Hi there!

I don’t identify as a “journalist” these days. As a consultant, I now sit squarely on the business side of music, working directly with industry clients. The core work I do is better described as market research, competitive analysis, and strategic advisory in music tech.

That said, I recently got my hands on a scoop that hasn’t been reported anywhere else in the press, and that feels quite critical for Water & Music readers to know.

This is about Udio’s upcoming licensed AI app — which has its own name, Starstruck. The shape that Starstruck is taking is so deeply tied to topics I’ve written about recently, including AI remixing and superfandom more broadly, that I felt compelled to publish a special issue about it.

We’re also kicking off a new survey on AI and artist services, and hosting two exciting events in NYC as part of New York Music Month. More details below.

We’ve had 100 new subscribers join us since the last issue. If you’re one of these folks, welcome! Feel free to reply directly to this email to introduce yourself — I read everything that comes through.

Thank you so much for your support!

Best,

Cherie Hu
Founder, Water & Music

Announcements

New survey: Artist services & AI

I’d love your help in filling out Water & Music’s latest survey on AI music trends, especially if you are a professional artist or music creator.

The AI music landscape has hit an inflection point, with a new wave of licensing deals moving real money — and shaping the terms of training, attribution, and royalty flows — faster than most artists can track. What does this moment look like for individual recording artists, songwriters, and producers, especially those whose labels, publishers, or distributors might be involved in these deals?

We’re running a 100% anonymous survey to find out. This survey is open to recording artists, songwriters, producers, engineers, and the wider creative community — whether signed, self-releasing, freelance, or somewhere in between.

As a thank-you, every respondent will receive a special copy of a keynote presentation we gave in December 2025 on how AI music attribution and compensation works ($2,000 value).

This survey closes on Wednesday, June 3 at 11:59pm ET. Please contribute your perspective and/or spread the word to your artist, songwriter, and producer peers!

We’re back for New York Music Month

Thrilled to share that we are once again collaborating with the New York City Mayor’s Office of Media and Entertainment to put on two events for the local music-tech community, as part of New York Music Month.

We brought together over 300 people for our NYMM events last year, and are excited to connect with faces old and new for Year 2!

June 9: NY MusicTech Meetup and Demo Night @ Public Records — We’re back for an evening of curated startup demos and focused networking at one of NYC's most renowned music venues and event spaces. (If you’re interested in showcasing your startup, please apply to present here by 11:59PM EDT on Wednesday, May 27.)

June 17: Music Tech Investment Panel @ Perkins Coie — Our second annual panel on music tech capital featuring the investors on the front lines of today's cutting-edge deals, including Robert Frech (Managing Director, The Raine Group), David Hua (VP, Creator Partners), and Raj Gopal (Partner, Vesper).

A scoop on Udio’s upcoming app, Starstruck

You can also read this article via our newsletter archives or on LinkedIn.

Three months ago, I argued that the AI music market was on a collision course, with product, licensing, and rights management all converging toward the same narrow set of bets.

On the licensing front in particular, major labels and indie bodies like Merlin, Believe, and Kobalt have been working to push AI music apps in the same direction — away from producing generic AI songs en masse, and toward remixing existing catalog and IP, inside closed environments where rights holders can track usage and collect royalties accordingly.

Turns out, Udio is now building exactly that, in an upcoming product called Starstruck.

Udio is currently one of the leading AI music apps when it comes to licensing agreements, inking deals with Universal, Warner, Merlin, Kobalt, and Believe over the last year. Many people in the industry have been on the edge of their seat, awaiting details about how these licenses will actually translate into concrete product changes for Udio and AI music at large.

Thanks to one of my peers, I got access to a recording of a private webinar that Udio co-hosted with Kobalt on April 30, 2026, in which Udio CEO Andrew Sanchez gave a live demo of Starstruck and answered a few questions from Kobalt members. Kobalt’s CEO Laurent Hubert and Chief Digital Officer Bob Bruderman briefly chiming in as well.

This webinar has not been reported by the press. Below is a summary of what was shared, the wider industry context around it, and my own take on whether Starstruck will succeed.

What is Starstruck?

tl;dr Starstruck is Udio’s forthcoming consumer AI app that lets fans remix, cover, and reimagine songs by their favorite artists, with royalties flowing back to rights holders every time a fan creates. It is set to launch later this year as a mobile-first app, aimed at everyday fans rather than professional artists or producers.

The app will offer four modes for fans to interact with music from opted-in artists and songwriters. Each of these will present a wonderfully different puzzle for the lawyers among you:

• Cover generates a new artist's version of an existing song, using the original lyrics and composition (for instance, hypothetically, a Charli XCX-style cover of a Taylor Swift song).

