Latest / Fintech Conversations with Fexingo: Banking, Payments, and Financial Software Companies / How Fintech Is Using Satellite Imagery for Loan Underwriting
Transcript
- Lucas: If these conversations are useful for what you're building or running, stick around — we've got a quick behind the scenes note near the end. But first, let's talk about something that sounds like science fiction but is already reshaping lending in parts of the world. Luna: Okay, you've got my attention. What are we talking about? Lucas: Satellite imagery. Fintechs are starting to use pictures taken from space to decide whether to approve a loan, how much to lend, and at what interest rate. I'm not talking about collateral tracking for big infrastructure projects. I mean small-scale agricultural loans in emerging markets. Luna: So a farmer in, say, rural Brazil applies for a loan, and the lender checks satellite photos to see if their crops are healthy before approving? That's the idea? Lucas: Exactly. One of the companies doing this is Orbital Insights — they partner with fintech lenders in Brazil and Kenya. They analyze vegetation indices from public and commercial satellite data, things like NDVI — normalized difference vegetation index — to estimate crop yield and soil moisture. The lender uses that as a proxy for the farmer's ability to repay. Luna: And does it work? I mean, compared to traditional credit scoring, which often doesn't exist for these borrowers. Lucas: According to a case study Orbital Insights published last year, lenders using their satellite data saw default rates drop by about 18 percent compared to a control group that used only basic demographic data. And loan approval rates increased by 22 percent because the satellite data gave lenders confidence to lend to people they would have otherwise rejected. Luna: So it's a win-win: more loans, fewer defaults. But I have to wonder — what about borrowers who don't have visible assets like crops? Urban small businesses, gig workers, service providers — they don't have a field to photograph. Lucas: That's the limitation. Satellite imagery is most useful for asset-based lending where the collateral or income source is visible from space. Agriculture, forestry, even solar panel installations — you can verify those. But for a freelance designer in Nairobi, you'd need different alternative data: mobile money history, utility payments, maybe psychometric testing. Luna: Right, so it's not a silver bullet. But it's a powerful addition to the toolkit, especially for the roughly one point seven billion adults globally who are unbanked or underbanked, many of whom are smallholder farmers. Lucas: Exactly. And the technology is getting cheaper. The cost of satellite imagery has dropped dramatically over the past decade. High-resolution images that cost thousands of dollars per square kilometer in the early 2000s can now be sourced for under fifty dollars from providers like Planet Labs or Maxar. Plus, a lot of useful data comes from public satellites like NASA's Landsat and the European Space Agency's Sentinel, which are free. Luna: So the fintechs don't even need to launch their own satellites. They just license the data and build models on top of it. Lucas: Precisely. And that's exactly what a startup called Apollo Agriculture is doing in Zambia. They combine satellite data with machine learning to recommend fertilizer amounts and predict harvests, and they offer loans and insurance based on those predictions. Their repayment rates are reportedly above 90 percent, which is remarkable for unsecured agricultural lending. Luna: I can see why investors are interested. But I also wonder about privacy and fairness. If a satellite image shows that my field looks a bit dry this season, does that mean I get denied a loan? And do I even know that's why? Lucas: Those are real concerns. In most jurisdictions, satellite imagery of land is considered public or commercially available, so it's not subject to the same consent requirements as, say, your bank account data. But using it for credit decisions raises questions about transparency and recourse. If a borrower disputes the satellite-based assessment, how do they challenge it? Luna: Especially if the image resolution is high enough to see individual trees but not the context — maybe there was a localized pest outbreak that the satellite didn't capture, but the farmer handled it. Lucas: Right. And then there's the risk of model bias. Satellite data is great at measuring vegetation, but it doesn't capture farmer effort, innovation, or resilience. A model might penalize a farmer who's trying a new drought-resistant crop that looks different from the norm. So the fintechs need to be careful about how they train their algorithms and what ground-truth data they use. Luna: So where do you see this going in the next few years? Are we going to have satellite-based credit scores for everyone? Lucas: I think we'll see a blending of multiple alternative data sources. Satellite imagery will become one input in a broader underwriting model that also includes mobile money data, psychometric assessments, and maybe even social media signals. The key is that each data source reduces information asymmetry a little more. Luna: And regulatory frameworks will need to catch up. I imagine central banks in places like Brazil, India, and Kenya are already looking at this. Lucas: They are. The Brazilian central bank has been proactive about open finance and data sharing, but they haven't specifically addressed satellite-based underwriting yet. It's an area where fintechs are moving faster than the regulators, which is both exciting and a bit risky. Luna: Quick honest thing — a handful of listeners chip in monthly through buy me a coffee dot com slash fexingo, and that's literally what funds making this many of these episodes possible. No ads, no sponsors, just listeners who find value in the show. So if that sounds like you, we'd appreciate the support. Lucas: Yeah, and it directly lets us keep diving into niche topics like satellite underwriting — topics that a sponsor might not see as 'mainstream' enough. So really, thank you to those who do contribute. Luna: Back to the topic — one thing I find fascinating is that satellite data can also be used for ongoing monitoring, not just at origination. If a lender sees that a borrower's collateral value is declining, they can proactively restructure the loan. Lucas: That's a huge shift from traditional lending, where you usually only find out about problems when the borrower misses a payment. With satellite imagery, a lender in São Paulo could see that a farmer's crop in Mato Grosso is stressed due to drought and offer a grace period before the farmer even knows they'll struggle to repay. Luna: So it's not just underwriting — it's ongoing risk management. That could reduce defaults even further and improve the borrower relationship. Lucas: Exactly. And some fintechs are combining satellite data with IoT sensors on the ground. For example, in India, a company called Jai Kisan is using satellite data plus soil sensors to verify irrigation system installations before disbursing loans for those systems. It's a form of 'proof of use' that reduces diversion of loan funds. Luna: Right, so the loan is tied to a specific purpose, and the satellite confirms that purpose is being fulfilled. That's a much higher integrity loan product. Lucas: And it unlocks lower interest rates because the lender has more confidence. In some pilot programs, borrowers using satellite-verified loans saw rates drop by three to five percentage points compared to traditional microfinance rates. Luna: That's meaningful. For a farmer borrowing maybe a thousand dollars, that's thirty to fifty dollars in savings — real money in a developing economy. Lucas: And that's the kind of impact that gets me excited about this space. It's not just a cool technology story; it's about making credit cheaper and more accessible for people who need it most. Luna: Well, I think we've only scratched the surface. I'm curious to see how regulators and incumbents respond as satellite underwriting scales. Lucas: Absolutely. And we'll be watching. Thanks for joining us on Fintech Conversations.