The layer Udio’s Starstruck cannot reach

Udio's forthcoming consumer app, Starstruck, is the clearest signal yet that the licensed AI music market is real. Rights holders are willing to participate. Publishers are negotiating terms. Royalties are flowing back to artists and songwriters every time a fan hits create. The adversarial phase between the music industry and AI platforms has not ended, but a constructive phase is underway, and Starstruck is its most concrete expression to date.
That is progress.
It is also, as it turns out, a precise map of where the problem remains unsolved.

The admission that matters

During Starstruck's launch webinar, someone asked whether Udio could identify when a generated track was structurally or stylistically influenced by copyrighted works in the model's training data, beyond the song the user had explicitly selected. The answer was essentially no. Model-level traceability, it was acknowledged, is fundamentally difficult given how current architectures work.

That is the most important point made at the webinar.

Not because it reflects poorly on Udio, but because it precisely defines the boundary of what product-layer licensing can achieve. It is an honest account of where the technology stands. Starstruck can attribute and monetize the catalog a user explicitly invokes. It cannot reach the layer below: the training data that shapes every output in ways no user ever consciously selected.

The question the industry has moved past is whether to pay rights holders. The question that remains unsolved is how to do it accurately, automatically, and at every layer of influence, not just in the catalog the user picked.

CopyrightChains and the layer beneath

CopyrightChains is built to solve exactly the problem Udio's legal team said could not be solved at the product layer. It operates not at the surface, where users make selections, but at the composition and recording identity layer, where works are registered, ownership is encoded directly into the asset, and settlement executes automatically upon a use event. No bilateral deal to renegotiate. No administrator reconciling reports at month-end. The logic runs at the moment of generation.

• The infrastructure exists.
• It is not a proposal or a roadmap item.
• The work has been done.

What it is not, and what it has never been, is something the music industry has come looking for.

The marketing problem no one wants to talk about

Building rights infrastructure for an industry that did not ask for it creates an unusual problem. The music industry is not known for proactively adopting technical solutions to structural problems it has not yet felt acutely enough to act on. It moves when the pain becomes undeniable, usually toward whoever shows up with the loudest argument at the right moment.

That means the path to adoption runs through marketing, sustained, expensive, industry-facing communication that keeps CopyrightChains visible in the rooms where these decisions are ultimately made.
NIM chose not to take that path, as there is no way to recoup marketing spend from the music industry.
There is no licensing fee that covers the cost of convincing an industry to care about its own infrastructure problems.
The investment is a condition of entry, not a recoverable line item.

This was a reason to stop and to be clear-eyed about what the work actually is.

Starstruck's launch webinar put the problem on record, in plain language, with Udio's own counsel saying it out loud. Every platform that follows, including Spotify's rumored super-premium AI-remixing tier, will eventually face the same admission. The patchwork of bilateral deals that works for ten platform relationships will not work for a hundred.
The administrative overhead compounds.
The independent artist with no label representation remains excluded.
The model-layer attribution problem remains structurally unaddressed.

The solution is there

The question for CopyrightChains is not whether the market exists or whether the infrastructure works.
Both are settled.
The question is when the industry decides to ask for what already exists, and whether CopyrightChains has stayed loud enough, long enough, to be the answer it finds when it does.

Udio did the industry a service by putting the gap on the record.
The infrastructure that fills it is not waiting to be built.
It is waiting to be called.