When Prosper booked its first loan in 2006, the idea was almost utopian: ordinary US savers would fund ordinary US borrowers and the bank in the middle would disappear. The reality of peer-to-peer lending in the US is less romantic. According to the Federal Reserve’s Survey of Consumer Finances, most US consumer credit still flows through traditional banks, credit unions and credit-card issuers. The original retail-investor model has shrunk; the marketplace-lending model it inspired has not.
What the term means now
The label “peer-to-peer lending” is technically inaccurate in the US market of 2026. The model that survived is marketplace lending, where institutional capital, not retail savers, funds most loans originated through fintech platforms. The borrower experience still looks peer-to-peer: an application, a quick credit decision, money to a bank account in a few days. The funding side is dominated by hedge funds, asset managers and bank balance sheets buying whole loans or securitisations. Calling it P2P is a marketing artefact rather than a structural truth.
The mechanics matter because they explain who is exposed to which risk. In the original P2P model, a retail investor took the credit risk of an individual borrower, with the platform as a matchmaker. In the marketplace-lending model, the platform underwrites the loan, an institutional buyer takes the credit risk in bulk, and the retail-investor pool is incidental. The losses that hit retail P2P investors in the 2015-2017 vintages, and the regulatory questions raised by suitability disclosures, helped push the industry toward the institutional model that now dominates.
The shift happened gradually between 2014 and 2018, when the largest US platforms realised that retail capital was slow to mobilise and expensive to service. LendingClub, the bellwether of the original era, eventually acquired a US bank and now operates as a chartered lender. Prosper retains a marketplace structure but draws most of its capital from institutional buyers. Newer entrants such as Upstart and SoFi never operated as pure P2P; they were marketplace lenders from day one.
What the volume picture looks like
US marketplace lending volumes have recovered well from the 2022-2023 rate shock. Origination figures across the major personal-loan platforms returned to growth through 2024 and 2025, helped by improvements in credit-decision models, the cost-of-funds peak being behind the market, and a renewed appetite from institutional buyers for whole-loan exposure. The product mix has also widened: personal unsecured loans remain the largest single category, but small-business lending, auto refinance, student-loan refinance and Buy Now Pay Later all sit inside the broader marketplace-lending tent.
The dispersion across platforms is what makes 2025 interesting. Upstart leans hard on machine-learning underwriting and reports approval-rate gains over traditional FICO-only models. SoFi has built a vertically integrated bank-plus-marketplace stack. LendingClub uses its bank charter to retain a portion of its origination on balance sheet, smoothing volatility for buyers of the marketplace tranche. Each is solving a slightly different version of the same problem: how to deliver consumer credit at a unit cost that traditional banks cannot match without giving up the funding flexibility a marketplace structure provides.
The credit performance picture is the one institutional buyers watch most closely. Net charge-off rates rose sharply in the post-pandemic period as consumer debt stress increased, and several platforms tightened underwriting through 2023. The 2024 vintages have so far performed better than the 2022 vintages, but the lift has been uneven across credit bands. Lenders that pulled back early on subprime exposure are showing the cleanest results, and the buyers funding their flow are pricing that discipline into the spread.
The regulatory shape of the US market
The US regulatory framework for marketplace lending is a patchwork. Lending platforms typically partner with a chartered bank that originates the loan and immediately sells it to the platform or to institutional buyers, which preserves federal preemption of state interest-rate caps. The Madden v. Midland Funding decision in 2015 challenged that structure for non-bank purchasers, and the regulatory response since has been a mix of OCC and FDIC clarifications attempting to restore the “valid when made” doctrine. The Consumer Financial Protection Bureau has, in parallel, been more active on disclosure and fair-lending compliance.
State-level money-transmitter and lending licences add another layer. A marketplace lender that wants to operate in all 50 US states either obtains state-by-state licences, partners with a federally chartered bank, or acquires a bank itself. LendingClub took the third route in 2021 when it bought Radius Bank. Several other platforms have explored the same path. The trade-off is straightforward: regulatory clarity in exchange for capital and operational obligations that a pure-play technology platform does not face.
The next regulatory question is the boundary between marketplace lending and other fintech credit products. Buy Now Pay Later, which the CFPB now classifies as a credit-card adjacent product, has pulled new entrants into a regulatory perimeter they previously avoided. Embedded-credit offerings, which appear inside non-financial apps, are facing similar scrutiny. The dynamics look similar to the ones driving why fintech needs to build for federal employee benefits in the public-sector benefits space, where the line between fintech feature and regulated financial product has had to be redrawn in real time.
Where the retail-investor option still exists
Direct retail participation in US marketplace lending has narrowed but not disappeared. Prosper still offers a retail-notes product. Yieldstreet, Fundrise and a handful of other platforms package marketplace-lending exposure inside accredited-investor or REIT-style wrappers for the affluent retail segment. Several broker-dealers offer access to securitisation tranches of marketplace-lending pools for accounts that meet suitability thresholds. The overall slice of marketplace-lending capital coming from retail sources is small relative to the institutional base, but it has stabilised after a decade of contraction.
Retail-investor returns in the surviving channels have been more modest than the early P2P pitch suggested. Net of defaults, servicing fees and time required to redeploy capital, a typical retail buyer of marketplace-lending notes has earned mid-single digits per year across cycles, sometimes less. That is still competitive with traditional fixed-income alternatives in many years, but it does not match the early marketing claims of double-digit yields, and the realised volatility around the mean has been higher than the original product description implied.
The broader retail conversation has shifted toward tokenised credit, where blockchain-based platforms wrap marketplace-lending exposure in token form for global investor access. The parallels with the the tokenized treasuries market crossing $15 billion rollout are deliberate. The instruments are different, the regulatory wrapper is different, but the underlying argument is the same: a tokenised version of the asset can reach buyers and trade in venues that the original retail-notes structure could not.
| Indicator | 2023 Survey of Consumer Finances | Primary source |
|---|---|---|
| US median family net worth | $192,700 | Federal Reserve SCF 2023 |
| Share of US families holding debt | 77.2% | Federal Reserve SCF 2023 |
| Share of US families holding credit-card balance | 45.2% | Federal Reserve SCF 2023 |
Source: Federal Reserve Survey of Consumer Finances 2023 (released October 2023), linked above.
Where the model goes from here
Three forces will shape the next phase of US marketplace lending. The Fed’s rate path determines whether the cost-of-funds advantage marketplace lenders enjoyed in the 2010s returns. AI-driven underwriting, particularly large-language-model use in document review and alternative-data scoring, is changing both decisioning speed and the cost structure of origination. And the embedded-finance trend is pulling credit issuance into platforms that did not originate as lenders, which both expands the addressable market and complicates the compliance picture.
The competitive question for the surviving platforms is who owns the customer relationship after the loan is funded. Banks that historically dominated the consumer-credit space had decades of cross-sell to draw on. Marketplace lenders are still building the next-product roadmap that turns a refinance customer into a primary-banking customer, and the firms that get this right will probably end up looking more like neobanks than like loan marketplaces by the end of the decade.
By the end of 2026, US marketplace lending will likely look more like a structured credit asset class with a fintech distribution layer than like a retail-savings alternative. The original P2P promise is unlikely to come back at scale. What replaced it is a more efficient capital-markets pipeline that ships small consumer loans into institutional balance sheets faster, cheaper and with better data than the legacy bank channel can. That outcome is less romantic than the founding vision, and on most measures it is also a better one.
