Stability AI is being rebuilt around tools for professional music creation, with Sean Parker and chief executive Prem Akkaraju steering the company toward audio products supported by the major record labels. TechCrunch reported on October 2 that the strategy follows a $76 million financing announced in late August. Sony, Warner and Universal were among the backers, and the labels also licensed their catalogs for training as part of the transaction.

The shift follows a difficult period for the image-generation startup. Parker, the Napster co-founder, joined an $80 million rescue of Stability AI two years ago after overspending and internal turmoil contributed to founder Emad Mostaque’s departure, according to TechCrunch. Akkaraju, described in the report as Parker’s longtime friend, became chief executive. The music push is the clearest public indication yet of what the pair have been building since that intervention.

Three music archives feed licensed material into a shared AI training chamber.
Three major record groups backed the financing and licensed catalogs for model training.

Catalog licensing is central to the new direction. Rather than describing an audio model trained without the industry’s participation, the reported arrangement makes the three largest record companies both financial supporters and providers of licensed training material. Parker told The Information that he is approaching the music business through established rules this time, while acknowledging that his earlier permission-later posture did not work well.

Stability AI has already released three new audio models and music-editing software, TechCrunch reported. The tools can produce complete instrumental tracks or shorter musical passages from written prompts. That combination points to uses beyond generating a finished song in one step: professionals could also create smaller pieces of audio and modify material as part of a broader production process.

Text, humming, and rhythmic pulses become editable instrumental music layers.
Stability AI’s tools generate instrumental audio, with humming and beatboxing controls planned.

Parker said a coming update is expected to add more direct forms of musical input. Users would be able to hum a melody or beatbox a rhythm to guide the system, according to the report. The feature has not yet been released, so its timing, performance and exact controls remain uncertain. If delivered, it would give musicians a way to direct the model with sound rather than relying only on text.

The rebuild does not by itself establish whether professional musicians will adopt Stability AI’s products at scale, and the report did not provide usage or revenue figures for the new tools. It does, however, show a notable alignment among a prominent AI company, a veteran internet entrepreneur and the three major music groups. Stability AI is betting that licensed training material, specialized editing tools and more intuitive musical controls can turn that alliance into a durable creative platform.