After 11 years, Tala still isn’t profitable. But it has revamped its technology and developed a new underwriting approach, and it aims to enter six new countries and break even by early 2026.
In March 2020, the Covid pandemic was beginning to devastate the finances of consumers in the Philippines. The country’s government had imposed what became one of the longest military-enforced lockdowns in the world, making paying bills impossible for many citizens. Across the ocean in Santa Monica, the consequences were dire for Tala, the fintech company that makes small loans of up to $500 to low-income consumers in the Philippines, Mexico and Kenya.
Historically, 10% of Tala’s customers had regularly failed to pay back their loans. That rate tripled in the second quarter of 2020. While most American fintech companies were enjoying a spike in digital transactions, Tala all but shut down most of its business, temporarily going into what entrepreneurs call “cockroach mode.” It had been lending out $80 million a month, but abruptly slashed that to $3 million.
Shivani Siroya, Tala’s 43-year-old founder and CEO, says one of her biggest fears was that she wouldn’t be able to pay back Tala’s lenders, which could have disastrous ripple effects. She thought at the time, “We have 600 people at the company. How are we going to protect their jobs?”
Tala laid off 20% of its customer service staff in the Philippines and Kenya and cut other costs dramatically. The emergency measures paid off–after a year, Tala was able to return to its pre-pandemic lending levels. It raised funding in 2021 at a valuation of $800 million from investors including fellow fintech lender Upstart, Kindred Ventures and Revolution Growth.
Since then, the company has been growing steadily. Revenue is up 35% from a year ago and is now running at an annualized rate of $340 million. It has 1.8 million active, revenue-generating customers. Against that background, Tala has launched an ambitious plan to double its lending by the end of 2027 through expansion to new countries, and it has new technology for assessing risk that it says allows it to safely approve more individual applicants.
A lot is hinging on this. After 11 years of lending, Tala is still losing money, though Siroya says the company “could be profitable at any time,” if it gave up on its growth ambitions. In classic Silicon Valley-style, it’s betting it can grow its way to profitability and do so in a business that is inherently risky. (Siroya expects Tala will break even by the end of the first quarter in 2026.)
Some major fintech lenders have followed the same strategy. Affirm launched in 2013, had one profitable quarter in 2020 and just recently became profitable again, in the second quarter of 2025. SoFi has a similar story, launching in 2012 and becoming consistently profitable at the end of 2023.
After operating in just three countries for most of its history, Tala this week entered a fourth, Guatemala. Over the next three to six months, it plans to go live in five more: the Dominican Republic, Panama, Peru, Vietnam and India. It plans to do that with a rebuilt tech infrastructure that lets it make more personalized risk assessments and expand faster into new countries.
Have a story tip? Contact Jeff Kauflin at [email protected] or on Signal at jeff.273.
When Siroya founded Tala in 2011, she set out to build what sure sounds like one of the hardest businesses in financial services to make profitable: lending money to low-income consumers in developing nations. Even though credit reports weren’t available for potential borrowers, and verifying their income was difficult, she devised a creative way to assess their risk: requiring access to their cell phone data and using that data, including text-message receipts, to extract information such as whether people paid their bills on time and how old their phones were.
Tala assigned consumers its own home-grown credit score and placed them into broad risk buckets, ranging from low to high risk. Its data was novel, but the conservative tiering approach was “manual” and “not personalized,” says Kelly Uphoff, Tala’s chief technology officer who joined the company four years ago, after nearly a decade at Netflix leading its content and marketing data strategy. “We knew we were turning people down who would have been good customers,” Uphoff says.
After starting at Tala, Uphoff led the effort to build a new data engine, moving Tala away from imprecise credit-score buckets and toward personalized risk assessments. The new system utilizes open-source tools Uphoff used at Netflix and a statistical approach called causal inference, which involves continuously running a large number of small tests to draw conclusions without having access to experimental data and a control group.
Beyond its new statistical models, Tala has started using different data sources for evaluating borrower risk. That’s partly because it doesn’t want to rely on information from Google’s Android mobile operating system, since the tech giant could decide to revoke access to that data. Now Tala scrutinizes more internal information on how consumers use its app, such as whether they speed through the loan application, in addition to factors like a person’s education level and whether they already have other outstanding loans.
So far, Tala has rolled out this new underwriting model only in Mexico, though eventually it plans to use it in every market. Over the past year, the new system has led Tala to boost its loan approval rate from about 40% to up to 80%, and defaults have fallen when comparing consumers with similar credit profiles. Siroya says Tala, which lends out $145 million per month, can now build a lending model for a new country in three months, as opposed to 12 under the old approach.
