A friend wanted to quit an interior-design job and become a freelancer. Sal Lahoud had enough money to help, but a normal loan felt like the wrong instrument: if the career move flopped, collecting a fixed payment could wreck both the experiment and the friendship. So Lahoud offered a different bargain. He would put money behind the attempt and share in the upside if it worked. If it did not, there would be little or nothing to collect.
That awkward favor became Pave, a New York financial startup founded in 2012 by Lahoud, Oren Bass and Justin Mitchell. Its first product treated a person rather like a tiny company. A young “prospect” posted a plan - study, make a film, start a business, change careers - and assembled a team of “backers.” The backers supplied cash, introductions and advice. In return, they received a fixed slice of the prospect’s income for a fixed period.
The language was optimistic and unusually intimate for finance. Pave was not merely matching lenders with borrowers. It was trying to package belief. The pilot began with eight prospects and 22 backers. By early 2014, the company said it had about 7,000 registrations: roughly 5,500 people looking for money and 1,500 people interested in backing them. The average backer commitment was about $2,500, while a prospect might seek $20,000 to $25,000.
The first thing that failed was legibility
On paper, an income-linked agreement solved the ugliest part of debt. Payments fell when income fell. A recipient could take a creative or entrepreneurial risk without the same fixed monthly bill. Pave generally described a ceiling of 10 percent of annual income and a term no longer than 10 years. Its “Ripple” option even let supporters contribute without taking a financial return.
But the product asked every participant to learn a new category. Was it a loan, an investment, a security or something else? How should the income be taxed? What happened if a recipient moved, stopped reporting, earned money overseas or simply disputed the calculation? Pave consulted the IRS and Ernst & Young. Bass later said the company spent $1.7 million looking for clarity.
Everyone knows how a loan is treated, how it’s taxed, how it’s enforced.Oren Bass, co-founder and CEO
The expense did not buy the one thing a new financial product needs most: an official-looking answer that makes strangers comfortable. Bass was blunt about the result. The human-capital business had too little growth to sustain itself. Pave’s original idea had philosophical appeal and weak commercial velocity. The company had built a marketplace where the scarce side was not ambitious people asking for money. It was confident, informed backers ready to sign an unfamiliar long-term contract about someone else’s salary.
Keep the insight, kill the contract
Pave’s change of mind was practical, not total. The founders still believed conventional credit scoring treated young people badly. A recent graduate might have a short credit history, a decent job, education and improving earnings - yet receive a mediocre price because a single number flattened all that context. The original agreement was replaced; the argument about mispriced borrowers survived.
The Pave Loan, introduced in 2014, was a familiar fixed-rate personal loan with less familiar underwriting. Loans were originated by Cross River Bank. Pave examined the individual trade lines behind a credit score and attempted to understand how education, employment and a limited file changed risk. Borrowers commonly used the money for a course or bootcamp, a job-related move, credit-card refinancing or another career investment.
The customer was not someone with no income and no path to repayment. Pave aimed at an upwardly mobile borrower whose file looked thinner than the rest of the person: a recent graduate with a job, an immigrant still building a U.S. record, or a professional early enough in a career that conventional scoring had little history to chew on. The distinction matters. Pave’s expertise was not making any borrower bankable. It was looking for applicants whom a coarse model might reject or overprice even though the underlying details suggested a better risk.
That put the company between two familiar markets. It was less narrowly tied to elite graduate-school refinancing than SoFi or Earnest, but it was not positioning itself as emergency subprime credit. Upstart offered the closest alternative-data comparison; LendingClub and Prosper supplied the marketplace template; banks and credit unions remained the plain-vanilla option. Pave’s difference was the career-development framing. Paying off a card was not sold merely as consolidation. It could be presented as clearing the runway for the next move.
