applications. The applicant pool shrank hard after the system moved to counting unique beneficiaries. Eligible registrations fell from 758,994 in FY2024 to 343,981 in FY2026, a drop of roughly 54.7%, while the selection rate climbed to 35.3%. A smaller, cleaner pool rewards a candidate who aims at the right employers and the right wage levels.

The employer's filing history tells you it can sponsor. The wage level tells you whether it will help you win.

The three public datasets and what each field proves

Three public sources carry everything the list needs, and each proves a different thing. Use all three; none is sufficient alone.

| Source | What it is keyed on | What it proves | | DOL OFLC LCA disclosure files | SOC code, wage level, worksite, employer name | The role-level filing: occupation, location, and wage level | | USCIS H-1B Employer Data Hub | Employer name, last four of tax ID, fiscal year | Approval and denial counts at the employer level | | PERM and Prevailing Wage files | Same OFLC page, same fiscal year | Green-card intent and wage benchmarks |

The OFLC LCA file is the spine. Each labor condition application contains the job title, SOC code, prevailing wage, actual wage offered, wage level (1 through 4), worksite location, and employer information, and practitioners report each file carries 75 or more columns per record. That wage level field is the one the old guides ignore and the new rules reward.

The USCIS H-1B Employer Data Hub lets the public search petitioners by fiscal year, NAICS code, employer name, city, state, or ZIP code, and it identifies employers by the last four digits of their tax identification number. It reports initial and continuing approvals and denials, but first-decision data only, so pending cases are excluded. Use it to confirm an employer actually gets petitions approved, not just filed.

A warning before you join them. USCIS and DOL files lack a shared unique ID, so you cannot cleanly link an LCA record to a Hub record by key. Expect approximate joins and validate on the tax-ID last four from the Hub and the full legal employer name from the LCA. A naive name-join merges two different employers and inflates both your approval rates and your filing counts.

Update cadence, and how to avoid a stale list

OFLC publishes disclosure files quarterly and annually by federal fiscal year, which runs October 1 to September 30, in Excel. A recent release covered final determinations for PERM, LCA, H-2A, H-2B, CW-1, and prevailing wage requests processed between October 1 and June 30 of the following calendar year. The USCIS Hub is updated quarterly, with the first fiscal quarter (October to December) appearing around April. Third-party tools lag one to three months behind because the DOL only publishes quarterly. Before you rank, confirm the OFLC quarter and the USCIS Hub fiscal year you are working from.

The per-role sponsorship math

Wage level is the number that now decides whether a posting is worth your application. DOL slots every H-1B job into one of four levels based on the experience, education, supervision, and judgment the role actually requires. The four levels correspond to fixed percentiles in the OES wage data for the occupation and location, and the only thing that moves the dollar amount is your SOC code and your metro area.

From FY2027, each registration receives between one and four entries into the selection pool according to the wage level the offered salary meets or exceeds. Level IV positions receive four times the weight of Level I positions. That weighting is why an employer who files your occupation only at Level I is a quiet rejection in disguise, even with a long history.

| Wage level | OES percentile | FY2027 lottery entries | DHS est. selection odds | | L1 | 17th | 1x | 15.29% | | L2 | 34th | 2x | 30.58% | | L3 | 50th | 3x | 45.87% | | L4 | 67th | 4x | 61.16% |

Read Table A as a priority order. Level IV odds are roughly four times Level I odds, which matches the entry weighting exactly. When you tag each employer in your list, record the wage level they typically file for your SOC and metro. Prefer Level III and IV. Treat Level-I-only employers as low priority unless they are cap-exempt, which changes the game entirely.

The two fee and lottery escapes: cap-exempt and the $100,000 trigger

Two facts sit above the wage-level math and can reshape your whole list. The first is cap-exempt status. The second is the $100,000 fee.

H-1B workers petitioned for or employed at an institution of higher education, its affiliated nonprofit entities, a nonprofit research organization, or a government research organization are not subject to the annual numerical cap. These employers file off the lottery and year-round. That means no 35.3% gamble and no waiting for the March registration window.

Cap-exempt is also the only path that escapes both the lottery and the fee. The $100,000 payment applies to certain H-1B petitions filed on or after September 21, 2025, but H-1B extensions, amendments, transfers, and cap-exempt filings are not affected. So a cap-exempt employer removes two of the hardest obstacles at once.

Verification matters, because the word "nonprofit" alone proves nothing. An affiliated nonprofit must be tax-exempt under Internal Revenue Code section 501(c)(3), (c)(4), or (c)(6) and be approved by the IRS as a tax-exempt organization for research or educational purposes. A nonprofit training company that occasionally places interns at a university generally would not qualify. Verify through the employer's IRS tax-exempt status, Higher Education Act accreditation for universities, and a written affiliation agreement for affiliated nonprofits. A placed-at worker can also qualify if the beneficiary will spend the majority of work time performing duties at a qualifying institution and those duties directly and predominantly further its essential mission.

