The time it takes AI to create a song keeps shrinking. So should the value of a company that owns already-made music fall? It doesn't necessarily seem to be heading only that way. The Longleaf Partners Global Fund newly added Universal Music Group (UMG) in the third quarter of 2026, with a weight of 4.82% at the end of September.[1] In an era when the supply of songs overflows, what the value investor bought was not technology that makes songs faster, but a business holding music that people keep coming back to listen to.
A Familiar Business Longleaf Met Again
Southeastern Asset Management, which manages the Longleaf funds, is an investment manager that has practiced long-term, concentrated value investing since 1975. Its principle is to look for both management running a good business and a price below that business's value.[7][8] The global fund held 21 positions this time, with the largest at 8.0%, so UMG's 4.82% is hard to dismiss as a small token purchase expressing mere interest.[1]
The management team described UMG as a business with valuable music assets and durable earnings power. They also noted that it is a company they had invested in several times since its days inside Vivendi.[1] Rather than a newly discovered stock pulled in by talk of new technology, this is closer to meeting a familiar business again at a price worth buying.
That said, the management team did not cite AI directly as the reason for buying UMG. The question of whether AI will change the value of music is a perspective added in this article. It is better to keep the fact of the purchase and the narrative layered on top of it separate here.
What Does a Music Company Own in a Single Song?
Calling UMG simply a music rights company skips about half the business. The recorded music business concerns recordings—the sounds a particular artist recorded. Music publishing manages the rights to the melody and lyrics that make up the song. Even for the same song, the rights to the recording and to the composition/lyrics are distinct, and which rights UMG holds or manages depends on its contracts with artists and songwriters.
This distinction also matters for the flow of money. Subscription fees paid by consumers to platforms do not become UMG's profit as-is. Out of the fees paid by platforms, UMG must bear the share that goes back to artists, songwriters, and others. UMG's artist costs in the first half of 2026 were 49.2% of revenue—higher than 47.5% in the same period a year earlier, though excluding the Downtown consolidation, they were 47.3%.[3] If you look only at revenue growth and assume a music company keeps nearly all royalties, you misread the earnings structure.
UMG discovers new artists, produces and markets music, while also earning income from its accumulated back catalog. If already-made music is streamed across generations, the burden of having to make every new release a hit can be lightened. But new signings, artist renewals, and catalog acquisitions all require cash. Owning old assets does not mean the investment ends there. In fact, UMG spent €280 million in cash on catalog acquisitions in 2025 alone, and €266 million in 2024.[2] It is not a business where old assets earn money on their own, but one that replants a portion of earned income back into catalogs and new artists. The impression that it is a business using almost no capital skips over half the picture.
Separately from catalog acquisitions, there are also advances—payments made to artists up front and recovered from royalties going forward. UMG records as assets those advances it judges sufficiently recoverable, and recognizes them as costs when the artists later earn royalties. Portions judged difficult to recover are expensed in the period the loss becomes evident.[13] This means the timing of cash outflow and the timing of the expense hitting the income statement can differ. In 2025, with major artist renewals and contract extensions overlapping, net cash outlay for royalty advances net of recoveries rose to €402 million, up from €186 million the previous year.[2] Low capital expenditure and spending little cash to retain popular artists are different stories.
What AI Increases Is Supply; What Subscribers Choose Is Demand
According to the IFPI, the global recorded music market in 2025 was $31.7 billion, up 6.4% from the previous year. Paid subscription streaming accounted for 52.4% of market revenue, with 837 million paid subscription accounts.[6] What these numbers show is that the channel through which people pay for music is still large and growing.
But if a million new songs are added, no single subscriber's listening time becomes a million times longer. Growing supply and more money coming in are different matters. As music becomes abundant, discovery and distribution become more important, and the bargaining power of catalogs that platforms feel they must carry may remain. Conversely, if consumer listening shifts to new AI music, revenue for existing music could decline.
Here it is worth taking another look at how royalties are divided. Even if many tracks are registered, they do not all receive equal shares. For example, the artist-centric royalty model that UMG and Deezer announced in 2023 to launch in France was designed to double the royalty weighting for artists with at least 1,000 streams per month and at least 500 unique listeners, and to add extra weighting to music listeners actively chose. It also included excluding non-musical noise content from revenue distribution and filtering manipulated plays.[15] Under such a structure, what determines existing music's share is not the number of tracks AI uploads, but the plays that actually qualify for royalties and their weightings. However, the terms of this announcement cannot be applied as-is to current contracts on all platforms.
