Suno shipped a new music model, V6, on the morning it went on air, and shipped it in partnership with several of the largest music companies in the world.
The last three years of AI music have been argued in court and in press releases about training data. Mikey Shulman's framing is that the argument has moved: rights holders are now counterparties on products, and a share of the revenue attributable to a fan playing with an artist's catalog goes back to that artist when the artist opts in.
"And the thing that's so exciting is that for the first time in a long time, there's actually product differentiation."
Shulman co-founded Suno and runs it. Before the partnership announcements, his company was the defendant the record industry pointed at.
I listened to the full segment so you can skip it. 25 minutes of audio, 11 minutes of reading.
Here are the 10 takeaways that matter.
👤 Guest: Mikey Shulman, co-founder and Chief Executive of Suno, the AI music generator that released its V6 model the morning of this interview
🎙️ Hosts: John Coogan and Jordi Hays, who present TBPN live on X and YouTube every weekday
📰 Published: 9 September 2026 on YouTube (TBPN)
🔴 YouTube | 🟣 Apple Podcasts | 🔗 Show notes | ⏱️ 25 min | ✅ Time saved: 14 min
Key Takeaways
A host told Suno's chief executive on air that the model is excellent and the software is bad
His answer was that the company should stop shipping models while the hosts are live
V6 ships with the record companies as partners, not as opponents
Artists who opt in take a share of the revenue attributable to fans playing with their catalog
Loop speed is a measurable driver of an entertainment product, not a nice-to-have
Suno can measure how much faster returns matter and knows the size of the effect
Music has no benchmarks, so post-training is a question of taste rather than score
V6 Wild is deliberately worse at complete songs and more interesting
The market Suno opened is songs with an audience of one or two people
The product is now reachable from inside iMessage
Shulman expects AI in every part of music production within a year and no AI acts at the top of the charts
A song is a point of view, and a point of view is hard to produce in one shot
1. The Hosts' Bug Report
Shulman came on the day of a model release and the first thing he got was a critique.
One host set it up as a compliment sandwich and then delivered the filling. Suno is the AI product that has produced the most laughter for the team, and it gets pulled up in the host's own home. Then: "the product is just like amazing. It's wonderful. I get so much joy out of it. I think the software quality is bad."
The complaint is the app, not the model. The hosts were explicit that the model side is excellent and unbelievably fast, and that the problems are in navigation and bugs
The diagnosis they offered was speed of development. One host said it feels like the team has been building too fast with AI coding tools and asked for cycles spent cleaning it up
Shulman's reply was to accept the note and blame the release schedule. "We should probably stop releasing models while you guys are on air so you can actually have some time to test them out."
The hosts pushed back on their own framing halfway through, saying they had only managed one slice of bread rather than a sandwich
2. V6 and V6 Wild
The release is two models rather than one. "So V6 is better, it's faster, the songs are way better, it's more expressive."
The second is deliberately less reliable. "We have V6 Wild, which is, a model that makes just some of the most interesting stuff, even though the songs themselves may not be quite as good." Shulman said he is having a great deal of fun with it
He framed the day as an industry event rather than a product event. "And it's actually like a big day for Suno and actually for the whole music industry where this is done in concert in partnership with some of the best music companies in the world."
The hosts' own experience of the previous model was that generation had already stopped feeling slow: they would fire off a prompt, put the phone down expecting five minutes, and the song would arrive in about twenty seconds
3. The Label Partnerships
The hosts named Warner Music Group and BMG as counterparties and asked how the economics settle.
Shulman's account of what everyone wanted to know. "Everyone's been interested in how the economics will flow, what happens to songs that are used in training." He named the two poles of the argument: fair use on one side, and a demand for a royalty on everything forever on the other
His claim is that these are long partnerships, not settlements. New business lines and new revenue for rights holders, artists and creatives, and the ability to build products jointly with the largest music companies
The product consequence is the part he emphasized. "And the thing that's so exciting is that for the first time in a long time, there's actually product differentiation." His argument is that twenty years of digital music trained everyone to treat it as one undifferentiated thing
The mechanism, in his words. "When those artists opt in, when users play with their content, the revenues that are attributable to those users, a portion of that will flow to those artists."
He was open that the design is unfinished. Every side acknowledges it is a start, nobody knows what all the products should look like, and the plan is to build them together
One host's read was that this moves the industry conversation forward by at least one step
4. Loop Speed Is Measurable
Shulman's most concrete operating claim was about latency. "I think the thing that people don't realize is that when you're building things in entertainment, actually the loop speed, how quickly you give something to your users ends up mattering a lot."
It is not a hunch. He said Suno can measure the effect directly and knows how much it matters
The hosts questioned why more speed was needed at all, since generation already felt instant. Shulman's answer was that the slow case is editing — being caught by a song and wanting to change one word in the lyrics
5. Suno's Job Is to Inspire
The hosts asked about the vertical video feed inside the app, where other people's tracks arrive as fully animated music videos, and whether Suno makes those.
Right now it does not. "Right now, that is mostly creators bringing outside technologies to bear."
The feed is designed to be exited, not scrolled. "The purpose of Suno is actually to inspire you to do things that you haven't done and not just to do the old thing of scrolling again." A viewer can remix a track straight off the video
The barrier he says the product attacks is self-image. People do not think of themselves as creators, and getting somebody over that line is what he called the magical spark
He credited the effort even when he knows the mechanics. When one of the hosts sends him a Suno song, Shulman said, he still reads it as the sender arranging ideas, even knowing it came from a handful of sentences
6. Songs With a TAM of One
The hosts' theory was that Suno works because the cost of creation went to zero on markets too small to serve. Their example was a song about a corporate lawyer, which they said would not exist in any other timeline, and a Fourth of July track about the Revolutionary War somebody else had made
Shulman took the market smaller still. "you can go all the way down to a TAM of one or two people." TAM is total addressable market — the whole population a product could ever sell to
The distribution answer to that is messaging. "You can actually access Suno right from iMessage if you have the latest version of the app."