There is a reachable pool here. In Refolk's index, 2,648 current Software Engineers in the United States have an employer in the Higher Education sector, a cap-exempt-leaning group. The top employers in that pool are the University of California San Francisco, the University of Cincinnati, Florida State University, Arizona State University, and the University of Texas at Austin.

| Metro | Count in sample | | San Francisco Bay Area / San Francisco | 8 | | Austin, Texas | 2 | | New York City metro | 2 |

Table C is a 25-record sample of that 2,648-person pool, so read it as directional concentration rather than population counts. It tells you where to point a cap-exempt search first, not how many jobs exist.

Refolk can turn your own history into a resume and then score how well you fit each cap-exempt posting, which removes the manual work of checking whether your experience lines up with a university research role before you apply.

Reading posting language: confirm, rule out, or ambiguous

Once you have an employer shortlist, every live posting sorts into one of three buckets based on its own words. This is the fastest filter you have, and it is free.

Here is what each bucket looks like, and what each signal looks like when it lies.

• Confirms (count as yes): direct wording such as "visa sponsorship available," "sponsorship considered," "work permit support provided," or "relocation package offered." When it lies: a boilerplate benefits list that mentions relocation but the recruiter later says no work-visa support. Verify on the first call.
• Rules out (hard reject): an employer may state in its postings that it will not sponsor applicants for work visas, and phrasing like "applicants must already hold the right to work" is a clear no. This rarely lies; take it at face value and do not apply.
• Ambiguous (do not count as yes): phrases like "international candidates welcome" or "global team" do not mean sponsorship is available. When it lies: it reads warm and inclusive and means nothing about visas. Keep an ambiguous posting only if the employer has a recent filing matching your SOC and metro, and confirm with the recruiter.

Build the shortlist, step by step

This is the full procedure, in order. Each step has a target time and a definition of done. Role-seekers should start from OFLC rather than the USCIS Hub, because OFLC carries the SOC code and wage level you need; the Hub carries approval rates you add afterward.

For the ranking in the final step, use a simple scoring template so every row is comparable and every rejection is explainable. Copy this and fill one row per employer.

Employer (full legal name):
Tax ID last-four (from Hub):
Filing recency (last quarter with a matching LCA):
SOC + metro match (exact / partial / none):
Typical wage level for your role (L1 / L2 / L3 / L4):
Cap-exempt (yes / no / unverified):
Hub approvals vs denials (current + prior FY):
Posting language (confirm / ambiguous / rule-out):
Rank score (recency + match + level + cap-exempt):
Reason code (APPLY / WATCH / REJECT-stale / REJECT-mismatch / REJECT-L1only / REJECT-language):

Once you have the OFLC-matched employers and want a live posting for each, Refolk can tailor your resume to a specific posting and score how well you actually fit it, so the ambiguous rows resolve faster.

How this goes wrong: failure modes and false positives

Most bad shortlists fail in one of eight ways. Each has a false positive that looks like a match and a check that catches it. Treat this as the most valuable part of the guide; the rejections you get right are worth more than the applications you send.

| Failure mode | What the false positive looks like | The check | | "Ever filed" read as "will file for me" | 2019 LCAs, no recent activity | Require a filing in the last one to two quarters matching your SOC and metro | | SOC or metro mismatch | Sponsors Data Scientists in Seattle, you are an Engineer in Austin | Filter LCA rows to your exact SOC and worksite metro before counting | | Wage-level blindness | Employer files, but only at Level I (~15% odds) | Read the wage level field; prefer Level III or IV | | "Nonprofit" assumed cap-exempt | A training firm that places interns at a university | Confirm IRS 501(c) status and a formal affiliation or research mission | | Ambiguous language read as yes | "Global team" treated as sponsorship | Require explicit sponsorship wording; otherwise verify with the recruiter | | Stale data | List built on a file one quarter behind | Confirm the OFLC quarter and USCIS Hub FY before ranking |

Two more deserve their own paragraph. Join errors across DOL and USCIS merge two different employers with similar names, inflating counts; validate on the Hub's tax-ID last four and the LCA's full legal name. Ignoring the $100,000 fee trigger makes a technically sponsorable cap-subject role cost-prohibitive for consular-notification cases; assess whether your case is a new consular filing versus an extension, amendment, transfer, or cap-exempt filing, all of which are exempt.

Verify before you call the list done

Run this checklist against your shortlist before you spend a single application. If any item fails, fix it rather than apply around it.

Keeping the list current

A visa target list is not a one-time artifact; it decays with every new disclosure quarter and every closed posting. Rebuild the data layer when a new OFLC quarterly file or USCIS Hub update lands, which is roughly quarterly, and remember that third-party tools run one to three months behind the government release. Re-run the wage-level tag after each refresh, because an employer that filed your role at Level II last quarter may file it at Level III next, which changes its rank.

Watch the rules as well as the data. The FY2027 weighted-selection final rule is effective February 27, 2026, for the lottery in March 2026, and the registration now requires SOC code, area of employment, and wage level. Premium processing was listed at $2,805, increasing to $2,965 effective March 1, 2026. Where a figure is time-sensitive, do not memorize the value; re-check it at the OFLC performance page and the USCIS Hub glossary before each cap season. The method in this guide holds even when the numbers move, because it ranks on the mechanism, not on a snapshot.