Suno powers 92.8% of AI tracks that reach streaming, and its share is still growing

Suno made 92.8% of the AI music that reached streaming services between April and August 2026, according to the Sonic Intelligence Academy’s Mid-Year 2026 report. In the Q1 edition the figure was 90.4%. Everyone expected the tool field to spread out this year. It got tighter.

The report covers 1,743 verified releases from 886 artists in 54 countries. Every track was commercially distributed and every tool credit was disclosed by the creator who submitted it.

What Suno’s 92.8% share actually measures

SIQA’s pool is not all AI music. It is AI music that got released, went through a distributor, and was submitted to a chart. That distinction matters, and it cuts in Suno’s favour rather than against it.

My read goes one step past what SIQA prints. If 92.8% of AI music that reaches Spotify and Apple Music came from one tool, then roughly 9 in 10 AI tracks on the streaming services you release to are probably Suno output. SIQA does not make that claim and I can’t verify it against platform data. Nothing else in the report points anywhere else.

Why ChatGPT’s share of AI music workflows fell to 16.3%

ChatGPT is the only other tool with a real presence, at 16.3% of workflows, used mostly for lyrics and composition. In Q1 it sat at roughly 1 in 5.

SIQA reads that drop as creators finding Suno’s own lyric tools good enough to stop switching apps. That tracks with what I see when I run these tools: the fewer windows a workflow needs, the more likely someone finishes a song.

Everything else is a rounding error. BandLab appears in 3.4% of tracks for production and mixing help. ElevenLabs, Udio, Soundraw and Moises each sit below 1%.

Udio at under 1% is the number I’d look at hardest. Its walled-garden policy stopped users from downloading and monetising what they made, and a dataset built on commercial releases cannot see a tool whose output never reaches a distributor. A restriction on export reads, in data like this, as absence.

What one tool at 92.8% means for the rest of the music industry

When a single company makes 9 in 10 releases in a category, its product decisions stop being product news and become supply news for everyone downstream.

Three consequences follow, and none of them are hypothetical:

• Every lawsuit against Suno is a lawsuit against most AI releases. The Munich GEMA ruling and the ongoing US cases land on the tool behind almost the whole category, not on one vendor among many.
• Suno’s licensing deals set the terms. Its agreements with Warner, BMG and Believe decide what a licensed AI track even looks like, because there is no second tool at scale to set a different standard.
• A pricing or policy change moves the whole market. Suno capped downloads on September 3 and retired every older model on September 9. Both changes hit most AI releases at once.

What the next SIQA report will show about Suno v6

The window closed on August 20, so none of this reflects Suno v6, which went live on September 9 and retired v4.5, v5 and v5.5 on the same day. It also predates the download caps Suno introduced 6 days earlier.

Both changes give the full-year report something to measure. Download caps put a ceiling on how many finished tracks a paying user can actually release, and a release cap is the one thing that could move a submission share.

If you release music made with AI, the practical read is narrow. You are almost certainly building on one vendor, and that vendor’s policy changes reach you before any regulator does. I’d watch Suno’s terms page the way you’d watch a distributor’s, because for most people releasing AI music this year, it functions as one.