10,000 Spotify Users Sold Data for AI Tools

Lead

In June 2025, roughly 10,000 members of a collective called Unwrapped voted to sell portions of their Spotify listening preferences to AI developers, drawing a swift legal and policy pushback from Spotify. The group—organized through the decentralized data platform Vana—collected about $55,000 for the dataset, with individual contributors receiving roughly $5 in cryptocurrency. Spotify says the project violates its developer terms and risks infringing the company’s Wrapped trademark; Unwrapped counters that users are exercising legally protected data-portability rights.

Key Takeaways

  • About 10,000 Unwrapped members (part of a broader >18,000 sign-up pool) voted in June 2025 to sell a subset of listening-preference data to Solo AI for $55,000.
  • The sale passed with 99.5% of participating votes; contributors averaged approximately $5 in crypto tokens each.
  • Spotify sent a notice to developers saying Unwrapped breaches its Developer Terms (prohibiting use of Spotify content to train ML/AI) and warned of trademark concerns around the name “Unwrapped.”
  • Unwrapped and Vana say they received no direct warning and insist the collective enables users to exercise data portability under existing privacy frameworks.
  • Vana co-founder Anna Kazlauskas said onboarding bottlenecks limit daily additions to about 300 users and that crypto payouts deter some participants.
  • Digital-rights advocates (including the EFF) emphasize user control over exported data but caution about privacy risks when selling personal information to AI firms.
  • Vana is partnering with Flower AI on a planned 100-billion-parameter model; proponents argue decentralized pools could rebalance power in AI development.

Background

Spotify’s annual Wrapped product, introduced in 2015, aggregates a listener’s top artists and tracks and has become a cultural moment each year. That popularity has driven demand for richer, more personalized analyses that Wrapped does not produce, such as mood-tracking across months or cross-platform behavioral correlations.

In February 2025, a group of users organized under the Unwrapped name on Vana to pool exportable listening data and vote as a bloc on whether to license aggregated datasets to third parties. The model mirrors earlier experiments in data co-ops and data dividends that attempt to give individuals collective bargaining power over personal information.

Main Event

In June 2025, Unwrapped members approved selling a narrowly scoped dataset—artist-preference signals—to Solo AI; the sale generated $55,000 in total. Vana’s governance token framework recorded a 99.5% affirmative vote among participants at the time, and Kazlauskas described the payout as symbolically important despite modest per-user payments.

Spotify told Ars Technica that it mailed a legal-style notice to the contact details listed for Unwrapped developers, saying the collective potentially infringes on Spotify’s Wrapped trademark and breaches developer rules that forbid using Spotify content to train machine-learning models. Spotify’s statement also emphasized that users can export their personal data but that collection, aggregation, and resale to third parties contravenes its terms.

Unwrapped’s organizers counter that they never received Spotify’s notice and that their infrastructure only facilitates users exercising existing data-access rights. They maintain the project does not redistribute Spotify content nor impair Spotify’s service, but rather enables collective monetization of personally generated data.

Analysis & Implications

The dispute sits at the intersection of three tensions: corporate platform control, individual data rights, and the growing appetite for training data among AI builders. Spotify’s developer policy explicitly limits how platform-derived data may be used in AI training; enforcement of that policy against user-directed exports raises novel legal and practical questions about who controls derived data.

From Spotify’s vantage point, allowing pooled sales could create competitive risks (models trained on aggregated user behavior might replicate or undercut Spotify features) and legal exposure if user exports include copyrighted content or metadata that Spotify licenses. For users and privacy advocates, the case highlights data portability as a user right that can be exercised collectively to command economic value.

Economically, the transaction value ($55,000 total; ~ $5 per user) illustrates both the current bargain status of raw behavioral signals and the coordination costs facing decentralized pools. Low per-capita payments and crypto-based settlement systems may limit mainstream participation unless onboarding and payout mechanisms become more familiar and fiat-friendly.

Policy shifts could change the calculus quickly. Utah’s model for near-real-time API access and other data-rights laws would make portability smoother and reduce technical gatekeeping. Conversely, stronger enforcement by platforms or new terms limiting resale could dampen the prospect of user-led data markets.

Comparison & Data

Metric Wrapped (Spotify) Unwrapped (Vana)
Launch year 2015 February 2025
Members (approx.) Hundreds of millions (users) >18,000 sign-ups; ~10,000 at sale
Sale (June 2025) $55,000 to Solo AI; ~$5/user
Developer policy Protects platform content and restricts ML training Enables user data export, collective sales

The table underscores scale differences: Spotify’s consumer reach is global and measured in the hundreds of millions, while Unwrapped is a nascent, niche collective. That gap shapes both technical friction (rate-limited exports, API controls) and legal leverage.

Reactions & Quotes

“All of our users can receive a copy of their personal data to use as they see fit,” Spotify said, adding that third-party aggregation and sale may breach developer terms.

Spotify spokesperson (company statement relayed to Ars Technica)

Context: Spotify emphasized data portability in principle while warning that aggregation and resale of streaming data to third parties violate its Developer Terms and could infringe trademarks.

“A single Spotify user can’t meaningfully sell their individual data—pools are what make this work,” said Vana co-founder Anna Kazlauskas, framing Unwrapped as collective bargaining for data value.

Anna Kazlauskas (Vana co-founder, quoted to Ars Technica)

Context: Kazlauskas characterized the $55,000 sale as proof of concept and noted UX and crypto-payment barriers that limit broader participation.

“Listeners should have control of their own data, which includes exporting it for their own use,” said Jacob Hoffman-Andrews of the Electronic Frontier Foundation, while warning about privacy risks when data is sold.

Jacob Hoffman-Andrews (EFF Senior Staff Technologist)

Context: EFF endorses user control and portability but cautions that monetizing sensitive personal records can create privacy harms.

Unconfirmed

  • Whether Spotify’s notice was actually sent to Unwrapped developers or whether delivery failed: Unwrapped says it never received the contact; Spotify says it sent a letter to listed contacts.
  • Extent of Spotify’s technical interference with export tools: Unwrapped alleges repeated shutdowns of simple export paths; independent verification of each incident is limited.
  • Exact downstream uses Solo AI will make of the purchased dataset and whether additional purchases or larger-scale licensing deals are imminent remain unverified.

Bottom Line

This episode crystallizes a growing fault line: users and decentralized platforms want to reclaim economic value from personal behavioral data, while large platforms insist on controls to protect product features, licenses, and trademarked properties. The initial $55,000 transaction is small in corporate terms but large enough symbolically to trigger a legal and policy response.

Expect further friction unless regulators clarify rules for portability and resale or platforms adjust developer terms to explicitly permit certain user-directed uses. Meanwhile, technical frictions (API throttling, crypto payouts, onboarding complexity) will likely keep most data-pool experiments niche unless legal and UX barriers fall.

Sources

Leave a Comment