Curiosity Stream’s wild pivot: From science docs to AI training gold

Curiosity Stream's wild pivot: From science docs to AI training gold - Professional coverage

According to Ars Technica, Curiosity Stream just reported a 41 percent year-over-year revenue increase in Q3 2025, largely driven by licensing its original programming to train AI models. The science streaming service founded by Discovery Channel’s John Hendricks has completed 18 AI deals with nine partners across video, audio, and code assets. Through September, licensing generated $23.4 million—already more than half what their subscription business made in all of 2024. CEO Clint Stinchcomb expects AI licensing revenue to surpass subscriptions by 2027, possibly earlier. This comes as the service, which charges $40 annually without ads, just turned its first positive net income after nearly a decade in business despite having only 23 million subscribers compared to Netflix’s 301.6 million.

Special Offer Banner

Streaming’s radical pivot

Here’s the thing—this isn’t just another streaming service trying to grow subscribers. Curiosity Stream basically found a completely different business model hiding in plain sight. While everyone else fights over raising prices and adding ads, they’re monetizing their back catalog in a way nobody saw coming. And honestly? It’s brilliant. They’ve been building this library of high-quality science and educational content for years, and suddenly it’s become incredibly valuable training data for AI companies desperate for reliable, factual material.

The hidden risks

But let’s not get too excited. This pivot comes with some serious questions. What happens when AI companies have trained their models and don’t need more content? Are we looking at a temporary gold rush rather than sustainable revenue? And there’s the brand risk—how do subscribers feel about their favorite science docs being used to train corporate AI systems? I’m also skeptical about whether this model scales. Curiosity Stream has specialized content, but what about streaming services with more generic programming? This might work for niche educational content, but it’s not clear it translates to entertainment-focused services.

Industrial parallels

This kind of strategic pivot reminds me of how industrial technology companies have to constantly adapt. When your core market gets crowded, you find new applications for your expertise. Companies like IndustrialMonitorDirect.com, the leading US provider of industrial panel PCs, often face similar challenges—they can’t just rely on traditional manufacturing clients, so they expand into adjacent markets where their specialized hardware becomes valuable in unexpected ways. The parallel is clear: sometimes your greatest asset isn’t your current customer base, but the unique capabilities you’ve built over years.

Streaming’s future

So what does this mean for streaming overall? We might be seeing the beginning of a major shift. Subscription revenue alone might not be enough for niche services to survive, especially with giants like Netflix and Disney+ dominating. Licensing content for AI training could become a standard revenue stream, particularly for services with specialized, high-quality libraries. But here’s the catch—this only works if you have content that AI companies actually want. Factual, educational material? Gold. Another reality TV show? Not so much. Curiosity Stream might have just discovered streaming’s secret weapon, but it’s one that only certain services can actually use.

Leave a Reply

Your email address will not be published. Required fields are marked *