‹ Back

Spotify says Engineers Haven’t Coded Since December

 

 

Spotify’s senior engineers have not manually written software code since December 2025, after the company deployed an internal generative artificial intelligence system to perform most coding tasks, executives said during the firm’s fourth-quarter earnings call.

 

The shift reflects a deep integration of AI into Spotify’s engineering workflow, with the technology handling bug fixes, feature creation and deployment. Spotify characterized the change as a strategic evolution in its development operations and cited enhanced product output in 2025 as evidence of the new process.

 

AI Replaces Manual Coding in Spotify’s Engineering Workflow

 

Spotify disclosed at the February 10 earnings call for the fourth quarter of 2025 that its most experienced software developers “have not written a single line of code since December,” according to co-CEO Gustav Söderström. Rather than composing code manually, engineers now issue instructions to an internal AI system named Honk, which generates the required code and assists in deployment.

 

The system, built on Anthropic’s Claude Code technology, is integrated with the company’s internal tools, enabling engineers to direct tasks via Slack, including bug fixes, feature implementation and app updates, from remote locations. Once Honk completes the tasks, updates are pushed back to engineers to review and merge into production.

 

Spotify executives emphasized that engineers remain involved in the process, but their role has shifted from hands-on coding to supervising and validating AI-generated output. This includes architectural decision-making, product specification and quality review, with the AI handling execution and repetitive coding tasks. According to Söderström, this arrangement has allowed the company to reallocate human effort toward higher-order tasks within the software development lifecycle while leveraging artificial intelligence (AI) to accelerate routine work.

 

Honk System Architecture and Operational Impact

 

The Honk AI platform lies at the heart of Spotify’s new development process. Designed to augment engineering productivity, Honk enables real-time code generation and deployment through a natural language interface. Engineers can prompt the system via Slack, specifying required changes or new features, and the AI produces the appropriate code, which is then automatically tested and prepared for deployment. Spotify described the workflow with practical examples, such as an engineer on a morning commute instructing Honk to fix an iOS app bug and receiving a merge-ready update before arriving at the office.

 

Spotify credited Honk with contributing to a prolific product release cadence in 2025, during which the company shipped more than 50 new features and modifications to its streaming platform. Notable updates included AI-powered Prompted Playlists, the Page Match feature for synchronizing physical print to audiobook experiences, and About This Song, which provides contextual information about tracks. Executives stated that such velocity would not have been feasible without deep integration of generative AI into the development pipeline.

 

Engineering Roles Evolve with AI Assistance

 

Although manual coding has receded among senior developers, Spotify maintained that engineers remain central to its product and technology strategy. In earnings remarks, the company explained that AI handles execution, but human engineers focus on direction, strategy and oversight. This model reframes the traditional role of software engineers from crafting code by hand to defining complex problems, shaping system architecture and supervising AI outputs for quality and alignment with business objectives.

 

Spotify officials also highlighted the strategic value of proprietary data in training and improving AI models. The company is building extensive datasets tied to user listening habits and preferences that they believe cannot be easily replicated by general large language models. By training AI on this specialized information, Spotify aims to create systems that deliver more nuanced personalization and context-aware recommendations, an advantage the firm says sets its AI capabilities apart from competitors.

 

Executives described the current state as an early phase in a broader transformation, rather than a culmination, suggesting further integration of AI into software development practices is planned. Spotify’s leadership framed the shift not as a replacement of human engineers, but as a redefinition of their contributions in an AI-assisted development environment.