EB-5 capital funds US development by filling a gap in a project's capital stack, normally as mezzanine debt or preferred equity sitting between the senior bank loan and the developer's own money. What it has built, by volume, is real estate: hotels, rental apartments, student housing, senior living, mixed use blocks and office conversions in large metros. Since the EB-5 Reform and Integrity Act of 2022 rewired the visa queue, a growing share of new money has moved into rural deals such as food processing, manufacturing, agriculture and energy. Every dollar arrives with the same string attached, ten full time jobs for each investor, so project selection is driven by job arithmetic as much as by construction economics.
Why real estate dominated EB-5 for three decades
Three features made construction projects the natural home for EB-5 money. First, the regional center model lets a project count indirect and induced jobs produced by an economic model, and construction spending plus stabilised operating revenue are exactly the inputs those models reward. A large hard cost budget can generate a large modelled job number without a single worker ever appearing on the new commercial enterprise's own payroll.
Second, real estate has a defined completion date and a repayment event, usually a refinancing or a sale, which gives the loan a plausible exit. Third, developers already understood mezzanine finance and were happy to take capital priced below what a private credit fund would charge. The statutory framework in the USCIS EB-5 Immigrant Investor Program pages never mandated real estate, but the incentives pointed there, and critics have fairly described the result as cheap subordinate financing for buildings that would have been built anyway.
The rural shift after the 2022 reforms
The Reform and Integrity Act reserved a slice of the annual visa supply for particular project types: 20 percent for rural projects, 10 percent for projects in high unemployment areas, and 2 percent for infrastructure. Those reserved visas sit in their own queues. For an investor born in a heavily backlogged country, choosing a rural project has meant a materially shorter queue than the unreserved line, and capital followed that difference almost immediately. The reserved queues are not immune to backlog either, so check the current bulletin rather than a figure quoted to you last year. The statute itself is published as Public Law 117-103 on GovInfo.
Rural has a hard definition. It means a place outside any metropolitan statistical area and outside any city or town with a population of 20,000 or more, which is why sponsors lean on the Census Bureau metropolitan and micropolitan area definitions when they document a claim. High unemployment areas are argued from labour market data, most often the series published through the Bureau of Labor Statistics local area unemployment statistics. Ask for the underlying tables, not the summary sentence in the marketing deck. A targeted employment area claim that falls apart later can take your set-aside priority with it.
What actually counts as an infrastructure project
The 2 percent infrastructure set-aside is narrow. It contemplates a project where a governmental entity is the job creating entity, contracting with the enterprise to finance public infrastructure. Very few offerings genuinely fit. If a sponsor markets a private toll operator or a private utility as infrastructure, read the structure documents closely before you accept the label.
How a project turns $800,000 into ten jobs
The mechanics differ sharply by structure. In a direct investment, only jobs on the payroll of the enterprise count, and you will eventually prove them with W-2 forms and payroll registers. In a regional center project, the sponsor commissions an economic report that converts construction spending, tenant revenue and operating expenditure into direct, indirect and induced jobs using an input output model. The definitions that govern all of this, including what a full time position means, live in 8 CFR 204.6 on eCFR.
The number you should ask for is the cushion. Divide the total modelled jobs by the number of investor slots. A project offering barely ten jobs per investor has no margin at all, and construction budgets come in under plan more often than sponsors like to admit. Our guide to EB-5 economic impact reports and how job numbers are calculated walks through where those models are strong and where they are optimistic, and the broader picture is in the programme's overall economic impact.
What the public record can and cannot tell you
There is no federal database that lists every EB-5 funded project with verified job counts and investor outcomes. USCIS publishes aggregate petition and adjudication figures through its immigration and citizenship data reports, and the agency identifies approved regional centers, but the performance of an individual deal is private information held by the sponsor.
That gap is why marketing decks are full of famous buildings. A skyline photograph proves that a sponsor once raised EB-5 money for a project in that city. It does not prove investors were repaid, does not prove the modelled jobs materialised, and does not prove I-829 petitions were approved. Treat a landmark reference as a starting point for questions, never as a result.
Reading an EB-5 project the way an underwriter would
- Where does the EB-5 money sit in the capital stack, and who is senior to it?
- How much developer equity is genuinely at risk, in cash rather than in contributed land value?
- What is the ratio of modelled jobs to investor slots, and which cost lines drive it?
- What documents support the rural or high unemployment claim, and who signed them?
- Is there a guaranteed maximum price construction contract, and who bears an overrun?
- What is the repayment source and date, and what happens if visa queues outlast the loan term?
- Has this sponsor taken investors all the way through to approved I-829 petitions before?
- Who controls the enterprise, and what can they do without asking you?
Most of these questions are answered, or conspicuously not answered, in the offering documents. If you are considering a deal with many co-investors, read how the structure allocates jobs and repayment in pooled EB-5 investments and joint ventures. Sponsor track record matters more than project sector, and the survey of major regional centers and their risk signals is a reasonable place to calibrate.
Rural is faster, not automatically safer
The set-aside changed the visa maths, not the credit maths. Rural projects can carry real advantages, including lower land costs, motivated local authorities and genuine unmet demand for processing or logistics capacity. They can also carry thinner labour markets, longer supply chains, fewer buyers if the asset must be sold, and sponsors with less institutional depth. A rural project that stalls leaves you with a delayed loan and a job shortfall in a county with little slack to absorb it. The rural renewable energy case study shows how differently these deals behave from a downtown hotel.
How the legal deadlines shape which projects exist
The regional center programme is authorised through 30 September 2027, with grandfathering protection for petitions filed by 30 September 2026, and the first inflation adjustment to the investment amounts is due on 1 January 2027. Those dates shape sponsor behaviour. Deals get structured, marketed and closed around them, and a rushed offering assembled to catch a deadline deserves more scrutiny than a project that has been in planning for two years. The wider policy picture is covered in how new EB-5 laws changed investor demand.
The honest summary
EB-5 has financed a real and substantial amount of American construction and, more recently, a meaningful amount of rural industrial capacity. It has also financed projects that failed, and the programme's own integrity provisions exist because some sponsors misused investor money. Your money is not buying a share of the programme's aggregate record. It is buying a position in one deal, with one sponsor, in one town. Judge that deal on its own documents.