Adobe seems obsessed with Firefly… but I think it could be on the way out

If you've visited an Adobe Max event recently, or streamed one online, you know the drill. Every presentation, every demo, every conversation revolves around one thing: Firefly. Adobe's "commercially safe" AI has been its obsession, pushed as the solution to every creative challenge.

At Adobe Max 2023, it was all about Firefly integration into Photoshop and Illustrator. Max 2024 brought more Firefly features, including video generation. The message was clear: Adobe had built the future of creative AI, and it was called Firefly.

Trained only on licensed content, it was the responsible choice. No copyright nightmares, just clean, ethical AI generation.

But then something odd happened.

Cracks appear

Earlier this year, Adobe started integrating third-party AI models into Firefly. First came Google's Gemini 2.5 Flash, then partnerships with OpenAI, Ideogram, Pika, Black Forest Labs and Runway. This month's integration of Luma AI Ray3 continues the pattern.

Think about this. Imagine if buying Photoshop in 2010 also gave you access to GIMP and Paint.NET. You'd think Adobe had lost its mind. Yet here we are, with Adobe essentially admitting that its audience might need an alternative to its flagship AI; one that it doesn't actually make.

Adobe is spinning this as a positive thing, of course; it wants Firefly to be "the creative AI ecosystem of the future". But strip away the marketing guff, and what you're seeing is Adobe recognizing that people want alternatives to Firefly – specifically, alternatives that aren't "commercially safe."

If Adobe truly believed in Firefly's superiority, why promote rival models? Does it fear that, well, Firefly just isn't good enough?

Here we go again

To this hackneyed tech journalist, it all feels familiar. Once upon a time, Adobe XD was going to revolutionize digital interface design. Adobe invested heavily, built communities, pushed it as the future. Then Figma destroyed it so thoroughly that Adobe tried to buy it for a cool $20 billion.

When that acquisition failed, did Adobe double down on XD? It did not. XD was quietly abandoned while everyone kept using Figma.

Before that, the same thing happened with Flash. Adobe had spent years evangelizing Flash as the future of web animation, building massive developer communities around it. Then the iPad happened, Steve Jobs wrote his famous anti-Flash letter, and Adobe couldn't distance itself fast enough.

So here's the thing about Adobe: it's brilliant at making products, and it's brilliant at building enthusiasm for them. It'll spend years convincing you that its latest innovation is essential… but then it may just quietly shuffle it aside when something shinier appears.

Firefly shows all the signs of being another Adobe U-turn. Initially, it was convinced it was the answer to creative AI. Now, barely two years later, it's hedging bets by partnering with every AI model maker available.

The commercially safe pitch is already being watered down, with disclaimers about user responsibility for tracking which model generated what.

Don't get too attached

I really like all the people I've met at Adobe. They've treated me well over the years, and I think it's an excellent company that makes excellent products. But I have to be honest: when companies start promoting rival products alongside its flagship offering, it does give me cause for concern.

So my advice to anyone who's bought into Firefly is, enjoy it whilst it lasts – but don't get too attached. Adobe's current AI darling may possibly be heading for the same fate as its previous "revolutionary" products.

These third-party AI integrations just mean that Adboe is expanding its offering to customers, the company claims – and I'm sure it genuinely believes that. But it kinda feels like it's also preparing its exit strategy. So, if that exit does come, don't say I didn't warn you.

You might like...

Whatever happens next, make sure you're ready for it with the best photo-editing laptops and the best laptops for video editing.