Spotify plans Q1 2027 launch for AI covers at $4–$5/month. Why distributors should read this closely.

The Financial Times reported today (October 30) that Spotify plans to launch an AI-powered covers and remixes tool in Q1 2027, with pricing weighed at $4–$5/month on top of existing subscription fees. The report cites people familiar with the plans and places the unveiling around the Grammys on February 7, 2027. Spotify declined to comment, and the FT's sources cautioned that neither the timing nor the pricing is finalized.

Under the proposed plan, paying subscribers would use AI to make their own versions of songs by artists who opt in — reworking a folk track as dance, or rendering a cover that sounds like a different artist. Royalties on fan-made remixes would flow to the original artists, and remix plays would roll up to the original tracks on Spotify's charts.

This is a split, not an about-face

For most of 2026, the DSP posture around AI music read as one thing: block, remove, enforce. Spotify alone removed 75 million AI tracks in twelve months. Deezer flagged a majority of daily uploads as AI and began retroactive takedowns. FUGA and Qobuz turned on AIGC-flag enforcement at ingest.

The Spotify covers tool does not reverse that posture. It carves a sanctioned lane next to it. Two categories of AI are now operating under different rules on the same platform:

• Sanctioned AI — the original artist opted in, the AI output is generated inside Spotify's product surface, royalties flow back, chart credit routes to the original. Monetized.
• Unsanctioned AI — uploaded through a distributor as if it were an original work, no opt-in, no royalty route back. Still gets removed, as the 75M number keeps proving.

The Believe/Suno partnership in September had already made the licensed-vs-unlicensed split operational at the distributor layer. Spotify is now building the same split at the consumer and chart layer. The market is bifurcating the AI question into "with whose permission" instead of "yes or no."

Why the distributor sits on the wrong side of this split by default

A distributor cannot see Spotify's opt-in registry. It cannot see the AI engine that produced a cover upstream. What it sees is a file arriving in the pipeline with metadata. If that file is a Spotify-sanctioned AI output being re-uploaded as if it were an original — or an unsanctioned Suno cover of an opted-in artist's track — the distributor is the party that delivered it. The chargebacks, the takedowns, and the inevitable publisher-side claims land on the distributor's ledger.

The economic shape of the problem gets sharper under the new model. If royalties on a sanctioned Spotify cover route back to the original artist, and an unsanctioned cover of the same work ships through a distributor claiming the uploader as the author, the original artist now has two things they didn't have before: a monetary interest in the cover ecosystem, and a direct comparison between what the sanctioned lane paid them and what the unsanctioned upload took from them. That is the shape of a dispute that lands at the distributor, not at Spotify.

Attribution, not blanket blocking

A detector that only answers "is this AI, yes or no" cannot survive this market. The answer in the sanctioned lane is yes-and-fine; the answer in the unsanctioned lane is yes-and-liability. The distinction has to be made at ingest, with evidence, before anything ships.

Our detection stack is built around that distinction. It listens to the audio and places it into four categories: human-performed, licensed AI (partner model output), Suno-generated, and unknown AI generator. The classifier does not need to see the opt-in registry — it needs to tell the distributor what kind of AI the audio is, so the compliance workflow upstream (consent record, partner attestation, chart-eligibility flag) can be applied against evidence instead of trust.

What to do this week

1. Pull the subset of your 2026 catalog where the uploader declared "cover," "remix," or "version of" and the acoustic profile looks machine-generated. Those are the tracks most likely to collide with a sanctioned-lane equivalent once the Spotify tool ships.
2. Run acoustic detection and keep the four-way attribution (human / licensed / Suno / unknown). The signed verdict is per-track evidence of what the file was on the day of ingest, independent of what the distribution contract or the metadata claimed.
3. Attach the certificate to the release record. When a chargeback or takedown lands in Q1 2027 against a track that collides with a Spotify-sanctioned cover of the same work, the per-track audit is what separates "we accepted this in good faith with evidence" from "we delivered without checking."

The FT report is still provisional — pricing may land at $3 or $6, launch may slip past the Grammys, the opt-in registry may ship in phases. The directional point does not depend on the specifics. The market is splitting AI into sanctioned and unsanctioned lanes with different royalty routes, and the distributor is the party that catches the cost of placing a file in the wrong lane. Per-track audit is the piece that converts that exposure into something auditable.

Sources

• Financial Times — reporting on Spotify's AI covers and remixes tool, Oct 30, 2026
• DistroShield blog — Believe reverses on Suno: the licensed-vs-unlicensed split at the distributor layer
• DistroShield blog — Spotify removed 75M+ AI tracks in twelve months
• DistroShield blog — Deezer: AI is a majority of daily uploads