Spotify is betting heavily on Artificial Intelligence. Gustav Soderstrom, CEO of the company, revealed that its top developers haven't written a single line of code since December thanks to AI.

While the announcement sounds extremely eye-catching, it reveals how the development and advancement of this technology is changing the way programmers implement new features and develop applications.

Spotify's premise is that AI will handle a large part of the code writing, while developers make the necessary adjustments to implement the improvements. This way, they can reduce costs and accelerate the time it takes to launch new features in the app.

What did Spotify's CEO say about not writing code thanks to AI?

The announcement was made during the fourth-quarter earnings call, where Soderstrom linked "zero code" with the use of AI to accelerate development.

The point isn't that people have lost their jobs, but rather—according to Spotify—that software is being built differently, with AI doing much of the typing while developers focus on review and decision-making. Spotify also noted that it introduced more than 50 features and changes to its streaming app during 2025, using this as a test of the product's "speed." And in recent weeks, it mentioned releases like AI-powered Prompted Playlists, Page Match for audiobooks, and About This Song, which arrived "in the last few weeks." What Spotify is implying is that AI is significantly helping them iterate faster, test more ideas, and deliver more improvements without getting bogged down in repetitive tasks. How are they using Honk and Claude Code to program faster at Spotify? Spotify explained that its engineers use an internal system called Honk to accelerate development and delivery speed. Within that flow, they mentioned Claude Code as part of the support for remote and real-time deployments with generative AI.

They gave the example of an engineer on their way to work who, from Slack on their phone, asks Claude to fix a bug or add a feature to the iOS app. When the change is ready, they receive a new version via Slack and can merge it into production before arriving at the office. Spotify stated that this "tremendously" accelerates the coding and deployment process.

The interesting thing is that this isn't just "a chatbot writing code," it's a complete integration into daily operations: messaging, automation, builds, and delivery connected to a feedback loop. When that pipeline exists, AI stops being a toy and becomes infrastructure; that's where you truly feel the productivity gains.

Now, the subtext that matters for any company is control. If you can deploy this quickly, you need clear boundaries, automated testing, well-defined reviews, and observability to detect regressions in time.

Spotify didn't go into all the details publicly, but the mere fact that they're talking about real-time deployments suggests that their approach isn't about "working magic," it's about making things happen. What does this mean for programmers and for the future of working with AI in 2026? Soderstrom also argued that Spotify can build a unique dataset that other models can't as easily commoditize, like open sources such as Wikipedia, because in music, many questions don't have a single, factual answer. He gave the example of "training music," which changes depending on the person and location, and said that they are currently building this dataset at scale and that it improves every time they retrain their models. This vision brings to a key point: AI isn't just for writing code faster; it's also for building better products if you have your own data and a continuous improvement loop. And Spotify capped it off with a phrase that's almost a statement of the times: for them, this "is not the end," it's "just the beginning" of AI development. In parallel, there were questions about AI-generated music. Spotify said it will allow artists and labels to indicate in the metadata how a song was made, while at the same time maintaining vigilance to prevent spam on the platform. That balance matters because it shows how a large platform is trying to open the door to the new without letting its house fill up with junk. If you work in programming, the takeaway shouldn't be "AI is replacing me," but something more practical: AI takes away mechanical tasks and pushes you toward the work that actually moves the needle—defining problems well, evaluating trade-offs, ensuring quality, and thinking about the user. In that world, writing code is still important, only the value is concentrated less in typing and more in directing the system, reviewing with discernment, and building products with intention.which is ultimately the true future.