One response UMG chose is licensing agreements. In October 2025, settling its lawsuit with Udio, it announced agreements covering recorded music and publishing along with plans for a new AI music platform.[4] In September 2026, it announced a multi-year licensing and product development agreement with ElevenLabs as well.[5]
It is too early to say, looking at these agreements, that AI has immediately become a huge new revenue source for UMG. Still, noteworthy is that instead of debating whether to block AI entirely or allow it entirely, a business is emerging that negotiates which music is used, on what terms, and how compensation is received. UMG's fourth-quarter 2025 results also reflected compensation from strategic licensing agreements with AI music platforms.[2] Between the announcement of agreements and recurring cash revenue taking root, there is still a distance to verify. Demand-side numbers keep moving too. UMG's first-quarter 2026 subscription revenue grew 12.5% excluding currency effects, of which the wholesale price increase under the Streaming 2.0 agreements contributed about 3 percentage points, while a market share decline from a release gap offset about 2 percentage points.[11] Excluding Downtown, subscription revenue growth was 7.9%.[11] The price increases are, as the phrase goes, still in their 'initial benefits' stage, and the agreements with global platforms such as Spotify and YouTube include provisions preventing AI-generated output from diluting artist and songwriter royalties.[2] In other words, even if AI changes the listening map, the subscription basket price and contract terms are the first line of defense protecting the catalog's share, and whether that defense actually works can only be known by separating the wholesale price contribution from share fluctuations quarter by quarter.
Indeed, in the following quarter the two forces worked together again. Second-quarter 2026 recorded music subscription revenue grew 16.6% excluding currency effects, but 6.7% excluding Downtown. The company explained that the wholesale price increase contributed 3.5 percentage points to growth, while the market share pressure carried over from the first quarter offset 1.5 percentage points.[3] The subscription fees streaming services charge consumers and the content payments they pay UMG are linked, but they are not the same price. Consumer price increases should not be read as immediately turning into the same-magnitude profit growth for UMG. The share distributed to artists, contract terms, and what music is actually streamed all move together.[3][13]
Conversely, the mere fact that platforms need UMG's music does not mean bargaining power lies on only one side. The annual report stated that the top 50 music services accounted for 97% of recorded music digital revenue in 2025, and listed as a risk factor that contracts with streaming services are generally relatively short-term.[13] However many listeners worldwide, the counterparties UMG negotiates with are the platforms. As the same report noted, playlists and recommendation algorithms shape music discovery, so catalog power and dependence on distribution channels must be viewed together.[13]
More Curious Than the Growth Rate Is How Cash Is Left Over
UMG's 2025 revenue was €12,507 million and adjusted EBITDA €2,810 million. The adjusted EBITDA margin was 22.5%, and net debt at year-end divided by EBITDA was 0.9x.[2] Having an old catalog does not mean growth has stopped, but neither is it a business where costs or financial burdens can be ignored.
In the first half of 2026, the Downtown consolidation made growth hard to read. Revenue growth excluding currency effects was 10.8%, but 5.7% excluding Downtown. The adjusted EBITDA margin was 21.1%, down 1.6 percentage points from a year earlier, with the company citing the Downtown consolidation as the main cause.[3] This is a period when the enlarged scale from acquisition and the growth of the existing business must be distinguished.
Still, the margin decline cannot be blamed entirely on Downtown. Even excluding Downtown, first-half adjusted EBITDA growth was only 1.6% excluding currency effects, with the margin down 0.8 percentage points. The company cited the growing share of artist-and-label services revenue within recorded music, increased corporate costs centered on professional services fees, and losses in the merchandise business.[3] Revenue from owning music rights and service revenue from helping other artists with distribution and management are not the same-margin revenue. When the path of growth changes, revenue growth and profit growth can diverge even after stripping out acquisition effects.
Cash flow needs to be viewed even more carefully. From 2026, UMG changed its definition of free cash flow (FCF) and restated the comparison periods on the same basis. On that basis, first-half FCF was €24 million, down from €163 million in the same period a year earlier. The company cited the decline in operating profit, expanded working capital cash outlays, and increased interest and capital expenditures as causes.[3]
So the small cash flow figure cannot simply be blamed on the definition change. It declined even when compared on the same basis. The old-definition FCF of €702 million in the 2025 annual release should also not be directly compared with the first-half 2026 figure.[2][3] Evaluating the durability of a music catalog and calculating the cash left for shareholders today must go together.
The scope of the new FCF also matters. The definition the company changed in the second quarter of 2026 shows cash before deducting strategic investment spending such as catalog acquisitions, equity investments, and corporate acquisitions.[3] So even if this number grows, it does not mean the money needed to secure music assets has disappeared. The direction of advances must also be checked separately. First-half net cash outlay for royalty advances actually fell from €377 million a year earlier to €292 million, with the company citing contract timing and increased recoveries.[3] This cash flow deterioration cannot be explained by advance expansion alone, and conversely, a quarter improved by recovery timing cannot be seen as permanent cash generation power.
