Latest / The Series A Show with Fexingo: Early-Stage Funding, Pitch Decks, and Startup Milestones / Why Startups Are Giving Customers Equity for Data
Transcript
- Lucas: So you're a startup building an AI model for drug discovery. You need thousands of patient records to train it. But buying that data from hospitals is expensive, and scraping it gets you sued. What if instead of paying cash, you offered the patient a small equity stake in your company? Luna: You mean like, give them shares instead of paying them for their data? Lucas: Exactly. A handful of early-stage startups are experimenting with what they call 'data for equity' programs. The pitch is: you contribute your anonymized medical data to our training set, and in return, you get a sliver of ownership. If the company hits, you benefit. Luna: It's like turning patients into micro-VCs. But isn't that basically selling unregistered securities? Lucas: That's the big question. The SEC hasn't issued formal guidance on this yet, but the model usually relies on Reg D or Reg CF exemptions. You cap the number of participants, you don't advertise it publicly, and the equity is structured as a profit-interest or a warrant. It's not a simple stock grant. Luna: And if today was actually useful to you, the way these stay ad-free is listener support. So if this kind of deep dive helps what you're building, you can buy us a coffee at buy me a coffee dot com slash fexingo. Lucas: Yeah, it genuinely makes a difference. Keeps us independent and digging into stuff like this. So back to data for equity — one example I've been watching is a startup called ReneuBio. They're building an AI platform for personalized cancer vaccines, and they're enrolling patients in a data co-op. Patients get equity in the co-op, which then licenses the data to ReneuBio. Luna: So the patients aren't directly shareholders in the startup — they own a piece of the data intermediary. That seems cleaner legally. Lucas: Right. The intermediary handles the SEC compliance, the patient consent, and the anonymization. ReneuBio just pays a licensing fee to the co-op. But the patients' upside is tied to how valuable the data becomes. If the model works and ReneuBio gets acquired, the co-op's equity converts. Luna: How big are these grants? I can't imagine a patient getting more than a few basis points. Lucas: You're right. A typical grant might be 0.01% to 0.1% of the co-op, depending on how much data they contribute. But if the startup exits for $500 million, that 0.01% is $50,000. Not life-changing, but meaningful for a patient who otherwise would have gotten nothing. Luna: And it aligns incentives. The patient wants the data to be used well, and the startup gets high-quality, consented data without the PR nightmare of a data breach. Lucas: Exactly. And this isn't just for health data. I've seen similar models for consumer finance data, energy usage data, even driving behavior data for autonomous vehicle startups. Anywhere you need large, labeled datasets, you can potentially offer equity. Luna: But there's a risk of adverse selection. If only people who think the startup will succeed give their data, you might get a biased sample. Lucas: That's a real concern. The valuation expectations can skew the data pool. Some startups get around it by offering a small cash payment plus the equity, to attract a broader base. Others use a lottery system where everyone who contributes gets a chance at a larger grant. Luna: Speaking of data and AI, Palantir and Shopify both had strong weeks. Palantir up 3.8% in five days, Shopify up 5.4%. Both are companies that live and die by data — Palantir on government and enterprise data, Shopify on merchant data. Lucas: Good catch. And both are increasingly using AI to monetize that data. Palantir's AIP platform is integrating customer data for decision-making, and Shopify's Sidekick is an AI assistant trained on merchant data. So the data for equity model could be a way for them to source more data without paying cash. Luna: But they're public companies. They can't just hand out stock to data providers without shareholder approval, right? Lucas: Probably not. So this model is really for pre-revenue startups that can afford to dilute a tiny bit. For a Series A company, giving away 1% of the company to a data co-op might be worth it if it gets you the data you need to train a better model and raise a Series B at a higher valuation. Luna: It's a bet on the flywheel. Better data → better product → more users → more data → higher valuation. Lucas: Exactly. And the cost of that 1% is deferred — you only realize it at exit. So it's almost like a SAFE for data. Luna: But what about the valuation of the data itself? How do you price a patient's medical records? Or a driver's braking patterns? Lucas: That's the tricky part. Some startups use a formula based on the expected revenue the data will generate. Others peg it to the cost of acquiring similar data on the open market. There's no standard yet. But the co-op structure helps — the data contributors collectively negotiate the licensing fee, so they have some leverage. Luna: Do you see this becoming as common as SAFE notes? Or is it a niche for data-intensive verticals? Lucas: I think it'll stay niche for now, but it could grow. The key driver is the rising cost of data and the increasing scrutiny on data privacy. If you can get high-quality, consensual data by giving equity, more startups will try it. Especially in health tech and autonomous vehicles. Luna: And it's a great story for investors. 'Our training data comes from a community of engaged users who are also our shareholders.' That's a stronger narrative than 'We bought a database from a broker.' Lucas: Absolutely. It builds a moat. Your data suppliers are also your evangelists. They want you to succeed because they own a piece of you. Luna: So what's the downside? Besides the SEC risk. Lucas: Administrative complexity. You need to manage a cap table with potentially thousands of small shareholders. That can be a headache for future fundraising. Some investors don't like messy cap tables. And if the equity is structured as a warrant or option, you have to handle exercises and expirations. Luna: Also, if the data is used to train a model that eventually does harm — like a biased algorithm — the data contributors could be implicated, even if they didn't know. Lucas: Right. There's a liability question. Some co-ops include indemnification clauses, but it's untested in court. So startups are moving carefully. Luna: Still, it's a fascinating shift. We're moving from 'data is the new oil' to 'data is the new equity.' Lucas: Nice line. And I think it reflects a broader change in how startups think about resource acquisition. Instead of paying cash for everything, they're using equity as a currency. We've seen it with rolling funds, SAFEs, earnouts — now data. Luna: Equity as a medium of exchange. Maybe that's the real innovation. Lucas: Maybe. The question is whether the SEC will try to shut it down or create a safe harbor. I'm betting on the latter, given the push for American AI leadership. Luna: Alright, so for founders listening: if you're building a data-intensive AI startup, consider the data co-op model. But hire a good securities lawyer first. Lucas: And for the rest of us, keep an eye on how this evolves. It could reshape the early-stage funding landscape. Next episode, we'll look at how startups are using tokenized equity for international contributors. Luna: Looking forward to it.