What Suno And Udio’s AI Licensing Deals With Music Majors Could Mean For Creators Rights

In the space of a year, the major record labels have shifted from legal crusaders to would-be business partners. When Universal Music Group, Warner Music Group and Sony Music filed copyright-infringement suits against AI up-starts Suno and Udio last summer, the industry assumed a bruising court fight was inevitable. Nine months later, the same companies are at the table hammering out AI music licensing deals that would let the startups keep training on label catalogues, provided the labels, and eventually their artists, get paid.

These talks are not just about settling a lawsuit. They are about setting the rules, or perhaps abandoning them, for how copyrighted music is used in training AI, how future licensing structures might look, and who gets to be in the room when those decisions are made. For many artists, this is déjà vu, and not the good kind.

From “Exploiters” to Partners

The pivot is striking. Just months ago, the majors accused Suno and Udio of having trained their models on copyrighted sound recordings “at an almost unimaginable scale,” offering prompts that could generate near-identical copies of existing songs. The Recording Industry Association of America alleged “mass infringement” and sought sweeping legal remedies.

Now, according to The Wall Street Journal, the labels are seeking licensing fees, compensation for past use, and minority equity stakes in both companies. In return, the startups would receive the right to continue using major-label catalogs to train their models with new controls and attribution systems in place.

Among the key conditions: a fingerprinting and attribution layer modeled after YouTube’s Content ID system. This technology, if feasible, would enable Suno and Udio to trace how and when songs influence AI outputs, allowing rights holders to track usage and collect revenue accordingly. The labels also want veto power over future AI music tools, including voice-cloning features and remix suites, a position one executive compared to the “controls labels already exercise in sync deals.”

But even as these negotiations accelerate, one truth remains unaddressed: labels don’t control everything.

As Gadi Oron, CEO of CISAC, the global body representing authors’ societies, points out: “Negotiating solely with the majors will not provide the full set of rights required by AI companies. Labels can only license the rights they control, which are the rights in the master recordings.”

Underlying compositions and lyrics, the lifeblood of songs, are typically managed by collective management organizations, and these rights have not been part of the current licensing discussions. Oron warns: “To use music lawfully, especially for training or generating new content, AI companies also need to obtain rights to the underlying compositions and lyrics. These rights are typically managed by CMOs on behalf of songwriters and publishers. Without separate agreements with CMOs for the compositions and lyrics, AI companies would infringe the rights of music creators, which is the current situation in the market.”

Loredana Cacciotti, founder of Future Play Music, a company specialised in digital licensing strategies for labels and distributors, echoes this concern, and places it in historical context: “History has taught us that the major labels often act based on short-term financial gain rather than long-term protections for artists and the industry as a whole. And that concerns me. We may once again find ourselves locked into licensing frameworks that fail to account for the deeper implications, both in terms of creative control and economic fairness for the independent community when a unified voice should be front and center when responding to disruptive innovation.”

She adds: “Yet, given the fragmented nature of our industry, there’s a real risk that we will once again be passively swept into a new era, one shaped by decisions in which much of the music community has had little to no voice.”

Who Speaks for the Creators?

To artist advocates and collectives such as the Musicians’ Union and the Ivors Academy, the exclusion of songwriters and performers is both predictable and dangerous. Many are sounding the alarm that history is repeating itself, only this time, it’s not just the distribution of music that’s being reshaped, but the right to exist as a creative identity.

Phil Kear, Assistant General Secretary of the UK’s Musicians’ Union, asked pointedly: “Will the consent of the music creators be sought? What share of the licensing revenue will they receive, if any?”

Meanwhile, Ivors Academy Chair Tom Gray warned that these agreements “appear to not offer creators an ‘opt-in’, an ‘opt-out’ or any control, whatsoever, of their work within AI.” Gray added: “The same companies who have stated they wish to ‘make it fair’ seem instead to be ‘on the make.’”

And Oron’s concern extends beyond who is in the room, it is about how the economic pie is being divided: “In the early days of the digital music market, some services operated without the required licenses, but at a later stage, negotiated deals with the major record labels in exchange for the lion’s share of income… With AI, the connection to the underlying musical works is even more essential, and should entitle creators to a larger share of the income.That makes it all the more important for composers and lyricists to be included in licensing negotiations from the start, with a clear stake in the outcomes. Failing to recognize this reality risks repeating past mistakes and marginalizing the very creators whose work underpins these technologies.”

