Lending to Creator-Led Businesses
Debt Structures for the Next Generation of Media
Executive Summary
Many creator-led media businesses born on YouTube and other digital platforms are generating recurring, predictable cash flow at scale and yet the segment remains one of the least-crowded corners of media credit. The reasons are understandable: the space is younger and can appear more platform- and key-person-dependent. However, amidst these valid concerns exists an investable ecosystem that more clearly resembles the better understood television market.
This paper is about lending against the cash flows derived from durable IP assets that are born out of creator-led brands and studios. We cover why these businesses borrow, how the revenue engine works, what the collateral actually is and how to control it, the three debt structures we see in the market, and how AI changes the forward view.
I. The Market
The creator economy now runs to hundreds of billions of dollars in annual revenue, with more than 200 million people creating content globally and tens of millions who treat it as a primary occupation.
Creators who have transitioned from individual proprietors to leading real operating companies, with staff, production infrastructure, and diversified revenue, have the working capital needs of any growing business. In addition, a class of acquirer-operators has emerged to roll up channels and catalogs, utilizing third-party debt and equity capital to finance their purchases.
For the right assets, debt fits the profile. While back catalogs of YouTube channels, online courses, and podcasts tend to decline faster than music or film royalties, those revenues (once stabilized) are similarly predictable and high-margin. When a lender can get comfortable that a meaningful share of cash flow comes from previously released content, the asset stops looking like a bet on a person and starts looking like a traditional IP library with a potential replenishment mechanism.
II. The Role of YouTube
Most of the financeable cash flows — those that are large enough, durable enough, and supported by a well-tested collection and reporting infrastructure — currently sit on YouTube, so that is where the analysis starts. By Nielsen's measure, YouTube is now the largest in-home television streaming service in the United States and it pays out the majority of the advertising revenue it collects to the rightsholders who supply the content. It is also a primary beneficiary of the migration of ad dollars from linear and cable toward connected TV, a shift still in its early innings.
A lender should not confuse the platform with the asset. Platforms and distribution technologies come and go. Television has moved through broadcast, VHS, DVD, cable, and streaming, but the value of deep catalog titles did not die with any of them. We think of platform risk as a potential temporary leak in a pipeline rather than the permanent loss of the oil. If the underwriting was correct, when one channel weakens, the content should still maintain or rebuild its audience via other channels growing to absorb it.
YouTube has become a place where genuine, durable intellectual property gets built in the first place: original characters, formats, and worlds that command an audience on their own merits, before any studio is involved. The clearest proof is that this IP increasingly "graduates" to the largest stages in entertainment.
In 2026, two horror films made by YouTube-native creators on shoestring budgets became outsized theatrical hits. Obsession, from creator Curry Barker, was made for approximately $750,000, with Focus Features acquiring distribution rights for a reported $15 million. It opened to $17.2 million domestically, then rose 30% to $22.4 million in its second weekend and another 22% to $27.4 million in its third. The film has now grossed more than $400 million worldwide. Backrooms, adapted by Kane Parsons from his YouTube series that had generated more than 190 million views, was produced for approximately $10 million. It opened to $81.4 million domestically and $118 million worldwide — the largest opening in A24's history — and has now grossed approximately $390 million worldwide.
These are outliers and no lender should underwrite a creator loan on the hope of producing the next one. The everyday version of this is far more mundane: a catalog developed on e.g. YouTube gets syndicated to an AVOD platform such as Tubi, licensed to a subscription platform such as Hulu, or re-packaged into a course sold on a website. The breakout films referenced above simply illustrate that the underlying assets, the IP and its audience, have value that exists independent of any single platform's payout policy. That is what a lender is ultimately secured by.
III. What You Actually Lend Against
Lending against a creator-led business such as a YouTube channel means lending against a bundle of IP, first-party audience data, and brand. At its core is the catalog of existing content, which should not be limited to any one distribution platform.
Around that may sit owned & operated distribution assets that can also survive any one platform, such as email lists, social handles, and owned web properties, plus brand and trademark rights, sponsor and product-placement relationships, and often adjacent businesses: courses, merchandise, paid memberships, podcasts. A coding YouTube channel can incubate a course business; a cooking channel may have expanded into a cookware brand, etc. As in any other business, these adjacencies are important diversifiers of the revenue mix, though lenders should obviously not extrapolate their existence if they haven't yet been launched.
