The MLC has good reasons to refuse Suno's data request. It should hand over a version anyway

Suno wants the Mechanical Licensing Collective's streaming data to fight the labels. The MLC doesn't want to give it. The MLC's objections are better than they look, and we still think a court should make it hand over a narrower, protected version.

The news hook

On 15 Oct 2026 Suno filed a motion to compel the MLC in federal court in Nashville. According to Digital Music News, Suno's April subpoena asked for monthly performance data, label by label, from 11 streaming services from 2021 onward; an August version dropped Meta, Snapchat and TikTok. Suno says the data shows what happened to the market before and after AI music arrived. A day earlier, UMG and Sony had asked a Manhattan court to make DistroKid hand over AI-detection counts, streams and royalties, per their proposed order. We covered both in The AI music war becomes a data war. This column takes one side of one fight: should the MLC have to answer?

The best case for the MLC saying no

Put it as the MLC would. It is a nonprofit set up under the Music Modernization Act to administer blanket mechanical licences and pay songwriters and publishers. It isn't a party to the Suno case. It says it doesn't collect data from labels or distributors and has no mandatory label fields, so a "label by label" breakdown asks for data in a shape it doesn't keep, per Digital Music News. The numbers come from streaming services, which objected to disclosure. And Suno has refused to limit the request to the plaintiff labels, so it reaches every rights holder's data in the market.

The law takes that seriously. Rule 45 tells a party issuing a subpoena to avoid "imposing undue burden or expense" on the person served, and lets a court quash or modify one that would disclose "a trade secret or other confidential research, development, or commercial information." A royalty body whose members trust it with their data has a duty not to hand that data to any litigant who asks. If every AI lawsuit can reach into the MLC's records, the MLC becomes a discovery warehouse for the whole industry.

That is a serious argument. On the scope of what Suno asked for, it mostly wins.

Where it breaks: market harm needs a market

The labels' theory, which the US Copyright Office's 2025 training report supports, is that AI output can dilute markets and royalty pools. That is a claim about the whole market. The labels are already trying to measure it with DistroKid's numbers, which come from one distributor and a third-party prediction model whose false positives the labels themselves ask about in their proposed order. If the court lets one side bring market evidence and blocks the other from testing it against wider data, the fair-use ruling rests on half a picture.

The MLC's best objections are about how much and in what form, not about whether. Rule 45 already has the tools: a court can modify a subpoena instead of quashing it, and the requesting party can be made to pay. Digital Music News reports Suno's lawyer said the MLC had agreed to produce data for a cost estimate of about $5,000 before reversing course. That suggests the burden of some version is manageable.

What we think

Narrow it, protect it, produce it. A sensible order would cover only what the MLC actually holds, at whatever level of aggregation it keeps, rather than invented label fields. It would put the data under an attorneys'-eyes-only protective order, with Suno paying the cost. And it would let the streaming services object to specific fields, not the whole request.

There is a second reason we lean this way, and it's about artists, not Suno. Whatever market numbers win this case will shape the AI licences and settlements that follow, including how royalty pools get split between human and AI tracks. Songwriters, whose money the MLC collects, have more to gain from those numbers being tested by both sides than from one side's figures going unchallenged. That's our reading of the incentives, not something the MLC or Suno has said.

What would change our mind: if the MLC shows that even an aggregated, protected production would expose individual songwriters' earnings, or would take far more than Suno's figure suggests, the burden argument wins. Until then, "we're not a party" is a reason to set limits, not a reason to keep the market data out of a case about market harm.