Cacciotti agrees. She warns, “These developments carry a distinct sense of déjà vu. We’ve been here before—most notably during the rise of streaming, when rights holders had to decide between fighting the tide or shaping it. But this time, the stakes are arguably even higher.”

This isn’t theoretical. GEMA, the German authors’ society and a CISAC member, has already filed suit against Suno, citing exactly this disconnect.

Suno and Udio: Why These Startups Matter

These are not trivial or speculative technologies. Suno and Udio have evolved from experimental demos to near-production-level toolsets. Suno’s June 2025 update allows users to upload full tracks, manipulate them with “weirdness” and “reference” sliders, and export 12 multitrack stems to a digital audio workstation. Udio’s most recent build added “intro/verse/drop” sectioning, faster generation times, and support for hybrid genre compositions.

But these features are only possible because the underlying models were trained on enormous datasets, including, by many indications, commercially released music without permission.

And while Suno claims its models don’t memorize or reproduce music, evidence from lawsuits shows that, when prompted, they’ve generated lyrics and melodies “identical or nearly identical” to protected songs.

Timing Is Everything: Why the Music Industry Is Moving Now

The urgency behind licensing negotiations between the major music companies and AI startups like Suno and Udio isn’t coincidental. Several structural pressures are converging to make this a uniquely combustible moment, one in which both the music industry and AI companies may see a narrow window to shape the future before external forces lock it in for them.

First, regulatory uncertainty looms large. The recent and abrupt firing of U.S. Copyright Office Director Shira Perlmutter, who had pushed back against broad “fair use” exemptions for AI training, sent a chill through the creative industries. Her removal has raised fears that a new Trump-appointed director could reshape federal copyright policy in favor of AI developers, weakening enforcement mechanisms for rightsholders and potentially legitimizing unlicensed dataset training.

Then there’s investor pressure. Suno’s $125 million raise in 2024, which valued the company at $500 million, reflects both excitement and risk. Venture capital firms increasingly want “clean” AI pipelines, ones backed by licensed data and clear rights frameworks. That means unresolved litigation is now a liability. For companies like Suno and Udio looking to scale or exit, licensing deals are no longer optional; they are the precondition for long-term capital access.

Finally, international policy is catching up. The European Union’s AI Act and the UK’s stalled exceptions for text and data mining both signal that the days of unregulated scraping in Western markets may be numbered. Compliance obligations, audit trails, and provenance disclosure could soon become mandatory. For Suno and Udio, this is likely the last best moment to secure cooperative licensing arrangements before governments impose restrictions that could limit how and what their models are allowed to ingest.

What the Majors Want from AI Music Licensing Deals

What’s emerging from these negotiations is a licensing framework that strongly resembles the major labels’ approach during previous tech disruptions, most notably their transition from suing Napster to licensing Spotify.

At the top of their list is the demand for fingerprinting at the model layer. Labels want systems that can not only detect direct sample reuse but also flag stylistic derivations within generative model outputs. The ambition is to move beyond surface-level detection and toward embedded attribution systems, although whether that’s technically feasible with current diffusion models remains an open question.

As Mike Pelczynski, Head of Licensing and Industry Relations at Sureel AI, which builds instant attribution systems for generative content explains: “Attribution systems are fundamentally more powerful than traditional content ID in the age of AI because the sheer scale and speed of new content creation make it impossible to track every instance of reuse manually. Only neutral attribution frameworks can identify relevant works, respect opt-outs, and give rightsholders real-time visibility and control.”

Next, the majors are pushing for commercial veto rights over product features. This would mean that any future tools released by Suno or Udio, from voice-cloning plugins to remix engines, would require prior approval. It’s a mechanism similar to the one labels have long enforced in sync and advertising licenses.

Financially, the proposed package includes cash settlements for past use, usage-based royalties going forward, and minority equity stakes in both AI startups. This echoes the labels’ early equity positions in Spotify, which later became highly lucrative, but also controversial, as artists had little visibility or participation in those deals.

One reason the majors are negotiating from a position of strength: many have likely registered copyrights with the U.S. Copyright Office that they believe Suno and Udio used for training. If proven, that could expose the startups to statutory damages, potentially amounting to hundreds of millions in liability.