The catalog, or "back-book", is content already produced and monetizing at close to 100% margin. The new-content engine depends on the operator continuing to create. The back catalog is the primary source of recoverable value because it may continue generating revenue without additional production. The lender must nevertheless test how much of that consumption depends on traffic and engagement generated by new releases. The new-content engine is the growth, but is also where the key-person risk lives. The lender's first job is to measure how much of today's cash flow rests on each.
The harder part is control. A security interest in platform revenue is worth only as much as the lender's ability to redirect that revenue when the borrower defaults or tries to divert it. Many channels are paid not by the platform directly but through a multi-channel network that aggregates and remits the money. An irrevocable letter of direction instructs that network to pay into a controlled account rather than to the creator, intercepting the cash before it reaches the borrower's general funds. That account is governed by a deposit account control agreement in the lender's favor. A UCC-1 perfects the interest in the revenues and intangibles while a copyright mortgage perfects against the underlying content. If the borrower attempts to reroute revenue, third-party enforcement/monitoring software services help the lender capture the full stream.
For larger facilities, a bankruptcy-remote SPV can sit beneath this layer to isolate the collateral from the operator's broader credit. It carries cost and friction and is not always worth it at this asset size. The decision depends on deal size, counterparty quality, and how cleanly the revenue can be ring-fenced.
IV. Cash Flow: Cohort Decay and Durability
We forecast revenue the same way across media: by the vintage of release and a decline curve thereafter. Content is released, draws a burst of attention and revenue and decays as attention moves on. The open questions are how steep the early decline is and where the curve settles.
For digital video the shape is pronounced. Revenue from content under a year old falls very steeply as the release surge fades. The decline moderates over the next several years. Content past five years in age tends to settle into a shallow, slow decline. Genuinely evergreen material, content whose appeal is not tied to a moment, usually finds a stabilized base of consumption five to seven years after release.
The advantage over legacy media (e.g. film libraries) is the data. Viewership and revenue are reported granularly and close to real time. We can see or estimate what share of current revenue comes from content more than twelve months old, how fast each vintage declines, and whether the back-book stands on its own or is propped up by the audience that new content brings in. That last question is the most important input to the runoff analysis and in this asset class you can actually answer it with data rather than make educated guesses.
A library spread across hundreds or thousands of videos behaves like a portfolio: no single title's idiosyncratic decay threatens the whole and aggregate revenue is far more predictable than any one title. A channel whose revenue concentrates in a handful of viral hits carries the opposite profile, closer to a single-title film library where one asset's decline drives the outcome. Asset count and revenue concentration are therefore a credit input in their own right, and a deep, well-diversified catalog supports a higher advance rate than a shallow or top-heavy one, all else equal.
V. Comparing Assets
The market is large and fragmented. A lender cannot say yes to everything and needs a way to prioritize. Our framework for asset selection has three components.
The first is longevity: evergreen vs. ephemeral. Is the appeal of the content durable or tied to a moment/trend/news cycle? Evergreen content underwrites to a longer, shallower decay and a stronger runoff case. The second is key-person dependency: faceless vs. celebrity-driven. How much of the value is tied to one individual who cannot easily be replaced? The more the brand is the person, the more transition risk the lender carries. The third is new-content dependency: catalog-driven vs. new-release-driven. How much of current revenue is sustained by the back-book versus dependent on a constant stream of new releases? This axis is tied most directly to the runoff case.
Because the data sets in this industry are rich, these don't have to be qualitative calls. Each component is informed by measurable inputs: share of revenue from content more than twelve months old; audience retention; engagement; and subscriber and watch-time growth.
VI. Three Debt Structures
There are three principal ways to lend into this market. They differ in what secures the loan, how predictable the cash flow is, and what recovery looks like.
Catalog-secured loans
Catalog loans are secured by an identified library, including the underlying IP as well as exploitation rights, and sized approximately to the standalone cash flow of that back-book. Recovery on the loan should theoretically not depend on the operator producing anything new. While pure catalog loans in the creator space are rare, this underwriting approach is instructive for all deal structures.
Corporate loans
Corporate loans are made to the operating company, whether a creator business, a studio, or a roll-up, secured by a blanket lien over the business and its cash flows. This is the most flexible structure for the borrower and the right one where the borrower is a diversified, professionalized operator. The trade-off is that the lender takes some level of going-concern and key-person exposure alongside catalog value. Depending on the size of the loan, repayment may depend on the business continuing to function, not just on the back-book continuing to pay.
