For most of the last decade, "free coding bootcamp" described two entirely different financial arrangements that happened to share a word. One was training funded by donors, grants, and employers, where the student genuinely paid nothing at any point. The other was an income-share agreement: nothing upfront, then a percentage of salary once the graduate crossed an earnings threshold. The second was not free. It was deferred, and the pitch depended on the distinction being easy to miss.
In 2026 those two routes have visibly separated. The nonprofit route is intact and, by its own account, growing. The income-share route has largely vanished from the places you would expect to find it — including the price pages of the providers that made it famous.
What Actually Costs Nothing
Against an average bootcamp tuition of $13,584 reported by Course Report for 2026, the genuinely no-cost programmes are worth understanding on their own terms rather than as a discount.
Per Scholas provides no-cost training across seven tracks: AWS re/Start, cybersecurity analyst, cybersecurity with AI tools, data centre technician with AI tools, IT support, Salesforce administrator, and AI-native software development. The organisation reports having trained more than 30,000 people over roughly thirty years, and names employers including Amazon, Wells Fargo, Bank of America, and Infosys among those where alumni work.
LaunchCode states plainly that it never charges tuition, fees, or a deposit. Its funding comes from three sources it describes openly: donations, public and private grants, and revenue earned from employers after paying wages and benefits to apprentices. The apprenticeship is paid and includes healthcare for apprentices and their dependents, and LaunchCode reports that apprentices convert to permanent hires more than 75 percent of the time. Admission runs through an online application with coding puzzles, then a technical interview in which the applicant demonstrates something they built — assessment on aptitude rather than credentials, which is the point of the model.
The structural reason these can charge nothing is that the student is not the customer. Employers and funders are. That alignment is genuinely favourable, and it also explains the constraints: cohorts are limited by funding rather than demand, admission is selective, and geographic availability follows the funding rather than the applicant.
The Outcome Numbers Are Self-Reported
Neither organisation submits its outcomes to third-party audit. The Council on Integrity in Results Reporting, the sector's audited-reporting standard, currently lists Codesmith, Code Platoon, and Hacktiv8 among its reporting schools. Neither Per Scholas nor LaunchCode is among them.
That is not an accusation. Both are nonprofits with long operating histories and named employer relationships, and neither has an obvious commercial motive to inflate a figure it uses to attract funders who audit it themselves. But LaunchCode's 75-percent conversion rate is a self-reported figure using its own definitions, and Per Scholas does not publish a placement rate on its main site at all. The correct posture is the one we set out in our guide to what placement rates conceal: ask what counts as placed, over what window, and how non-responding graduates are treated. Those questions apply to free programmes exactly as they apply to paid ones.
The Income-Share Agreement Has Quietly Gone
App Academy was the flagship of the income-share model — pay nothing, then hand over a share of your salary once employed. Its current tuition page tells a different story. The options now presented are a $10,000 base tuition carrying a $6,500 discount for paying upfront, leaving $3,500 due at enrolment; an instalment plan at $189 a month over thirty-six months with no deposit; and deferred tuition through third-party financing partners, with payments postponed until three to nine months after graduation. Income-share loans appear as a mention, noting that a hard credit check is required, with no terms published on the page.
Read that list against the original pitch and the change is stark. A model built on the school taking a stake in the graduate's future earnings has been replaced by a discounted cash price, a conventional thirty-six-month instalment loan, and referrals to outside lenders. The financial risk has moved back onto the student and onto third-party finance, which is where it sits in ordinary tuition.
Why the model retreated — and when
The decisive regulatory moment is older than most 2026 coverage implies, and dating it correctly matters. On 7 September 2021 the Consumer Financial Protection Bureau took action against Better Future Forward, an income-share provider, in the first public federal enforcement action against an ISA company. The Bureau found that the provider had falsely represented that its ISAs were not loans and did not create debt, had failed to give the disclosures required under the Truth in Lending Act, and had imposed unlawful prepayment penalties. Its formulation was direct: regardless of the name on the label, these products are credit and must comply with federal consumer protections. No financial penalty was imposed, the Bureau citing the company's good-faith cooperation.
That single action removed the premise the marketing rested on. If an income-share agreement is credit under federal law, it must carry loan disclosures, cannot be described as debt-free, and cannot carry prepayment penalties — and once it is disclosed as a loan, its main selling point over a loan disappears. State-level supervision followed, including a California Department of Financial Protection and Innovation arrangement with the ISA servicer Meratas.
So 2026 is not the year the ISA was regulated. It is the year its absence became visible on the price pages, several years after the legal question was settled. Anyone reading a current article that frames ISA decline as breaking regulatory news has the sequence backwards.
What Replaced It
The gap the ISA occupied — students who cannot pay upfront and will not take a private loan — is now being filled from an unexpected direction. Workforce Pell became available on 1 July 2026 for qualifying short-term programmes, extending federal grant aid, which is not repaid, to training that had never been eligible for it.
| Route | What the student pays | Who carries the risk |
|---|---|---|
| Nonprofit, grant-funded (Per Scholas, LaunchCode) | Nothing at any point | Funders and employers |
| Workforce Pell at a qualifying programme | Nothing up to the grant amount; grant is not repaid | Federal government |
| Income-share agreement | Nothing upfront, a share of later salary | Shared, and legally a loan |
| Deferred tuition via a finance partner | Full tuition, beginning after graduation | Student |
| Instalment plan | Full tuition, spread over months | Student |
The top two rows are the only ones where "free" is accurate without qualification. The bottom three are payment timing, and describing them as free is the specific confusion the CFPB action was about. For students weighing a Pell-eligible programme, our eligibility guide and state-by-state tracker cover which programmes qualify.
The practical order of operations for someone who cannot pay upfront in 2026: check whether a nonprofit programme serves your area and field, since it costs nothing and admission is the only barrier; then check whether a Workforce Pell-eligible programme at a local institution fits, since a grant is better than any loan on any terms; and only then evaluate financed options as what they are, which is borrowing. An arrangement presented as free that involves a credit check is a loan, and the disclosure it must now carry will say so.
Are tuition-free bootcamps lower quality than paid ones?
Not by construction. Their revenue depends on employers hiring graduates, which is a direct incentive toward employable skills. The real differences are selectivity, cohort availability limited by funding, and geographic coverage. Curriculum currency varies by provider in both the free and paid segments and is worth checking directly.
Does an income-share agreement still make sense if I am offered one?
Evaluate it as a loan, because that is its legal status. Ask for the disclosures it is required to provide, then compare total expected repayment against a conventional loan and against the discounted upfront price, which providers now often price well below the nominal tuition. Check the income threshold, the payment percentage, the payment cap, and the maximum term — a capped ISA and an uncapped one are very different products.
Why do free programmes not report to CIRR?
The reasons are not published. Audited reporting carries cost and administrative burden, and nonprofits already report outcomes to funders under those funders' own requirements, which may make a second framework redundant from their perspective. It does mean their figures rest on their own definitions rather than an external standard.
Can I use Workforce Pell at any bootcamp?
No. The programme must run at least 150 and fewer than 600 clock hours over at least eight and fewer than fifteen weeks, be offered by an institution participating in federal student aid, and satisfy outcome and quality criteria. Many private bootcamps do not meet the institutional requirement at all, which is why community colleges dominate the qualifying list.
Is "free upfront" ever the best available option?
It can be, when the alternatives are no training at all and the terms are capped and disclosed. What has changed is that it can no longer be presented as an alternative to debt. It is a form of debt, and comparing it against other debt on total cost is the only sound way to assess it.
← Back to Bootcamps | Bootcamp Cost After Workforce Pell | Is the Job Market Cooling? →