Good morning producers 👋🏼

As we are knee deep into the holiday season, we'll keep it brief! This newsletter will be our update for both November and December, so don't expect to hear from us again until the January news update in February.

Spotify top 5 2025 were released... in 2024?

Be honest with me right now. If you were to sit and think of what the biggest songs this past year were… what would you say? I'm talking those big culture-wave type songs that take over the cultural zeitgeist for the year and become the only thing you can hear no matter where you go. If you're having a tough time coming up with them, you're not alone. You might think some song in the latest Taylor Swift album, or perhaps Golden from KPop Demon Hunters, but neither of those made the cut for top 5. The most streamed songs this year were:

• Die With a Smile — Lady Gaga and Bruno Mars (1.7 billion)
• BIRDS OF A FEATHER — Billie Eilish (1.5 billion)
• APT. — ROSÉ and Bruno Mars (1.5 billion)
• Ordinary — Alex Warren (1.4 billion)
• DtMF — Bad Bunny (1.2 billion)

This is not to say that these tracks aren't absolute bangers, but compared to previous years, 2025's top five is unusually dominated by songs released the year prior. In fact, the only tracks on this list that were released in 2025 were #4 Ordinary and #5 DtMF.

So what does this say about the state of our music industry? Maybe nothing. I mean, how important is it really for the top song to be released the same year? That being said, it's certainly uncommon, so let's dive into it.

In the era of TikTok and streaming services it seems that the only time people collectively sit down to watch or listen to something older is during the holiday season. Otherwise, it's always about the latest release on Netflix or the newest YouTube video discussing the most recent turn of events. The Spotify algorithm, however, seems not to follow the same set of rules as other platforms. A recent article by HEC Montréal examines the Spotify recommendation system and showcases the 3 main components that the algorithm takes into account:

• Content-based filtering: analysis of the songs themselves through:
  • Metadata provided by the artist or label
  • LLMs scanning all available content for the artists such as songs, lyrics, cover art, social media reviews, etc.
  • An audio analyzing model which analyzes the BPM, timbre, key signature, etc.
• Collaborative filtering: analysis of the consumer and how users interact with Spotify through active and passive listening.
• Feedback: once recommendations are made, the system optimizes itself through the data it collects from the user's actions.

Due to the nature of the algorithm itself, Spotify is likely to recommend songs it has more data on. This is resulting in recommendations of already popular music. The interesting thing is that although music recommendation is becoming less diverse, music consumption is simultaneously becoming more fragmented — meaning music is becoming less diverse within each listener's world while becoming more fragmented across listeners. The result is that breakout hits are becoming more rare, while evergreen hits are more prominent. Spotify's top 5 is testament to that on 2 fronts. Firstly, as we have already discussed, the majority were released the year prior, making them more evergreen, and secondly, the top 3 are extremely established artists.

In the radio era, music was controlled by the record labels and the DJs. Labels paid heavily to promote singles, sometimes transparently, sometimes not. But DJs still had a real influence. They championed records they believed in, broke artists regionally, and occasionally took risks on something unproven. That combination of centralized promotion and human curation created conditions where breakout hits could happen quickly, and when a song caught on, it did so everywhere at once. The audience wasn't segmented into millions of personalized feeds, it was largely hearing the same thing at the same time. The result was fewer total hits, but bigger ones, and a stronger sense of shared cultural moments.

By the way, we'd love to hear your thoughts on the newsletter. We see that about half of you are opening and reading it, which is awesome! If you ever want to drop a thought in response to a story feel free to e-mail us at hello@timeoff.audio or just press on the button below.

Hit us up

We need to talk about Suno

As much as we don't want to constantly talk about AI in our industry, it's dominating the news both in the music industry and broader, but we will try to keep it condensed as much as possible to allow space to talk about all the other important things going on.

One of the biggest stories in our industry this past month is that AI music generation software Suno raised $250 million in a Series C funding round at a $2.45 billion dollar post-money valuation according to the company. For those of you not intimately familiar with the startup funding model, here's a quick breakdown:

• Angel investors -> individuals who invest early on based on the belief of the founders or the idea before there's a real product or revenue.
• Seed round -> the first formal funding used to build the initial product, test the concept, and find early traction.
• Series A -> the first major institutional round focused on refining product-market fit, building the core team, and proving the business model.
• Series B -> a growth round aimed at scaling, expanding users, markets, infrastructure, and sometimes starting to build multiple product lines.
• Series C -> a late-stage round for accelerating dominance, international expansion, acquisitions, large team scaling, or prepping for an IPO or major exit.

The latest funding round is aimed to "redefine what is possible in music creation". Back in September, Suno released "Suno Studio", a browser-based DAW focused on an AI generated workflow. It functions like a multitrack DAW, with the ability to generate entire songs as well as individual instruments based on prompts and recordings. Now I'm about to say something extremely controversial, and we want to hear your thoughts as readers of this newsletter, so please reach out. Here is the controversy: this is a much more interesting AI generative future than the simple "prompt based whole song output".

Imagine this. You're in your DAW, you've written the structure for an entire song, and then you want to add guitar chords, violin pizzicatos, and synth stingers here and there. You can do all of that with your voice. And if you don't like the sound, you can export the data as MIDI and bring it back into your DAW to try different instrument textures. There's still a million and one issues at play here, don't get me wrong. From the training data being used without consent, to the incredible amounts of energy being used to generate the prompts, to the real professionals losing work to AI — all of that is still a problem. But for the first time since I heard the term "AI music", I can squint and see a glimpse into a future of AI generated music that is creative and collaborative.

Okay, that's great and all. But let's actually talk about the money now. It's kind of hard to wrap our heads around how much a $2.5 billion dollar valuation is, so let's compare this to other companies in our industry:

• AVID (Pro Tools, Media Composer): $1.4 billion, 2023
• Dolby Laboratories, Inc.: $6.3 billion, 2025
• Yamaha: $3.14 billion, 2025
• Focusrite: £132 million, 2025
• Universal Music Group (UMG): $46 billion, 2025
• Spotify: $115 billion, 2025

We tried finding data for Universal Audio, Ableton, and Native Instruments, but as privately held companies this is much harder to swing. That being said, Suno is a 3-year old company valued at $2.5 billion — bigger than AVID and about the same size as Yamaha. So how does this shake down when talking about annual revenue?

• AVID 2023: $400 million
• Yamaha 2024: $3 billion
• Suno 2025: $200 million

A few things to note here. Firstly, all of these figures are gross annual revenue. So Yamaha being an order of magnitude higher in terms of revenue does not necessarily mean much because hardware is a much less lucrative industry than software. And secondly, Suno's revenue is self-reported, so take that with a grain of salt. What really matters of course is net revenue, which is much harder to find, especially when discussing privately-owned companies. That being said, all of this still begs the question "what is Suno's business model?" Is it mass market appeal or professional production? As with everything else in the AI world, the business model still needs to be proven. $200 million in annual revenue might sound impressive, but the truth is that those operating costs are probably sky-high as well, so $200 million might not be profitable. And if the goal is to continue developing AI tools to improve multi-track production generation, it might not be for a while.

Is that song using my voice?

On October 29, HAVEN (the artist name of Harrison Walker) released his debut single i run. The track had all the makings of a TikTok breakout: two minutes long, algorithm-friendly pacing, minimal production, etc. That is, until listeners started asking a very specific question:

"Is that Jorja Smith?"

The vocalist on the track was uncredited, but to many ears, the voice sounded unmistakably like the British R&B artist. Close enough that online speculation quickly turned into something more serious. Jorja Smith is now reportedly seeking a share of the royalties, despite not writing, recording, or producing the song. The track was removed and re-recorded with Kaitlin Aragon. You can still hear the original one on YouTube and hear for yourself.

On one hand, this isn't a new problem. Voice imitation, soundalikes, and deepfakes existed long before generative AI became mainstream. Labels have used "sounds like ___" singers for decades. On the other hand, tools like Suno and Udio have dramatically lowered the barrier to generating convincing vocals, and more importantly, to doing so at scale.

Traditionally, copyright protects compositions and recordings, not timbre. You can copyright a performance, but not a voice. That legal gap is now being stress-tested in real time. If a song uses a vocal that sounds like a famous artist, but no copyrighted material was directly sampled, what exactly is being infringed?

There's an interesting parallel to draw here with film and television. Actors do own their likeness — you can't just use an actor's appearance, voice, or recognizable identity to sell a product without permission. Using someone's likeness in a way that implies endorsement or association is protected under what's known as "rights of publicity". This is why studios will negotiate likeness rights, digital doubles require contracts, and why recent strikes by SAG-AFTRA focused so heavily on protections against unauthorized digital replication.

Music has never had an equivalent protection, and the problem is now impossible to ignore. If an AI-generated voice clearly evokes a specific artist, triggers public confusion, and drives commercial success, the question becomes "are you using their likeness?"

This dispute lands at an especially awkward moment for the industry. While the "big three" labels have spent the past year suing Suno and Udio for large-scale copyright infringement, that posture just shifted. As of this month, the major labels and the AI companies have announced a closed settlement and a new "strategic agreement." They're no longer fighting, they're collaborating.

Labels are moving toward licensing deals with generative platforms, while individual artists are left to personally defend their voices. We may be heading toward a future where AI voice generation is legal, licensed, and normalized. Grimes started this back in April 2023, publicly inviting people to use her AI-generated voice and offering a 50% split of royalties. Now it's going to happen at scale.

Honourable mentions

• The studio that recorded Michael Jackson, Whitney Houston, and Ariana Grande is up for sale
• Beyoncé: the highest grossing solo tour of the year
• YouTube ditches the Billboard charts

Thanks for taking the time, and as always, feel free to reach out via email or Discord for any questions or feedback!