Advances and license-based structures
Advances are recoupable against future revenue, closer in form to a music advance or a film minimum guarantee. While advances can (and should) be secured by the underlying catalogs similar to term loans, there is typically no maturity and no explicit amortization schedule. The funder advances a sum and recovers a fixed dollar amount, the advance plus a defined fee, through a sweep of a set share of revenue from a specified pool, ending when that dollar cap is reached.
Since the dollar return is fixed, the multiple of invested capital is roughly fixed across performance scenarios. What moves is the timing, and therefore the IRR. Strong performance hits the cap sooner and lifts the IRR; weak performance stretches the timeline and lowers it.
Additional downside protection features can include an escalating sweep (e.g. a baseline rate in the ordinary course, but a higher rate if the creator misses minimum content-output requirements, up to 100% capture on default or sustained underperformance), posting covenants and cross-collateralization triggers, which let the funder reach across all of its deals with a counterparty when performance deteriorates.
| Corporate loan | Catalog-secured loan | Advance / license | |
|---|---|---|---|
| What secures it | Blanket lien on the operating company | Identified catalog, IP & monetization rights | Revenue from a defined pool |
| Collateral scope | Entire business & cash flows | Back-book and its rights | Explicit share of revenue |
| Cash flow predictability | Going-concern dependent | Back-book driven | Variable timing; performance-driven |
| Key-person exposure | Higher | Lower | Moderate |
| Recovery profile | Going-concern or restructuring | Runoff of the catalog | Fixed-dollar cap via revenue sweep |
| Maturity & amortization | Term + amortization | Term + amortization tied to decay | No maturity; ends at the dollar cap |
| Right sizing lens | Leverage / LTV | LTV vs. runoff value | Recoupment coverage & sweep velocity |
VII. Sizing and the Runoff Case
Sizing starts with an estimation of an asset's market value, which frames the attachment point. Single channels tend to price in a mid-single-digit multiple of earnings. Diversified businesses built on top of channels command meaningfully more.
These starting points to an analysis must be triangulated with a runoff case. Assume new-content activity winds down, let the catalog decline along its curve, exit the residual at a conservative multiple, and discount back. Size the loan so that this stream, rather than the going-concern projection, covers principal. In our experience, this supports a starting point around three times earnings, or roughly a 60% loan-to-value assuming a reasonable 5x value of the asset, varying with catalog durability, revenue diversification, and new-content dependency.
In the illustration below, a business with $300 of Year 0 revenue and $240 of operating cash flow runs off to a present value of $676 — implying roughly 2.8x leverage against current earnings and a 56% loan-to-value against a $1,200 asset value.
| Year 0 | Year 1 | Year 2 | Year 3 | Year 4 | Year 5 | |
|---|---|---|---|---|---|---|
| New-content revenue | $100 | $20 | $16 | $14 | $13 | $12 |
| YoY decline | -80% | -15% | -10% | -10% | -10% | |
| Catalog revenue | $200 | $145 | $132 | $121 | $113 | $107 |
| YoY decline | -28% | -9% | -8% | -7% | -5% | |
| Total revenue | $300 | $165 | $148 | $135 | $126 | $119 |
| Operating cash flow | $240 | $132 | $118 | $108 | $101 | $95 |
| Terminal value (5.0x × Yr 5 OCF of $95) | $475 | Current earnings (Yr 0 OCF) | $240 |
| PV of Yr 1–5 operating cash flow @ 12% | $407 | Implied leverage (676÷240) | 2.8x |
| PV of terminal value @ 12% | $270 | Asset value (5.0x × Yr 0 OCF) | $1200 |
| Runoff present value | $676 | Implied LTV (676÷1200) | 56% |
Illustrative, round numbers. 80% OCF margin; 12% discount rate; 5.0x terminal and asset multiple.
For an advance structure, sizing relies on simulating whether the expected cash flow sweep recoups the fixed cap inside an acceptable window and how that recoupment holds if performance disappoints. Because the dollar cap is fixed, MOIC stays at roughly 1.4x across every scenario; what changes is how long recoupment takes, which moves the IRR from around 30% on a fast sweep to roughly 14% on a slow one.
| Scenario | Collection dynamic | MOIC | IRR | Time to recoup |
|---|---|---|---|---|
| Outperformance | Fast sweep → cap reached early | ~1.4x | ~30% | ~2.3 yrs |
| Base case | On-plan sweep | ~1.4x | ~20% | ~3.2 yrs |
| Underperformance | Slow sweep → cap reached late | ~1.4x | ~14% | ~4.5 yrs |
Illustrative, round numbers. $100 advanced against a $140 fixed repayment cap, 60% revenue sweep.