• Reimagine keeps the lyrics but rewrites the musical composition entirely, as if a different artist had written the song.

• Remix applies genre or style shifts to existing recordings (e.g. hypothetically, a reggae version of a country track).

• Create lets users write their own lyrics and pair them with a selected artist's voice, subject to guardrails around topic, language, and style.

Whichever mode users ultimately pick, they will always have to select a specific artist and song; generic, unattributed AI music isn't on the menu. Artists and songwriters opt in via their respective label and publisher partnerships, and royalties will flow back to artists, songwriters, and rights holders every time a user hits "create" on these modes.

Bruderman shared on the webinar that “given the overall structure of this deal… the songwriter [will be] paid significantly more than they are in the traditional streaming construct.” This is similar to how Kobalt achieved 50/50 parity on publishing/recording payouts in its deal with Udio competitor ElevenLabs.

Interestingly, the intention with most of these interaction modes is that no new copyrights will be produced — at least on the composition side.

During the webinar Q&A, a Kobalt member asked who owns the recording of music generated on the platform. Sanchez explained that the generated recording would be owned by the rights holder of the participating artist — which could be the artist’s label, publisher, or perhaps the artist themselves via their distributor. In other words, the fan is not being positioned as the owner of the AI-generated recording; they are paying for the right to create and listen inside Udio’s controlled environment.

The focus on covering, remixing, and reimagining existing IP also ties to Starstruck’s constrained approach to attribution. Another Kobalt member asked whether Udio could identify when a generated track was stylistically or structurally influenced by specific copyrighted works in the core model’s training dataset, beyond the song or artist the user selected. Sanchez’s answer was essentially no: he described that kind of model-level traceability as fundamentally difficult, and perhaps impossible, given how current model architectures work.

Instead, Starstruck can only attribute and monetize the catalog that users explicitly invoke. It does not appear to solve the more contentious question of how training data influences outputs in ways the user never directly selected. So while Starstruck may represent a licensing breakthrough at the product layer, it is not yet a full rights-management breakthrough at the model layer.

Starstruck’s pricing model will be subscription-based across two tiers, Standard and Pro, each capped by a different number of monthly creations. This mirrors the current credit-based pricing plans for both Suno and the legacy Udio app. Sanchez also floated a more campaign-oriented version of the tiering model — where higher-tier users could, for instance, potentially get earlier access to new features, or even earlier access to remix new releases from participating artists.

Importantly, every generation on Starstruck will also live exclusively inside the app’s “walled garden” — i.e. songs cannot be exported or distributed elsewhere.

Sanchez described three layers of control that Udio is building to enforce these protections: stream encryption to make ripping harder, inaudible watermarking on every output, and fingerprinting so distributors and streaming services can check uploaded tracks against Starstruck creations and block any matches. That last piece matters, as it presents a massive coordination challenge in addition to a technical challenge. Udio’s walled garden will only be as strong as the surrounding music distribution ecosystem’s willingness to recognize and enforce these fingerprints.

Will Starstruck work?

One of the core questions I asked in my collision course article was: do fans actually want to use an AI remixing app like this? While the jury is still out, new consumer data suggests there is at least an engaged niche market in play.

In April 2026, MIDiA Research published a report on the state of AI music that gives Udio some reasons to be hopeful. According to MIDiA, only 6% of consumers in the US, UK, Canada, and Australia use generative AI music platforms on a monthly basis — but that 6% over-indexes hard on virtually every signal that defines a "superfan." Gen AI users pay for music subscriptions at nearly double the rate of the average consumer (68% vs. 36%), go to gigs at more than double the rate (45% vs. 19%), and buy merch at more than triple the rate (41% vs. 13%). Among AI voice tool users specifically, 42% frequently engage with fan-made content, compared to 15% of consumers overall.

The MIDiA data suggests that Starstruck isn't fishing in a vague, undefined pond. While the AI music market is still niche, the people inclined to use these tools are the same people who already orient their identities around music — the ones most likely to want a remixed version of their favorite song, and most likely to pay for it. This is a meaningful counterpoint to the unilateral "AI will kill the music business" narrative that has pervaded industry discourse for the last few years.

Of course, overcoming the deeper structural obstacles to meet these fans where they are will be much easier said than done.

The superfan subscription trap

My previous newsletter issue, which made the rounds in the industry, examined the mounting evidence that superfan subscriptions are dying out.

My argument was not against the superfan business case — which is strong as ever — but against the notion of recurring subscriptions as the default container for how artists and superfans should relate online. tl;dr most artist and fan activity is “spiky” around tentpole events like album releases, tours, anniversaries, and viral moments, and does not operate on a neat monthly schedule. Fans still want closer relationships with artists, but do not want to pay every month for access. The right answer will be giving artists tools to reach, recognize, and monetize fans when the moment actually matters.

