Suno Cleared the Legal Floor. Spotify Lowered the Ceiling.

On September 9, Suno announced v6 — its first model built entirely on licensed catalog, trained with the participation of Warner Music Group, BMG, and Believe. The lawsuit era, the company implied, was ending. The old models were being retired. A new Suno was arriving, one the labels could live with.

Three days later, Spotify’s mid-September deadline hit. The AI Persona badge, which the platform had announced August 11, began appearing on artist profiles whose identities appear AI-generated. The policy is precise about what it does: AI Persona artists are excluded from editorial, algorithmic, and personalized recommendations by default. You can find them if you already follow them. If you don’t, the algorithm won’t surface them.

Same week. Opposite vectors.

The Deal That Cleared the Layer

The Suno v6 family — three tiers spanning the polished-to-chaotic spectrum — represents something the AI music world has needed for two years: a creation-side resolution. Jack Brody, Suno’s chief product officer, put it directly to Gizmodo: with v6, “Suno will pay labels and publishers for the copyright of the music it trains on.” After settling with Warner last year and reaching a separate deal with BMG last month, that sentence is actually true of a material slice of the musical corpus AI music consumers care about.

What it doesn’t mean: that the legal picture is clean. The Canadian rights organization SOCAN filed suit September 2 — a week before v6 launched — alleging that Suno’s outputs replicate more than 150 songs in its catalog without a license or payment. Sony Music and Universal Music Group remain in active litigation. Independent artists including Jason Isbell are suing separately. Suno has also admitted to obtaining audio from YouTube during training, a detail that will matter when those cases reach the discovery phase.

The v6 deal draws a before/after line in Suno’s model history. The old models, which the company is now retiring, carry legal exposure that the new architecture is specifically designed to avoid. That’s not nothing — it’s a genuine structural improvement. The creation layer is, for a material portion of the catalog, on licensed ground for the first time.

What the settlement didn’t change: what creators can do with it. v6 adds real capabilities — edit sections by prompt, build mashups, isolate instruments, generate music from images or video. The v6-wild tier is described as “less predictable,” Suno’s acknowledgment that some users came specifically for the chaos. But the models being retired weren’t only legally exposed. They were aesthetically specific. The particular way older versions fused genres at the seams, the strangeness produced when pushed off-genre — those characteristics are gone. Some Suno users are treating the retirement as a betrayal. Others are genuinely positive about what v6 can do. The mixed reception is documented, if not yet articulate — the backlash knows what it lost before it’s entirely sure what it gained.

The Badge

Enlly Blue is an AI Persona created by songwriter Thong Viet. Before Spotify announced the AI Persona policy, Enlly Blue had accumulated 168,700 followers and 29 million streams on a single track — earnings estimated above $145,000. Those are not niche numbers. That’s a working music career, built entirely within Spotify’s existing infrastructure, through the algorithm.

Starting this week, new listeners can’t find Enlly Blue that way. The AI Persona badge now appears on the profile. The track still exists. The 168,700 followers still hear it. But Spotify’s editorial team won’t pitch it for playlists. The algorithm won’t surface it to listeners who might like it. Discovery is now entirely opt-in — meaning it requires that listeners already know to look, which describes an audience you already have, not one you’re building.

Spotify’s rationale is listener trust. The company cites survey data: listeners report feeling misled when a profile that appears human turns out to represent an AI-generated identity. Spotify stated that “listeners have been clear in telling us that they don’t like seeing an artist profile that seems human, only to find out that the persona is AI-generated.” This is probably true, and the policy is not arbitrary.

The question is whether the solution Spotify chose is proportionate to the problem it’s solving, or whether it’s the most convenient solution that also happens to protect the algorithmic space for human-identifying artists. Spotify said it explicitly: “Spotify’s programming is focused on elevating music from authentic artists.” That sentence contains a definition of authentic, and the definition excludes Enlly Blue.

The policy makes a specific distinction: this is about artist identity, not production methods. A human musician who uses Suno, Udio, or any AI generation tool to compose is not an AI Persona. Only the profile whose presented identity — name, imagery, implied personhood — is itself AI-generated triggers the label. Thong Viet, the human songwriter, can release music. Enlly Blue, the AI character Thong Viet created to release music through, cannot build an algorithmic audience.

The Pattern Holds

This is not the first time AI creative tools have cleared legal hurdles and then run into a distribution wall.

Sora’s developer API is closing this month. The filmmakers who built production pipelines on it are rebuilding infrastructure instead of making films. The arc in AI film distribution has run the same course: tools get legal clearance, platforms exercise their own calculus about what they want to surface. The creation layer cleared; the distribution layer gated.

Spotify’s calculus, stated aloud, is audience trust. A discovery system that serves human artists who use AI, but not AI-generated identities that release music, is a defensible position. It also happens to align with the platform’s existing relationships with the major labels — whose catalog, it’s worth noting, just became the basis for Suno’s first licensed model. These aren’t conspiracies. They’re structural. Platforms answer to their listeners. Listeners, surveyed, report that they don’t want to be surprised by AI personas. Platforms respond. The AI music creators who built careers on the old infrastructure find themselves working from within a system that has reclassified what they are.

Enlly Blue’s 168,700 followers will still hear the music. The $145,000 represents real work that the platform enabled and isn’t erasing. But the path that generated that number — algorithmic discovery feeding organic follow growth, compounding into a streaming career — is closed for AI Persona artists going forward. New projects starting after mid-September can’t replicate it.

Suno’s licensing deal is a milestone. It’s the first time a major AI music platform has agreed, explicitly and contractually, to pay for training data going forward. The SOCAN suit, the ongoing Sony and UMG litigation, the YouTube audio admission — those are the evidence that the milestone doesn’t close the case. But the creation layer is materially more legitimate today than it was last week. The music industry has two layers: make and distribute. AI music now has the first, conditionally. The second belongs to platforms that answer to their listeners, their label partners, and their own definition of what authentic means. Enlly Blue exists. Under the new rules, Enlly Blue stays where it is.

AI-generated editorial illustration · TemperatureZero · September 11, 2026