VIII. The Impact of AI
No current assessment of creator-economy credit is complete without addressing generative AI.
On the supply side, a steep increase in AI-generated content volume appears inevitable, but its long-term commercial viability is unproven. Audiences vote with their time and their money, and the test for any AI catalog would be the same test we apply to all content. After enough time for the initial cycle to decay and revenue to stabilize, is it still generating steady royalties? If the answer turns out to be yes for a meaningful subset of AI-first content, that becomes a new asset class worth underwriting on its own merits. We are not there yet, and no one has the data to underwrite that question today.
The more relevant question for existing credit is how AI content affects demand for the human-made catalogs already in the ground. We distinguish demand that is pulled by consumers, audiences seeking out a specific creator, personality, or franchise, from demand that is pushed by platforms through algorithmic recommendation. Pulled demand is significantly more insulated from AI substitution. Pushed demand is not, because a recommendation engine that begins surfacing AI-first content can redirect attention that currently flows to human creators. Assessing that mix for a given asset is becoming part of diligence. The related risk is that AI content compresses attention and accelerates catalog decay, which strikes directly at the runoff case, since a steeper curve erodes exactly the back-book the loan relies on.
For now, we see no evidence that the major platforms are systematically promoting AI-generated content over human-made work. Several, YouTube among them, have taken public positions that favor the discoverability and monetization of authentic content. That posture could change and a lender should monitor it rather than assume it. Established back-books retain their position in the recommendation systems that drive a meaningful share of passive revenue today.
AI is not only a risk. It is also a margin and output lever for existing operators, lowering production cost, accelerating localization and dubbing into new markets, and extending the life of a catalog, all of which can strengthen a credit.
In summary, our approach is to weight the evergreen and pulled-versus-pushed assessments more heavily and lean toward shorter tenors/faster amortization where new-content and pushed-demand dependency runs high. We believe generative AI content proliferation is a reason for more rigor, rather than a binary reason to not participate in this sector.
IX. Risks and Mitigants
The central risk is faster-than-expected catalog decay, typically because new-content dependency was underestimated and the back-book turned out to be propped up by audience arriving for new releases. It is the risk that most directly threatens principal and can be managed with conservative decay assumptions, appropriate loan sizing/attachment point, and setting the pace of amortization to the catalog's vintage profile.
Key-person dependency can further be addressed through transition provisions and continuity covenants.
Idiosyncratic platform actions, from demonetization to algorithm and feature changes, are addressed through multi-platform distribution requirements, control over collection accounts, and the escalating sweeps and cross-collateralization triggers described above.
Most platform risk is gradual and diversifiable, but certain content categories carry binary, permanent-impairment risk. Content directed at children sits under heightened regulatory and monetization restrictions and can be demonetized en masse by a single policy change. Sexually suggestive, violent, or otherwise brand-unsafe content is perpetually one advertiser boycott or platform enforcement action away from losing its revenue base.
The cleanest mitigant is not to price around these categories but to screen them out at underwriting. We generally avoid lending against catalogs whose revenue depends on content that is brand-unsafe or targeted at children (other than premium/proven IP), because no advance rate adequately compensates for a risk that is binary rather than gradual.
X. Conclusion
Creator-led digital media is younger than music or film, but it is maturing quickly and the toolkit required to lend against it is similar. Find off-the-run cash flow with real asset coverage, separate the durable collateral from the going-concern story, control the cash flow operationally rather than on paper, and size the loan to what survives in a runoff. Platforms will keep evolving and AI will reshape parts of the supply and demand picture in ways no one can fully forecast. The catalog and the underlying IP remain the asset: content that has proven it can hold an audience and that increasingly travels across formats and platforms.
For those interested in discussing in more detail, please don't hesitate to reach out: info@mepcap.com
Cite as: "Lending to Creator-Led Businesses: Debt Structures for the Next Generation of Media." MEP Capital Management LLC, 2026. DOI: https://doi.org/10.5281/zenodo.22024082. The full paper is also available as a PDF on Zenodo.
This paper was prepared by MEP Capital Management, LLC, a U.S. SEC registered investment adviser and reflects the current opinions and estimates of the firm, which may change without notice. This report is for informational purposes only and nothing contained herein should be interpreted as official accounting of investment performance, a forecast of future events or a guarantee of future results. Figures contained herein are obtained from sources deemed reliable, but we cannot guarantee their accuracy or completeness. All contents copyright © 2026 MEP Capital Management, LLC.