Latest / Fintech Conversations with Fexingo: Banking, Payments, and Financial Software Companies / How Fintech Is Using Satellite Data for Agricultural Lending
Transcript
- Lucas: So if you're a smallholder farmer in rural Kenya, you don't have a credit score. You probably don't have a bank account that generates statements. But you do have a field — and increasingly, that field is worth more to a lender than any traditional financial data point. Luna: Because they can look at it from space. Lucas: Exactly. Satellite imagery is becoming a core input for fintech lenders targeting agriculture. And the poster child for this right now is a company called Apollo Agriculture, based in Nairobi. They've been operating since 2019, and they've financed over 200,000 smallholder farmers across Kenya and now expanding into Nigeria. Luna: And the hook is that they use satellite data to estimate crop health, predict yields, and then offer input loans — seeds, fertilizer, irrigation — all delivered through mobile money. Lucas: Right. Let's unpack how the satellite part works, because it's not just taking a picture and saying 'yep, that's a farm.' The process involves multiple data layers. You've got optical imagery from satellites like Sentinel-2, which is European Space Agency, freely available. That gives you vegetation indices — NDVI is the common one, Normalized Difference Vegetation Index. It basically measures how green and dense the crop canopy is. Luna: And that correlates with biomass and expected yield. So if a farmer's field shows a healthy NDVI trajectory over the growing season, they're a lower credit risk. Lucas: Exactly. But Apollo doesn't stop there. They also layer on weather data — rainfall patterns, temperature — and soil type maps. Then they run all of that through machine learning models that predict the probability of default for each individual farmer. The model outputs a credit score, and within minutes the farmer gets an offer via SMS or USSD. Luna: And the loan isn't cash — it's in-kind. Seeds, fertilizer, maybe a soil test kit. So the money is tied directly to the productive input. Lucas: Which reduces the risk of diversion. The farmer can't spend the loan on something else. And repayment happens after harvest, often in the form of produce sold through Apollo's market linkage partners. So the whole cycle is closed-loop. Apollo claims default rates around 5 to 7 percent, which is remarkably low for unsecured agricultural lending in that context. Luna: Compare that to traditional microfinance agricultural loans, which can have default rates north of 20 percent. So satellite-based underwriting is a real improvement. Lucas: And it's not just about default reduction. It's about reaching farmers who were previously unbankable. Traditional lenders require collateral or a guarantor. Satellite data removes that barrier. The field itself becomes the collateral, in a sense. Luna: But there are challenges. Cloud cover, for one. In tropical regions, you can go weeks without a clear satellite image. How do they handle that? Lucas: They use synthetic aperture radar, or SAR, from satellites like Sentinel-1. Radar penetrates clouds. It doesn't give you NDVI, but it can measure soil moisture and crop structure. So they fuse optical and radar data. Plus they use gap-filling algorithms to interpolate between images. It's not perfect, but it's good enough for a credit decision. Luna: And then there's the human element. The satellite says the field looks good, but what if the farmer is sick, or there's a family emergency? The model can't capture that. Lucas: That's a real limitation. Apollo does have field agents on the ground who verify a sample of farms and collect local knowledge. But you're right — pure remote sensing misses social and health shocks. Some lenders are now experimenting with combining satellite data with mobile money transaction history to get a more holistic view. So it's not just about pixels. Luna: This feels like something that could scale beyond smallholder farming. Could we see satellite-based lending for forestry, or even for carbon credit projects? Lucas: Absolutely. There's a company called Pachama that's using satellite imagery to verify carbon offsets from reforestation projects. And lenders are starting to use similar data to finance sustainable agriculture transitions — paying farmers to switch to regenerative practices, with satellite monitoring as the assurance mechanism. Luna: So the same technology that underwrites a loan for fertilizer today could underwrite a carbon credit tomorrow. Lucas: Exactly. And that's where the fintech opportunity gets really interesting. Because the same data pipeline — satellite images, weather, soil, machine learning — can serve multiple financial products: loans, insurance, carbon credits, supply chain finance. Luna: Speaking of which, a couple of dollars a month is genuinely what keeps these conversations going — if you've gotten something out of this episode, buy me a coffee dot com slash fexingo makes a real difference for a small show like this. Just