ElevenLabs Music v2: Multi-Genre AI Audio Generation

What has changed with ElevenLabs Music v2

On May 28, 2026, ElevenLabs announced the release of Music v2, the second generation of its AI-based music synthesis model. According to The Decoder, the model can handle transitions between musically distant genres — opera, heavy metal, rap — while keeping the track's harmonic and structural coherence.

Music v2 also introduces an audio inpainting feature. It lets you select a specific section of a generated track and regenerate it in isolation. The remaining parts stay intact. This makes the revision process surgical, not destructive.

Previously, any change to an unsatisfying section meant regenerating the entire track. As a result, creative control was limited and timelines stretched. Music v2 solves this bottleneck head-on.

The architecture of multi-genre coherence

The core technical challenge of Music v2 isn't generating individual genres. Earlier models — including the first version of ElevenLabs Music — could already produce stylistically coherent tracks within a defined genre. The leap in quality lies in the transitions between genres.

Keeping a track musically coherent while moving from opera to metal means managing timbre, tempo, harmony and rhythmic structure all at once. ElevenLabs says Music v2 can navigate these transitions without losing the track's narrative thread. From a model architecture standpoint, that's no small feat.

Similarly, the inpainting feature requires the model to understand the surrounding musical context before regenerating the selected section. In this sense, the model works with a form of musical context awareness that brings it closer to the behavior of a human editor.

Immediate impact for creative agencies and SMEs

For agencies producing video content, ads or digital campaign materials, original music is often a significant fixed cost. Licensing commercial tracks or hiring professional composers weighs on production budgets, especially for SMEs.

That's why tools like ElevenLabs Music v2 open up a different way of working. An agency can generate original music tailored to each format — reels, pre-rolls, podcasts, corporate presentations — with no recurring licensing costs. Inpainting also lets you quickly adapt the same track to different creative variants.

At SHM Studio, we work every day with Italian SMEs running multichannel campaigns. Shorter audio production time has a direct impact on speed to market. An agency that builds Music v2 into its workflow can deliver audio variants in hours, not days.

To learn more about how AI fits into content production, check out our pages on AI services and SEO copywriting.

Three real-world use cases for Italian SMEs

First, let's look at where Music v2 delivers the biggest operational return. Here are three representative scenarios for the Italian B2B and retail market.

• Social video campaigns: generating original music for reels and short videos eliminates the risk of copyright claims on platforms like Instagram and YouTube. Multi-genre coherence also lets you vary the musical style for different audience segments without needing separate tracks.
• Podcasts and branded audio content: SMEs producing corporate podcasts or industry interviews can generate custom jingles and musical stingers. In particular, inpainting lets you update individual sections as the brand evolves, without redoing the entire sound identity.
• Presentations and event materials: agencies running B2B events or industry trade shows can produce tailored soundtracks for every moment of the day — opening, networking, closing — with a single tool and in less time.

These scenarios tie directly into the digital marketing work and LinkedIn campaigns we manage for our clients. Audio becomes part of your content strategy, not an afterthought.

The competitive landscape: where Music v2 stands

ElevenLabs isn't the only player in this space. Still, the combination of multi-genre transitions and inpainting sets it apart from its most direct competitors. Tools like Suno and Udio offer text-to-music generation, but with less control over precise section-by-section revisions.

According to Gartner, generative AI applied to creative media is rapidly maturing. As a result, the gap between leading models and secondary ones tends to widen fast. Those who adopt the most advanced tools today build an operational edge that's hard to close later.

It's also worth noting that ElevenLabs already holds a strong position in the AI voice synthesis market. So its expansion into music follows a platform logic: a single provider for voice, sound effects and original music. For agencies, that means a simpler tool ecosystem.

If you run Google Ads campaigns with video components, high-quality original audio can improve ad engagement metrics. For YouTube-related SEO, original music also removes restrictions on monetization and distribution.

What the press releases don't tell you

It's wise to stay critical. ElevenLabs says multi-genre transitions happen without losing musical coherence. However, how good those transitions feel depends on the context of use and on what the final audience expects.

For content aimed at social platforms or corporate videos, Music v2's quality level is probably good enough. By contrast, for productions that need a distinctive, refined sound identity — TV spots, soundtracks for high-profile corporate films — you'll need to test it yourself before judging.

Also, the inpainting feature, promising as it is, comes with a new learning curve. Fine-grained control over an AI track takes some basic musical skills to use effectively. Even so, for a creative team with just a basic grasp of musical structure, the operational advantage is clear.

If you want to explore how to integrate these tools into a structured content strategy, our team is available through the contact page. And to stay up to date on the AI developments that matter most for Italian SMEs, the SHM Studio blog regularly publishes analysis on these topics.

What to do now: three operational priorities

For SMEs and agencies that want to evaluate Music v2 in a structured way, we suggest three immediate priorities.

• Map your existing audio touchpoints: identify every content format that currently uses purchased or licensed music. This inventory defines where the tool can be applied right away.
• Launch a pilot on one specific format: choose a low-risk format, such as social reels, and test Music v2 across a full production cycle. Then measure the time saved and the quality your client perceives.
• Build audio into your content strategy: treat original music not as decoration but as a brand identity variable. In particular, assess how well your musical style fits your brand positioning.

To dive deeper into building a content strategy that combines audio, video and text coherently, our pages on web services and digital marketing are a useful starting point. And if you're an SME looking to understand how AI fits into an editorial plan, the AI services section describes the adoption paths we follow with our clients.

Outlook: where AI audio generation is heading in 2027

Music v2 points the way forward; it isn't the finish line. According to McKinsey, generative AI applied to creative media has significant automation potential in content production processes. As a result, over the next 18-24 months it's reasonable to expect models that are even more controllable and easier to integrate into agency production pipelines.

In particular, the convergence of speech synthesis, music generation and AI video production suggests that by 2027-2028 it will be possible to produce complete audiovisual content (voice, music, visuals) in a single AI-assisted workflow. For Italian SMEs, this scenario redefines the relationship between production budget and the quality of the final content.

That's why building hands-on familiarity with tools like ElevenLabs Music v2 today is more than a tactical choice. It's an investment in your ability to compete in a content ecosystem that is being redrawn fast. If you wait for the market to settle, you risk starting out already behind.