“The crucial point in the Suno and Udio licensing discussions,” said Liz Cimarelli, Head of Business Development at Cosynd, a platform that simplifies copyright registration and ownership tracking for creators, “is that the major labels have likely registered copyrights with the Copyright Office that they believe these companies used for training their AI models. Given the potential statutory damages of $150,000 per willful infringement, this could serve as a significant negotiating advantage for the majors.”

But she also warned: “The risk for the wider industry is that the major labels might agree to terms that set a low standard for everyone else. Without new legislation, policy changes, or infrastructure, this could diminish the economic opportunities that AI should offer. Generative AI has already been predicted to cost music creators $22B in income over the next five years. How low can we go?”.

Lastly, there’s a nod toward creator control: artist opt-outs for certain use cases such as vocal cloning. But crucially, there’s no sign yet of a rights framework that would allow artists to license (or deny) their work directly, nor clarity on how royalties will be tracked or distributed at the artist level. For many creators, this feels less like consent and more like default inclusion with an escape clause.

And as Oron makes clear, none of this addresses songwriters’ rights: “AI companies must seek permission from all relevant rights holders, not just the labels. Without compositions and lyrics written by humans, there’s nothing for SUNO or Udio to offer.”

For attribution advocates, the stakes are higher than just tracking AI training inputs, it’s about future leverage. “Flat licensing without attribution is blind licensing,” said Dr. Tamay Aykut, founder and CEO of Sureel AI. “Artists (and their labels) would lose control and could end up competing against their own AI derivatives. Labels can’t price what they can’t measure, and AI can’t avoid what it can’t track. Attribution is the difference between knowing and guessing and if they are indeed pursuing licenses, then neutral attribution can only strengthen their hand with the AI companies who want to do the right thing.”

Aileen Crowley, co-president of Sureel AI, added: “As an industry, we must come together to demand that licensing only happen when an independent attribution system is in place. Any future licensing deals must guarantee that rights holders can clearly and effectively exclude their works, and only attribution technology can deliver this level of control and transparency.”

That sentiment was echoed by Benji Rogers, also co-president at Sureel AI: “Opt-in and opt-out rights must be non-negotiable, and only neutral attribution can provide this level of transparency and protection.”

The Road Ahead: Possible Futures

These licensing negotiations could define not only the outcome of current litigation but the licensing infrastructure for AI music globally. Three distinct scenarios are emerging:

In one scenario, licensed acceleration, the majors strike a deal this summer. Suno and Udio integrate attribution and payment systems, and AI-generated remix tools launch inside premium tiers. Labels win a new revenue line, and high-profile artists who embrace the tech gain visibility. But those who opt out, or were never consulted, get left behind.

In a second, stalemate, negotiations collapse and lawsuits drag on. If Trump-era regulators tilt toward AI-friendly fair use policies, case law may erode the legal basis for any future licensing obligations.

In the third, patchwork, one or two majors settle, others hold out, and AI companies develop regionalized tools trained on different catalogs. The result is a fragmented landscape that mirrors the dysfunctional world of sync licensing.

Five Unanswered Questions

Amid all the strategy, five foundational questions remain unanswered:

1. Can fingerprinting systems actually trace micro-stylings within a diffusion model, or will they only catch obvious replicas?
2. Will voice and likeness rights be negotiated separately from master and publishing rights, or will they be bundled in?
3. How will royalty statements for model training be structured? Per API call? Per generated second? Per dataset token?
4. Could this moment spark the creation of a compulsory AI training license, akin to mechanicals for reproduction rights?
5. And if labels take equity stakes in Suno and Udio, will creators see any share of that upside?

Consent is Not a Checkbox

This is not just a music industry story. It’s a proxy for how every creative sector, from writing to voice acting to film, navigates the shift from human artistry to machine synthesis. For the majors, this may feel like the inevitable next step in monetizing technological disruption.

But for the artists, songwriters, and composers whose music trained the machines, this is about power, authorship, and cultural survival. As with the rise of streaming, the question isn’t whether the business will change, it’s whether the people who make the music will be allowed to shape that change.

As AI music licensing negotiations between AI startups and major rights holders quietly unfold, songwriters, composers, and performers are once again fighting for a seat at a table where their work is the main asset, but their voices remain peripheral.

Universal Music Group and Sony Music Entertainment were contacted for comment. As of publication, neither had responded.