The Next 20 Years of Music Technology

Thinking about the future of music… June 2026

I’ve spent most of my career thinking about music from two very different perspectives.

As a musician, I’ve spent years writing songs, recording albums, playing live shows, and trying to understand why certain pieces of music connect with people.

As the CEO of Feed.fm, I’ve spent the last decade helping companies build music experiences for millions of listeners across fitness, wellness, gaming, and connected-device platforms.

What I’ve come to realize is that the music industry has largely solved one of its biggest problems.

Access.

Today, Spotify offers access to more than 100 million tracks, and Apple Music has surpassed 100 million songs in its catalog.

For those keeping score at home, that’s more music than any human could listen to in several lifetimes.

For most of the industry’s history, access was the challenge. Rights had to be negotiated. Royalties had to be tracked. Distribution was expensive and limited. Entire businesses were built around helping people discover and hear music.

Today, access is abundant.

The next challenge is understanding.

Music licensing is a cost center. Music intelligence is about understanding.

Understanding music at the scale of 100 million tracks requires more than a catalog. It requires a platform that can organize, classify, contextualize, and continuously learn from the content itself. It requires metadata, tagging, ontologies, behavioral signals, and feedback loops that transform songs into intelligence.

We’ve learned a lot about contextual data through operating our own platform for the past decade. Every day, Feed.fm processes more than 50 million music-related API calls and powers millions of streams across fitness, wellness, gaming, and connected-device experiences.

The most important takeaway is that when everyone has access to the same catalog, access itself is no longer a differentiator. Understanding is.

One example comes from research we recently conducted with parents. We found that 77% reported their children had been exposed to inappropriate music inside apps. More importantly, 84% said inappropriate music reduced their trust in the app or brand, and 82% said they would be willing to pay for trusted, family-friendly music experiences.

Those findings reveal something important.

Listeners don’t just care about content. They care about context.

They care about whether content is appropriate for the audience, environment, and experience.

That’s where music intelligence begins.

We’ve seen this story before.

Google didn’t become one of the world’s most valuable companies because it searched web pages. It became valuable because it indexed, organized, and helped people understand the web.

The intelligence layer became more valuable than the content itself.

Google’s advantage came from classification, relationships, and context. Building music intelligence requires the same foundation.

Historically, the music industry asked a simple question:

“Can I play this song?”

Music intelligence asks a different question:

“Should I play this song?”

That distinction changes everything.

Genre alone is no longer enough. A song can be uplifting, nostalgic, aggressive, family-friendly, motivational, relaxing, empowering, or culturally significant. It can be appropriate for a fitness experience and completely inappropriate for a healthcare waiting room.

Those distinctions do not exist in traditional music metadata. They emerge through tagging, classification, ontology, behavioral data, and human judgment.

Metadata describes individual attributes. Ontologies describe the relationships between those attributes and how they evolve over time.

Content may be abundant, but context is what creates value.

Without a shared vocabulary for describing music, personalization, discovery, brand safety, recommendation engines, and AI all become significantly harder problems to solve. With a common framework for understanding music, content becomes measurable, searchable, and actionable.

Despite recent advances in AI, human judgment remains essential. Some signals are still difficult to infer from data alone. The goal isn’t human or machine. It’s human and machine.

This principle applies everywhere.

A fitness app may want different music for cool-down than the core workout. A wellness experience may require different content for meditation than sleep. A teenager may want different music for studying than for partying. At least one of those activities is more likely to improve their grades.

The music may all be licensed, but it is not all appropriate for the same audience, environment, or brand.

The same is true for engagement, personalization, and recommendation. Understanding why a song resonates with one audience but not another requires a richer understanding of the content itself.

As generative AI accelerates content creation, this challenge will only become more important. The world does not need more content. It needs better ways to understand it.

Music licensing answers the question:

“Can I use this content?”

Music intelligence answers the question:

“What is this content, and why and when should I use it?”

We have largely solved the access problem.

The next challenge is understanding.

The companies that create the most value over the next decade will not be the ones with access to the most content. They will be the ones with the best understanding of it.

Jeff Yasuda is the CEO of Feed.fm and plays in Fuzz Collective.