U.S. visas for startup founders

The petitioner problem, and how founders solve it

The O-1A cannot be self-petitioned. Neither can the H-1B or the L-1A. All of them require a U.S. employer, or in O-1's case a U.S. agent, to file on your behalf. For a founder, that means your own company petitions for you — and USCIS then asks whether a real employer-employee relationship exists.

The test is control. Can the petitioner hire, fire, pay, supervise, or otherwise control the beneficiary's work? If you are the sole owner, sole officer, and sole decision-maker, the honest answer is no, and a petition filed on those facts is a recurring denial pattern rather than an edge case.

What actually works is a corporate structure where someone other than the beneficiary holds that authority. In practice that means a board with independent directors, or investors with the contractual right to remove officers, and documentation showing the arrangement is real rather than papered for immigration purposes: board consents, an employment agreement executed by someone other than you, and evidence of the board actually exercising oversight.

This is easier for a venture-backed company than a bootstrapped one, and that is not incidental. A seed round with investor board seats gives you the structure almost by default. If you have not raised, this is worth working through with counsel early — retrofitting governance after a denial is considerably harder than setting it up before filing.

The O-1 agent route is the other option and it is underused. Where a beneficiary will work for multiple entities, a U.S. agent can petition. It has its own requirements — the agent relationship must be genuine, and the itinerary and contractual terms have to be documented — but it can fit a founder whose work genuinely spans several companies.

• Sole owner + sole officer + no board = the hardest fact pattern to petition on.
• Independent directors or investors with removal rights create the control USCIS looks for.
• Document the governance as real: board consents, an employment agreement signed by someone else, evidence of oversight.
• The O-1 agent route can fit founders working across multiple entities.
• EB-1A and EB-2 NIW sidestep this entirely — both are self-petitioned.

What a founder's record actually proves

Founders often have strong evidence and file it under the wrong criteria, or fail to make the argument that connects it.

Funding is evidence, but of what? A priced round from recognised investors supports original contributions of major significance and can support a critical role, because sophisticated parties conducted diligence and concluded the work had value. It is not itself a criterion, and a petition that presents a funding announcement without an argument leaves the adjudicator to guess. Say what the round proves and about what.

Product adoption is the strongest founder evidence available. Named enterprise customers, revenue, user numbers, integrations by other companies, and third-party products built on your platform all demonstrate independent adoption beyond your own organisation — precisely what original contributions of major significance requires.

Press coverage satisfies published material about you only if it is about you and your work, with title, date, and author. A funding-announcement piece that mentions you in a list of founders is weak. A profile of your work in a major trade publication is strong. Coverage your own company placed is routinely discounted.

Critical or essential capacity is where founders sometimes overreach. Being the founder demonstrates the role; it does not demonstrate that the organisation has a distinguished reputation. That needs separate evidence — funding, market position, independent rankings, substantive coverage. A two-person pre-seed company will struggle here regardless of how central the founder is to it.

Judging the work of others is available to more founders than use it: accelerator selection committees, pitch competition judging, grant review panels, and advisory roles evaluating other companies all count.

When EB-2 NIW is the better founder route

The national interest waiver is frequently the right answer for entrepreneurs and it is under-considered, largely because founders assume EB-1A is the prestige option and stop there.

NIW is self-petitioned — no employer, no job offer, no labor certification — so the petitioner problem disappears. And the Matter of Dhanasar framework fits entrepreneurship unusually well. Prong three asks whether it benefits the United States to waive the job offer requirement, and self-employment is a direct argument: there is no employer to run a labor certification through, which is the point.

The prong founders get wrong is the first one. National importance attaches to the endeavor, not to the company. A petition describing what the startup does, however impressive, is describing a business. An endeavor names a problem, a beneficiary population, a scale, and an expected effect. "Building developer tools" is a market segment. "Reducing the cost of formal verification so that safety-critical software teams outside large defence contractors can afford it, and here is what that failure currently costs" is an endeavor an adjudicator can evaluate.

Prong two — being well positioned — is where the founder record does its work: prior successes, technical credentials, the plan itself, investor commitments, customer traction, and letters from people positioned to judge whether you can execute.

The trade-off is the visa bulletin. EB-2 is more oversubscribed than EB-1, and for founders born in India or China the wait for a visa number can be years longer even though the petition standard is lower. Some founders file both categories; a denial in one does not prejudice the other.

Sponsoring your own team

Most founders reading this eventually have the second problem: a critical engineer who needs status, and no appetite for losing twelve weeks to a boutique firm's billables.

The H-1B is cap-subject for most employers, which means a spring registration lottery and an October 1 start at the earliest. Missing that window costs a year. Plan hiring around it, and note the fee stack — the ACWIA fee, the Fraud Prevention fee, and, for larger employers, the Public Law 114-113 surcharge. Department of Labor rules bar passing the ACWIA and Fraud Prevention fees to the employee.

For someone already working for a related entity abroad, the L-1A or L-1B has no cap and no lottery. The new-office variant is how many companies establish a U.S. presence, and it grants one year before USCIS reviews whether the operation actually materialised.

For Canadian and Mexican professionals in a listed occupation, TN is the cheapest and fastest route in U.S. immigration — a Canadian can be admitted the same day at a port of entry. The catch is that the role must map onto a USMCA Appendix 2 profession, and there is no software engineer entry.

And for a genuinely exceptional hire, the O-1A has no cap, so it can be filed whenever the evidence is ready rather than when the calendar allows.

The categories in full

Each guide covers who qualifies, what evidence satisfies the standard, how the process runs, what the government charges, and the RFE patterns to avoid — with the regulations cited.

Common questions