(Almost) Daily Musings

Evening all

Just as I renewed my FT subscription, Bloomberg tries to pull me back in with emails like these:

Little does Mike Bloomy know that in the aftermath of this year’s Christmas Parties I am not tempted by his offer — even on a Thursday.

Energised by my 767th coffee, I share today’s musings.

Cheers
Philip

Today’s Stories

Spinoffs: Coming to a TV station near you soon

Warner Bros. Discovery is doing the corporate equivalent of spring cleaning, separating streaming from dusty cable networks. The official line reads “Enhance clarity and focus.” Feels like a pre cursors to an block sale or a strategic breakup. A move that screams How to Spinoff Your Problems 101.

Proof? Let’s check out the credits:

• Comcast (2024): Spun off NBCUniversal cable networks (minus Bravo)

• Lionsgate (2022-2023): Split Starz, sent it packing

• ViacomCBS (2019): Broke out streaming and legacy networks. Sold Simon & Schuster right after

• Time Warner Cable (2009): Spun off cable, sold it to Charter for $78B in 2016

Chart Art

Freedom AI

Discussing these two concepts in one is a big ask. Reid Hoffman did so in a recent London lecture. Some interesting highlights synthesised on how AI can enhance freedom as well as restriction. Drawing from Isaiah Berlin’s concepts of liberty, Hoffman framed AI as a tool that can either inhibit freedom (negative liberty: freedom from barriers) or empower individuals (positive liberty: the ability to act on one’s agency)

While sceptics like Yuval Noah Harari argue that AI poses an existential threat to human agency, Hoffman sees a different possibility: AI as an enabler of what he calls "superagency"—a decentralized tool that amplifies individual capability. Think of it as an enhancing companion — a sentiment echoed by Ashton Kutcher and Effi Epstein at a recent Sound Ventures Panel at TechCrunch Disrupt — AI as helping us become a plus version of ourselves (perhaps soon changing two and a half men to three and a half men)

But achieving the vision of liberating AI hinges on: Design and Governance.

Players in AI Ethics to keep tabs on

1. Mozilla Foundation: Builds open-source AI tools with a focus on transparency and privacy

2. EleutherAI: Community-driven research project emphasizing open-source model development, focussed on preventing AI misuse

3. NEC: Established the "NEC Group AI and Human Rights Principles", setting an ethical framework for AI

4. HuggingFace: Developed an ethical charter outlining AI use cases to avoid, including human rights violations

5. Fiddler: Offers tools for explainable AI, bias mitigation, and model monitoring

6. Arthur: Focuses on bias detection and explainability, reflecting a shift toward ethics in client priorities

7. Bluebash: Specializes in LangChain and ChatGPT integrations with an ethical focus

8. Adva Digital Solutions: A team-driven by morality and craftsmanship, designing AI with ethics at the core

Take a look at the funding data for AI ethics—it’s clear Europe is dominating this space, with the U.S. trailing and Asia barely registering. Regulatory-first mindsets (think GDPR) provide a potential clue on why Europe is leading the subvertical AI race on ethics — at least in deal figures. Governments that prioritize regulation tend to encourage innovation within ethical constraints. US remains laissez-faire, and Asia appears focused on other priorities

Software Private Equity

To continue my little exploration of neocloud bubbles, I continue on with a broader evaluation of how software, private credit and yield hungry sponsors created their very own bubble ecosystem.

2021 and 2022 were frothy across the board—not just in VC, but in tech, private equity, and credit markets alike. Software companies with questionable EBITDA metrics but “best-in-class” growth potential were trading at a hefty 9.1x top line multiple. The logic? Future growth would make up for the lack of profit today through the sponsor favorite value lever: Operating leverage (high fixed cost base spread over increasingly growing top line creating diminishing marginal unit cost).

Turns out, hockey stick projections did not stick.

And who said finance isn’t a creative profession? Lenders rediscovered their inner artists and got more than creative with the definition of covenants and the handing out of PIKs. PIKs, piling up to the sky where clouds and bubbles hang out, meant companies that took on this chunky leverage will soon face a massive lump-sum repayment. And nothing illustrates this more clearly than the credit maturity wall hitting opcos by 2028 and 2029—a cool $20 billion in software-related loans maturing in 2029 alone, mostly SMBs borrowing from BDCs.

A lot of this debt gets sold off to syndicators, CLO’d into the market, and—voilà—turned into someone else’s problem. Or at least that’s the theory. Big ‘08 déjà vus incoming. Double whammy isn’t just for baseball anymore…

Now, that’s all debt—but just how much PIK-related credit are we staring down? Check out this little gem of a skewed graph:

Just as expected—around $5 billion in PIK debt maturing in 2028. So, who’s holding the keys to these PIK notes? Fan and LME favorites: Blue Owl, Blackstone, New Mountain, and Clearlake (the latter deserving an honorable mention).

Blue Owl and Blackstone strike a particular chord with me, given their role in funding a significant chunk of the paper written to CoreWeave—the neocloud provider we spotlighted yesterday.

While the stated maturity for PIK notes—typically subordinated notes—falls within the 5-8 year range, the timeline for repayment can be significantly accelerated in certain circumstances. A notable example is a change of control, such as when the sponsor exits by selling the underlying asset.

In healthier exit markets (unlike the current drought), sponsors usually manage to offload assets within 5-7 years. This means that even though the loan's maturity might theoretically extend beyond this, the accrued liability often becomes due at the point of sale, effectively shortening the expected repayment horizon.

All things considered, this wouldn’t be a story if:

A) M&A and IPO markets showed any signs of life, and

B) The revenue growth and operational profile of the asset screamed “quality”

Looking at the revenue growth in the space, I am cautiously raising my eyebrows.

Cocktail: Dark and Stormy (2028 Edition)

Ingredients

• 1 part PIK leverage: Companies bought at record multiples, now drowning in slowing growth and comparatively weak EBITDA margins (if the expected revenue growth would have held up post 2021, the margins may have improved on operating leverage accounts but that bet did not pan out). Add a dash of frothless exit multiples (my Oxford Dictionary submission is pending).

• 1 part maturity wall: A hefty $20 billion in software-related loans coming due in 2029, with a $5 billion cherry of PIK debt maturing in 2028.

• A twist of distress: LMEs, distressed buyouts, and creditor closures may hit like a sudden squall.

Garnish

• A sliver of opportunity: Tactical buyout shops and DIP lenders poised to feast on the wreckage. If you enjoy distressed assets, 2028 might be your vintage.

Directions

Shake with as much Newtonian momentum as possible

Macro

European markets ended Thursday with a mixed performance, as the European Central Bank paired its fourth rate cut since June with a clear reminder: growth prospects aren’t looking great — for context: the ECB rolled out updated forecasts, cutting 2024 GDP growth projection to 0.7% (vs. 0.8%) and 2025 to 1.1% (vs. 1.3%). The benchmark deposit rate now sits at 3%, its lowest since March 2023, as the ECB takes a decidedly dovish turn.

• Stoxx Europe 600: Down 0.2%, with energy and mining stocks dragging the index

• FTSE 100: Up a modest 0.1%, keeping things stiff-upper-lip

• Dax: Gained 0.2%, a rare bright spot on the continent

• Cac 40: Unmoved—presumably sipping a café au lait, contemplating existentialism