Tala’s expansion includes four new countries in Latin America and two in Asia. After launching this week in Guatemala, it will open in Panama, the Dominican Republic, India and Vietnam by the end of December. In the first quarter of 2026, it plans to fully go live in Peru, where it launched a cryptocurrency-based digital wallet last year as a test product. The criteria for picking such spread-out targets? Tala is entering nations with large, underserved populations with similar income levels to the markets it already serves.
Tala has already had false starts in India, whose largest-in-the-world population of 1.45 billion makes it an obvious choice. Siroya talked about entering the country as early as 2018, and in 2020, Tala released its app there to a small number of consumers to test the market. But it didn’t have the proper lending license necessary to build an economically viable business–without it, Tala’s costs to lend were too high. It finally secured a Non-Banking Financial Company license earlier this year, and now is ready to fully launch there, Siroya says.
Making loans of $20 to $500 to customers who earn only $5,000 to $15,000 a year, on average, is a tough business for many reasons. Every loan, no matter how tiny, comes with a significant fixed cost, rendering microloans much less profitable for lenders. Investors are often leery of funding loans for poorer consumers, and such customers typically have a harder time paying back their loans. Today, about 10% of Tala’s global customers don’t pay their loans back, a default rate that’s more than 100% higher than Americans’ default rate on credit cards and 50% higher than for customers of Brazilian digital bank Nubank.
To make its business work, Tala has to charge very high interest rates. In the Philippines, it charges up to 15% per month, the maximum lenders can legally charge, leading to an annual percentage interest rate (APR) that can reach 183%. Its largest competitor there, GCash, charges a monthly interest rate of up to 7%, plus a 3% processing fee, according to its website.
In Mexico, Tala charges 24% monthly interest, or an APR of about 288%, a similar price as competitors Baubap and Kueski based on their websites’ disclosures, though Tala says it lends to lower-income customers than its peers. The company adds that, unlike credit cards, Tala’s loans don’t have revolving interest–fees don’t accrue beyond the time period of the loan.
Siroya insists that Tala can keep default risk in check. The company has 10 years of lending data to learn from, which includes a severe downturn during the first few months of Covid, she points out. Its loans are short term, about 30 days on average (though they can go up to six months), so Tala can quickly cut the amount of money it’s giving out if it wants to.
Moreover, most of its growth will come from markets it knows well; Tala spokesperson Lauren Pruneski says 75% of its revenue growth over the next two years will come from its three main markets of the Philippines, Mexico and Kenya.
Tala’s CEO defends her strategy of continuing to focus heavily on growth instead of prioritizing profitability. She points out that Tala hasn’t raised funding since 2021: “In that respect, we’ve been able to continue to pursue growth while keeping our costs sustainable.” She adds, “The mission that we’re after requires us to stay growth-focused.” She thinks the areas she has been investing most in–technology, infrastructure and data–will allow Tala to keep growing rapidly and break even by the end of the first quarter of 2026.
Branch, a San Francisco company founded in 2014 that makes small loans to consumers in India and Africa, has taken a slower, more conservative approach. After it laid off half of its staff at the start of the pandemic, founder and CEO Matt Flannery says he adopted a “self-imposed mandate to be profitable” and “applied profitability logic to every decision,” which was a dramatic change in how he ran the company. Today, Branch has 6.5 million monthly active users, 90% of them in India. It lends out $80 million a month, 55% of Tala’s monthly total, and brings in about $140 million in annualized revenue. Flannery expects Branch to make $20 million in net profit after taxes in 2025.
Like many tech CEOs, Siroya seems most concerned that her company isn’t moving fast enough. She says one thing that keeps her up at night is the growing availability of data and advancements in AI, the combination of which could allow well-resourced tech companies to launch lending businesses faster than ever before.
So she’s continuing to push Tala into new areas beyond its six-country expansion. She’s planning to launch a stablecoin-based wallet that can act like a bank account and earn interest. And she wants Tala customers to be able to receive international money transfers from senders in developed countries that Tala doesn’t operate in. One way it could do that: by partnering with other large fintech companies that already have big customer bases.
This long list of goals begs the question of whether (and more likely, when) Tala will need to raise more funding. Siroya declined to disclose how much cash Tala has in the bank. She says she’ll consider securing more financing as the company pushes to grow faster but that for now, it’s “not a core focus.”
More from Forbes
Source: https://www.forbes.com/sites/jeffkauflin/2025/09/25/unprofitable-microlending-fintech-makes-big-bet-on-global-expansion/