saying. Lucas: Appreciate that. It's listener support that keeps us ad-free and focused on the substance. So thank you. Luna: Okay, back to satellites. Lucas, you mentioned insurance — how does satellite data help with agricultural insurance? Lucas: Great question. Index-based insurance is a big area. Instead of assessing individual losses, you set a trigger based on satellite-measured rainfall or vegetation. If the index falls below a threshold, all policyholders in that area get an automatic payout. No adjusters needed. That dramatically reduces administrative costs and speeds up claims. Luna: And that's being done at scale in India and East Africa. The World Bank has backed several such programs. Lucas: Right. But the fintech twist is that these insurance products are often bundled with loans. Apollo, for instance, includes a drought insurance component in their loan package. If the satellite shows a severe dry spell, the farmer's loan is automatically forgiven or reduced. That protects both the borrower and the lender. Luna: So the satellite data serves as the single source of truth for both credit underwriting and risk mitigation. That creates a pretty resilient product. Lucas: It does. And it's bringing down the cost of capital for these lenders. Investors are more willing to fund a portfolio that has satellite-based monitoring and automatic insurance triggers. Some of these loan pools are now rated by credit rating agencies, which opens up institutional capital. Luna: But let's talk about the data privacy angle. Farmers are effectively having their fields surveilled from space. Do they have any say in that? Lucas: It's a valid concern. In practice, the satellite imagery used is typically low to medium resolution — pixels of 10 to 30 meters — so you can't see individual people or animals. It's more about the field as a whole. But the data is still personal in the sense that it's tied to a specific farmer's livelihood. Apollo and similar companies require consent as part of the loan application process. The farmer agrees to share their location and allow satellite monitoring for the duration of the loan. Luna: And what about data ownership after the loan is repaid? Does the lender retain the right to keep monitoring? Lucas: That varies by company. Some delete the field data after repayment. Others retain it for model training, but anonymized. There's no industry standard yet, and regulators are starting to pay attention. Kenya's Data Protection Act applies, and the Central Bank has issued guidelines on alternative credit scoring. So it's an evolving space. Luna: And it's not just Kenya. We're seeing similar models in India, with companies like Jai Kisan and SatSure, and in Latin America with companies like Kilimo. The concept is spreading. Lucas: It is. And what's interesting is that the technology is getting cheaper. Satellite imagery costs have dropped dramatically. A high-resolution image used to cost hundreds of dollars. Now you can get decent-quality data for a few dollars per square kilometer. And open-source data from Sentinel is free. Luna: So the barrier to entry for fintechs is lower than ever. But the machine learning expertise is still a bottleneck. Building a model that accurately predicts yield from satellite data across different crops and climates is hard. Lucas: No question. That's why many fintechs partner with agtech companies that already have the models. Apollo, for example, works with a company called aWhere for weather data, and they've built their own ML in-house. But you need a team that understands both remote sensing and credit risk — that's a rare combination. Luna: And that's probably the biggest constraint on scaling. But we're seeing universities and accelerators start to train that talent. Stanford's Center for International Development has a project called 'Satellite-Enabled Credit' that's publishing open-source models. Lucas: I didn't know that. That's exactly the kind of infrastructure that could accelerate the whole sector. Imagine a world where any fintech in Africa or South Asia can plug into a satellite-based credit scoring API, the way they now plug into identity verification APIs. Luna: That would be a game-changer for financial inclusion. Suddenly, a farmer in a remote village has a credit score based on their field's productivity, not on their proximity to a bank branch. Lucas: And that's the ultimate promise. It's not just about lending — it's about creating a financial identity for people who were invisible to the formal system. And the satellite is the enabler. Luna: Alright, so to wrap: satellite data is already reducing defaults, enabling insurance, and expanding access. The challenges are data quality, privacy, and talent. But the trajectory is clear. Lucas: Yeah. And the next frontier might be real-time monitoring — using constellations of small satellites to get daily images, rather than weekly. That would allow for dynamic loan adjustments during the growing season. But that's for another episode. Luna: Absolutely. Thanks, Lucas. Lucas: Thanks, Luna.