Spotify says its best engineers aren’t writing code anymore, AI is doing it

Artificial Intelligence

• Spotify uses AI to speed up coding and product releases
• Engineers can deploy code remotely via AI tool Honk
• Spotify claims unique music data gives it an AI advantage

Did our AI summary help?

Has AI coding hit its tipping point? Spotify appears to think so. During the company’s fourth-quarter earnings call this week, co-CEO Gustav Söderström said some of Spotify’s best engineers “have not written a single line of code since December,” underscoring how deeply AI has been woven into its development process.

The comments came as Spotify highlighted the pace of its product output. Over the course of 2025, the company shipped more than 50 new features and changes to its streaming app. In recent weeks alone, Spotify has rolled out additions such as AI-powered Prompted Playlists, Page Match for audiobooks, and About This Song.

At the centre of this shift is an internal system called “Honk,” which Spotify uses to speed up coding and product velocity. According to the company, Honk enables remote, real-time code deployment using generative AI, specifically Anthropic’s Claude Code.

Söderström described a scenario where an engineer, while commuting, can instruct Claude via Slack on their phone to fix a bug or add a feature to Spotify’s iOS app. Once the AI completes the task, a new version of the app is pushed back to the engineer on Slack, ready to be merged into production before they even arrive at the office.

Spotify said the system has “tremendously” accelerated both coding and deployment, and framed the current setup as an early stage rather than an endpoint. “We foresee this not being the end of the line in terms of AI development, just the beginning,” Söderström told analysts.

Beyond productivity gains, Spotify also argued it has a strategic advantage in AI thanks to the unique datasets it is building. Söderström said music-related data cannot be easily commoditised by large language models in the same way as factual sources such as Wikipedia, because many music questions are subjective rather than definitive.

Workout music, for example, varies widely by geography and personal taste. While hip-hop dominates overall listening in the US, millions prefer death metal. Across Europe, EDM is popular for workouts, but heavy metal has a strong following in Scandinavia. Spotify believes this kind of nuanced, preference-driven data gives it a defensible edge as it retrains its models.

Analysts also pressed the company on AI-generated music. Spotify said it allows artists and labels to disclose how tracks are made through metadata, while continuing to police the platform for spam and abuse.