At first glance, Udio's Starstruck looks like a clever sidestep around the failure modes I identified in that piece. For one, Starstruck does not ask artists to keep showing up on a content treadmill. The creative labor shifts entirely to the fan, who can generate covers and remixes on their own schedule, around their own moments. The artist’s catalog is simply the input.

The app also ostensibly formalizes existing behavior, since fans already remix and cover their favorites all over platforms like TikTok and YouTube. This feels closer to infrastructure that supports existing fan impulses, rather than a system that manufactures new obligations for artists and their teams.

And yet, to sustain itself financially, Starstruck will have to swim upstream against the relentless “spikiness” of fan behavior as well. Starstruck requires continuous active engagement: pick an artist, pick a song, hit create, listen, track your credit usage for next time. The recurring monthly bill that fans will get from a Starstruck subscription still imposes a flat cadence on irregular behavior, which is the same failure mode that drove the likes of Vault to drop paid subscriptions. Will fans really keep generating enough covers and remixes month after month to justify the charge?

On walled gardens and social platforms

The walled garden debate is the second structural obstacle, and arguably the more important one. While closing the ecosystem protects artist and songwriter economics, it also strips out the social shareability that drives meaningful fan behavior in the first place.

This question — whether AI music should stay within a walled garden or live freely on traditional streaming and social platforms — has been at the heart of the ongoing AI music lawsuits. In February 2026, Suno's Head of Music Paul Sinclair published a LinkedIn post arguing in favor of an open ecosystem. A few weeks later, UMG’s Chief Digital Officer Michael Nash declared on Billboard’s “On the Record” podcast that Suno’s disinterest in building a walled garden around AI outputs was a “hat-hanger” in their ongoing litigation.

Udio’s CEO Sanchez has been firmly on the other side. In an interview with Music Business Worldwide this week, he defended Udio's closed approach: "A highly controlled ecosystem is necessary for most rightsholders and artists to be comfortable with this extraordinary and potentially transformative technology."

The most important distinction in the walled garden debate may be between DSPs and social platforms. While rightsholder-approved remix and cover tools like Starstruck can solve the important monetization piece, it is on social and UGC platforms where songs actually breathe new, meaningful life and drive meaningful cultural impact at scale.

My colleague Lenny Skolnik (a.k.a. Yung Spielburg) recently wrote about this issue on LinkedIn, arguing that “SoundCloud and TikTok have resurrected more careers than any A&R department.” Without social sharing, the fan incentive to play with their favorite artist's music and voice may never reach critical mass. Wall off that behavior, Lenny argued, and you kill the thing that creates demand in the first place.

So far, the data backs this up. While Udio is clearly deprioritizing its web app to focus on Starstruck, the existing walled-garden setup on the web version has tanked its engagement: consumption on Udio is now orders of magnitude behind that on Suno.

The top trending tracks on Udio get only around 100 plays a month. Meanwhile, top tracks on Suno, which still allows social sharing, regularly clear 200,000+ streams in the same window — and that's only counting native Suno streams, not off-platform activity on YouTube, Spotify, or elsewhere. As of late February 2026, Suno has 2 million paid subscribers and $300 million in ARR.

To quote artist and technologist Mat Dryhurst: “pop music is a promise that you aren’t listening alone.” Fan remixing is not only about making a thing — it is about sending it, posting it, arguing over it, laughing at it, and watching other people react. Walled gardens can absolutely help labels, artists, and rights holders control what gets created and attributed. But if that control causes fans to abandon the platform, then it raises the question: what is the point? (inb4 “the point is to kill these platforms,” which is a debate for another time.)

Lastly, let us consider that Spotify has been openly teasing a similar AI remixing feature for its rumored "super-premium" tier, in collaboration with many of the same licensing partners as Udio.

If it ever ships, Spotify will be entering the same walled-garden game, but from a position of overwhelming scale — namely, over 290 million paying subscribers globally. Even a sliver of that user base would give Spotify a garden much larger and livelier than anything Udio or even Suno could realistically build from scratch. The social layer of sharing, posting, and reacting could plausibly happen inside Spotify even within a closed ecosystem, simply because the network effects are already in place.

All in all, based on the Kobalt webinar, Starstruck is a strong answer to a question of how to license AI remixing without diluting catalog economics. It is a much weaker answer to the question that actually determines whether any of this works: how do we get fans to do this at scale — not to mention, in a closed system that forbids the very behaviors that make fan creation meaningful?

The licensing puzzle largely has a solution at this point, but the consumer adoption puzzle does not. Udio's bet seems to be that solving the first puzzle buys them enough time to solve the second. We will find out soon enough, and perhaps feel starstruck along the way. ★