His best example was a viral one he called Oreo-gate. A woman turned her mother's text messages about her children eating Oreos into a song. Shulman's point was about the alternative: "she basically would have had to become a stand-up comic to bring this to life."
7. No Benchmarks in Music
The hosts had noticed a tell in the previous model. Songs kept returning to imagery of breaking glass, which they read the way people read a language model's stock phrases
Shulman said the absence of a scoreboard is the hard part. Unlike the large language model labs, music has no benchmarks, so questions of taste carry more weight and even measuring the problem is difficult
His team hunts the tells deliberately and applies judgment to how the models are fine-tuned and post-trained
V6 Wild is his answer to the flattening. A differently post-trained model, worse at producing complete songs on purpose, exists to widen what a user can reach
8. Where the Economics Land
The hosts framed the outer bound using Spotify, which they said passes roughly 30% of its topline to artists and is still a hundred-billion-dollar company, and asked where AI music settles.
Shulman split the question in two. One half is lowering the cost of the existing thing and handing it to far more people. The other is applying it to markets nobody could previously imagine
His argument against the old sizing of music is that the category itself changed. Investors used to pass because music was small; what Suno is changing is how people entertain themselves, which is a larger thing than recorded-music revenue
He declined to name an equilibrium. There are economic models now, there is enough alignment to build products and work out the economics attached to each one, and the pace of the technology removes the option of settling everything in advance
9. Suno Studio and the Pros
The hosts asked about professional workflows — splitting a generated track into stems, notes and sheet music so a trained musician can edit it, the way an image goes into Photoshop.
The product suite already spans both ends. A few words typed into iMessage at one end; Suno Studio, the company's digital audio workstation, at the other
The surprise was who uses the professional tools. Making those workflows more fun rather than only more powerful pulled ordinary users into them
His analogy was equipment, not skill. "I might buy the signature guitar of my favorite guitarist. I don't think I'm going to be a famous guitarist, but it actually makes me feel really good knowing that I use the same tools that the greats use."
He called the opposite assumption a failure of imagination — that people cannot fall in love with digital tools the way they fall in love with instruments
10. AI Music a Year From Now
One host laid out the pattern in AI video — a Harry Potter and Balenciaga mashup as the breakthrough of one year, a track that turned out to be AI-generated as the moment of the next — and asked what comes next.
Shulman's near-term answer was speed again. Video is getting more convincing and easier to use, and as iterating on a full-length video becomes affordable the constraint moves
He declined to name the creator. The technology is not the story; the story is somebody with a point of view getting the tools to express it, and which point of view lands is not predictable
He did commit to the event. "But there's going to be a big cultural moment involving music and technology."
A host's own prediction was that AI music becomes ubiquitous infrastructure rather than a chart act — in every studio and production, taken off the shelf by particular artists, without a fully AI-made album breaking through
Shulman said that is not too conservative, and gave the reason. "I think that you know a year from now it's actually quite mundane that there's a lot of AI in music." Some of it will be Suno and some will not
His argument against AI chart-toppers is about what a song is. A song is a point of view, refined over and over, and one-shotting a point of view that resonates worldwide is very hard. He drew a line between this and a verifiable, long-running problem such as proving a mathematical theorem, which he said is not what music is
Bonus Insights
On the foldable iPhone launched that morning, Shulman said Apple's motives were not his: "I don't think we're the reason that they're making that they're making that phone, but if you want to suggest that, I'm very happy to take it." A host said he was suggesting exactly that
Asked how a mobile developer should respond to a folding screen, one host's answer was that for many developers it will be a single prompt. Shulman's interest is different interaction moments — using the device without unlocking it, choosing between cameras — and the version of Suno he described as its final form is a headset app that lets a user conduct an orchestra
Shulman asked the hosts whether Suno has ever run billboards or wheat-pasting campaigns. Told it had not, he pitched one: a billboard reading "make a song with a sentence." He then caught the problem — the point is to stop people picking up their phones while driving
The hosts raised self-driving cars as the fix, since a passenger can look at a billboard. The counter on air was that it is a bear case: in a driverless car everybody looks down at a phone instead
Shulman's bottom line is that the fight over training data has turned into a commercial arrangement, and the product that comes out of it is one where a fan's play with an artist's catalog generates revenue that flows back to the artist — with the model itself becoming ordinary infrastructure across music production inside a year.
Products, Companies & Tools Mentioned
Suno (Released V6 and V6 Wild the morning of the interview — faster, more expressive, with the second model deliberately worse at complete songs and more interesting)
Suno Studio (The company's digital audio workstation, which Shulman says draws in ordinary users as well as professionals)
Warner Music Group and BMG (Named by the hosts as counterparties in the partnership announced alongside V6)
Spotify (The hosts' benchmark for how much of the topline reaches artists — about 30%, on a hundred-billion-dollar business)
iMessage (The new distribution surface — Suno is now reachable from inside it, which is how a song with an audience of two gets made)
Apple (The iPhone Duo launched the same morning; Shulman was asked whether Apple built it for music apps and said no)
Waymo (Raised as the reason billboards might work again, and then as the reason they might not)
Books & Resources Mentioned
Steve Jobs Would Have Loved AI, and Been Hated for It (The show's own notes for the day, listing the running order and the headlines behind the news block)
If this was worth your time, send it to someone closer to the industry than you are.
Get the latest market chatter as it happens: