There’s a quiet gold rush happening in the world of artificial intelligence, and it’s not happening in Silicon Valley boardrooms. It’s happening in your inbox, your photo library, and that stack of receipts sitting in your glove compartment.
Mode Inc, an AI training startup founded in 2019, has been on a shopping spree lately. The company just picked up Trimbox, an inbox management app, and QR Code Reader, a code scanning tool. That’s seven acquisitions in the past year alone. The strategy is straightforward: buy apps with massive user bases, then incentivize those users to hand over their data for cash or rewards.
“We’re not really focused on just going after the Mechanical Turk or the Outlier/Scale AI gig workers,” CEO Dan Novaes told Business Insider. “It’s really about everyday consumers and doing the things that they do every single day.”
Think about what that means in practice. Your Amazon and Walmart receipts, your streaming habits, your wearable device data. Mode is turning the mundane stuff you generate every day into a product that AI labs are desperate to buy.
The Numbers Are Staggering
The company now boasts more than 100 million monthly active users across its portfolio. To date, it’s handed out $1 billion in earnings, savings, and incentives. Novaes has raised over $80 million through crowdfunding and has publicly stated he intends to take the company public in the next two years.
This places Mode squarely in a crowded space alongside startups like Scale AI, Mercor, and Handshake, all of which pay hundreds of thousands of contractors worldwide to collect and label data for companies building self-driving systems, chatbots, and more.
What makes Mode different is its approach. Rather than relying on gig workers explicitly labeling data, it’s reaching directly into everyday behavior.
Why This Matters Now
Novaes believes the legal spotlight on AI companies using online content without consent will only increase demand for his model. Recent lawsuits against Anthropic and Perplexity over content scraping have made AI labs twitchy. The fix? Get data the old-fashioned way: with permission.
He gave a telling example. An AI lab recently requested handwritten documents. Doctors’ notes. Oil change receipts. Millions of them. Mode sent a notice to its users, offering compensation for samples, and presumably got what the client needed.
That’s the model in a nutshell. Massive scale, granular data, user consent. Whether you find that inspiring or slightly unsettling probably depends on how you feel about your digital footprint becoming someone else’s revenue stream.
The Bigger Picture
Novaes sees a clear path to a billion monthly users. “You can create the next Telegram or Twitter, or you can acquire 1,000 apps that have a million monthly active users each.” He’s clearly choosing the second option, and he’s not shy about it.
He’s reportedly looking into an app that could make users’ photo libraries available for AI training. Your vacation photos, your family snapshots, your random snapshots of receipts and documents. All fair game, theoretically, if you opt in.
The Technology implications are massive. As AI models get more sophisticated, they need more varied, more personal, more real-world data. The companies that can supply that data at scale are sitting on something valuable.
What remains to be seen is whether users will keep trading their data for rewards, or if the pendulum will swing toward privacy concerns that make this business model harder to sustain. For now, Mode is betting that convenience and compensation will win out. The Business of your data is only getting started.