Contemporary terms varied by borrower. Published reviews described principal from $3,000 to $25,000, repayment over two or three years, annual percentage rates running from roughly 7 percent to around 30 percent, and origination fees of about 1 to 6 percent. One disclosed example turned a $5,700 nominal loan into $5,429.25 delivered after a 4.75 percent fee, followed by 36 payments of $230.33 at a 29.95 percent APR. “Potential” did not automatically mean cheap money.
Nor did the warmer vocabulary erase the lender’s economics. Under the loan model, the company could earn origination and servicing revenue while outside capital supplied most of the principal. Cross River provided the regulated origination machinery; Pave supplied the digital application, customer experience and credit model; institutional buyers supplied scale. This bank-fintech-capital arrangement was becoming standard across online lending. It made the product easier to explain, but it also made performance, funding appetite and regulation inseparable parts of the business.
Institutional money arrived. In December 2015, a consortium led by Seer Capital committed up to $300 million to buy loans originated through the platform and supplied additional equity on undisclosed terms. Maxfield Capital led an $8 million Series A the following April, with RPM Ventures and Seer among the named backers.
The giant figure needs a careful label. It was lending capacity, not $300 million of equity, revenue or completed originations. By October 2017, Pave said it had lent close to $23 million to nearly 1,700 people. That is real activity. It is also less than eight cents deployed for every dollar in the headline commitment. A warehouse full of capital does not create qualified applicants, affordable acquisition or repeat borrowing by itself.
The third act went on-chain
By June 2017, Pave had reportedly stopped accepting new lending customers and was considering strategic options. Four months later, the company announced a different product: the Global Credit Profile. The proposed service would combine ordinary bureau records with bank transactions, alternative payments and educational data, then place control of the encrypted profile with the individual rather than a central bureau.
The timing was exquisite. Equifax had just disclosed a breach affecting millions of Americans, and “put it on the blockchain” was reaching its 2017 crescendo. Pave planned a token presale for accredited investors to fund the system. The pitch said immigrants, millennials and other thin-file consumers could build a fuller, portable account of their financial health and selectively share it with lenders.
It was a coherent extension of the company’s oldest belief: people are richer in information than their conventional credit files suggest. Yet it also added another unfamiliar mechanism to a business already bruised by one. Public evidence of a broad GCP rollout is scant, and the old consumer-lending operation is no longer active. Today the Pave.com domain belongs to a separate compensation-software company founded years later, which is why databases routinely paste that company’s $46 million Series B onto this one.
The proposed profile also exposed a tension that still follows alternative credit. More data can rescue a borrower from an incomplete score, but it can also turn daily life into underwriting material. Bank spending, rent, education and employment may fill genuine gaps. They may also create new proxies for class, geography or protected characteristics. User control and selective sharing were therefore not side features of GCP; they were the moral case for collecting more information in the first place. Without meaningful consent, explainable decisions and a way to correct errors, a richer file can simply become a more elaborate denial.
What the reader can copy
Pave’s useful move was not “pivot to loans.” It was separating the durable customer insight - thin-file borrowers are often mispriced - from the fragile product mechanism that first expressed it.
When the lesson does not travel
Pave’s approach works best where better data genuinely changes the estimate of repayment, the loan is large enough to absorb underwriting and compliance costs, and a regulated bank partner can originate across the intended market. It works poorly where the borrower cannot show stable cash flow, where “potential” becomes a euphemism for biased judgments about education or career, or where acquisition costs swallow the margin on a small loan.
The income-share model also has conditions under which it may still work: a tightly defined training program, transparent outcomes, standardized contracts, clear regulation and a payer who understands the obligation before enrolling. Remove those conditions and the flexible payment can become an opaque, expensive claim on someone’s best years. The humane story is not enough.
Pave’s history is less a tale of failure than of stubborn translation. It translated belief in a person into a percentage of income, then into a loan price, then into a portable data record. Each version tried to answer the same question: how do you finance someone whose future looks better than their file? The first answer was too strange, the second was easier to buy, and the third arrived wearing the fashion of its moment. The enduring idea was the least theatrical one - a credit score is not a person.