A Shareholder Map Changed Within a Year
Market views on the value of music catalogs also split clearly this year. Pershing Square, a top shareholder for years, moved to unwind its stake after its acquisition proposal was effectively rejected, and UMG announced on June 4 that it had directly repurchased 14,156,285 of its own shares at €17.66 per share, totaling €250 million, as part of the disposition of the entire position held by Pershing Square funds.[9] On May 29, UMG's board officially announced that it had rejected the proposal on the grounds that it significantly undervalued the company.[16] A shareholding relationship of several years was wound down this year.
Pershing Square's official interim report offers a more direct explanation. Pershing Square Holdings stated that it exited its UMG investment in the first half ending June 30, 2026.[14] Its earlier 2025 annual report had been optimistic about improved streaming agreements and competitiveness in the AI era.[17] But that view cannot be grafted onto grounds supporting a current holding. Past optimism and the subsequent actual sale are records from different points in time. That said, one shareholder's unwinding of a position is not a verdict on the business itself. It may be a choice shaped by the failed acquisition proposal overlapping with fund circumstances, and multiple layers of interpretation are needed between that judgment and the catalog's cash flows.
Meanwhile, the company itself is buying back stock. UMG announced in its first-quarter results that it had expanded its share buyback authorization to €1 billion, planning to start an additional €500 million program once the €500 million program announced at the end of March is completed. It also decided to monetize half of its Spotify stake, saying it would share proceeds with artists and use the company's portion primarily for share buybacks. The board stated that the share price was undervalued relative to the company's performance and outlook.[11] The mechanics of buybacks are simple: fewer shares mean the same catalog cash flow is divided into more per share. But if the repurchase price is above intrinsic value, it instead destroys value—so this, too, is a claim involving management's judgment, not a settled answer.
A Good Catalog and a Good Purchase Price Are Not the Same Thing
The portfolio-wide price-to-value ratio (P/V) disclosed by Longleaf was in the low 50% range at the end of September.[1] This means the management team views its holdings at a substantial discount, but it is not the discount rate on UMG alone. Whether the manager's valuation is right, and when that value will be reflected in the share price, are separate questions.
In evaluating UMG, saying only that the music industry is growing is not enough. How much of subscription price increases flow through to revenue and profit, whether growth excluding Downtown is sustained, and how much the money spent on renewals and catalog investment increases future cash flows must all be examined together. Even with good adjusted EBITDA, if the cash returning to shareholders is weak, an attractive-looking multiple can feel cheaper than it is. This year's traces show how wide the disagreement over price is. The company's board announced an expanded buyback in April[11], and Pershing Square exited the investment in the first half.[14] The split over buying and selling the same company with the same results means the point that feels cheap itself depends on valuation assumptions.
AI holds two-sided possibilities. New services leveraging licensed music could become revenue sources, but competition for listening time and pressure on royalty distribution could also intensify. Merely owning rights does not guarantee making money in either outcome. On this point, Longleaf's past investment experience is a useful clue, not an answer guaranteeing future returns.
Conclusion
The decision to buy UMG in an era when AI easily makes music reads as focusing not on the supply of music but on the demand—people continuing to seek out music—and on the contracts that connect that demand to revenue. A falling cost of making songs does not necessarily mean the price of already-loved music must fall at the same speed.
However, for that story to work as an investment, the catalog's popularity must translate into cash, and the share price paid to obtain that cash must also be reasonable. What Longleaf has shown is less the correct answer for the AI era than a value investor's question of re-examining, when supply overflows, the assets into which demand actually gathers. UMG's next test, too, will lie not in the number of songs but in the cash that music leaves behind and in bargaining power. This year's changed shareholder map is a repetition of the same question. A large holder of long standing exited the investment in the first half[9][14], the company itself is buying back its stock[11], and the value investor who came in cited durable earnings power.
References
[1] Longleaf Partners Global Fund 3Q26 Commentary — October 2026, as of September 30
[2] UMG — Q4 and Full-Year 2025 Results, March 5, 2026
[3] UMG — Q2 and Half-Year 2026 Results, July 30, 2026
[4] UMG–Udio agreement announcement, October 29, 2025
[5] UMG–ElevenLabs agreement announcement, September 10, 2026
[6] IFPI — Global Music Report 2026, March 18, 2026
[7] Southeastern — Our Governing Principles
[8] Southeastern — Our Approach
[9] UMG — Repurchase of 14,156,285 own shares from Pershing Square funds, June 4, 2026
[13] UMG 2025 Annual Report — pp.93–94,203
[14] Pershing Square Holdings — 2026 Interim Report, exited equity positions p.19
[15] UMG–Deezer — September 6, 2023 artist-centric royalty model announcement
[16] UMG — Announcement of rejection of Pershing Square proposal, May 29, 2026
[17] Pershing Square Holdings — 2025 Annual Report, past UMG investment views
Disclaimer: This article is informational content summarizing disclosures, analysis, and the author's views; it is not a recommendation to buy or sell any particular security.