I think about the beginning of my copywriting career very often. Back in the 2000s, many things we do by default now were a genuine hassle. Finding images for campaign proposals depended mostly on how much patience you had and how much storage you could spare. In most agencies there were entire folders packed with stock photos from previous campaigns, recycled across briefs because there wasn't much alternative. I arrived in the agency world in 2007, when designers were manually removing watermarks from Getty Images and other photo banks in Photoshop, just to show a client a proposal. If the campaign was approved, the photo was bought. If not, the watermark had been removed for nothing.

In 2010, Freepik changed that entirely. Suddenly, millions of free images were available to everyone, not only to agencies with licensing budgets. The deal was simple: use what you need, and credit Freepik in return. For us, that meant no more Photoshop workarounds and no more hours of searching. It changed the way creative work moved.

By 2021, Freepik had 41 million monthly users and more than half a million paying subscribers. A successful company in every measurable way. Then OpenAI's DALL-E 2 appeared in 2022 and the ground shifted in a way that felt familiar: the same disruption Freepik had delivered to the image banks in 2010 had now arrived at Freepik's door.

When DALL-E 2 appeared, most of the stock image industry reacted the way established companies usually do. They looked at what they already owned and asked how to protect it.

Freepik itself was doing very well. It closed 2022 with €80 million in revenue, almost 25% more than the year before, and more than 550 employees. Nothing in the numbers said it needed to change.

Cuenca himself was not an early believer. He was sceptical of the first image models, which failed at basic things, and he was convinced the creative spark was something very human. DALL-E 2 changed his mind. "It's like the first ChatGPT," he said later. "It hits you." It hit him hard enough to go all in.

For many managers, that moment would have been frightening, especially for those emotionally attached to their positioning, their niche and all the work done before. Cuenca was a different profile. Before Freepik, he had co-founded Panoramio, a photo-sharing site that Google bought in 2007 and folded into Google Maps. He knew what it was to build something and then watch it become part of something bigger. He was not afraid of changing direction.

So instead of freezing, he moved the company towards the very thing that was threatening it. Freepik started building generative AI tools, first for images and then, in 2025, for video. It was a bet that required dismantling significant parts of what had made Freepik successful, the stock asset model, the freemium architecture, the brand positioning, before those things stopped working commercially.

It was not painless. The company went from around 550 employees to 400, and Cuenca has admitted there is "some pain in any transformation". But he made the change while the old business was still growing, when there was still money and time to do it properly.

The journey from Freepik to Magnific is nowhere near a typical rebrand. As part of the restructuring Freepik went through in the wake of AI, it acquired Magnific in May 2024.

Magnific had been built in Murcia by two people, Javi López and Emilio Nicolás, with no investors, no team and no advertising. What López did have was a year of posting about generative AI on X, simply because he was curious about it. When they launched their upscaler in late 2023, more than 30,000 people joined the waiting list in the first 24 hours. Some early users complained it was not a "real" upscaler, because it invented detail instead of just enlarging the image. That was exactly why designers loved it. Five months after launch, it had 725,000 registered users.

Cuenca knew about them because of their Spanish origin and their success, and decided to close the deal. What nobody knew at the time was that the product would become the brand less than two years later.

On April 28, 2026, Freepik announced it was becoming Magnific. By then, Freepik had sixteen years of brand equity, while Magnific had two and a half years of momentum. The conventional decision would have been to absorb the acquisition into the parent brand: Freepik AI, Freepik Pro, Freepik Studio, or something similar. Instead, the parent brand disappeared into the acquisition.

The vision Cuenca and the board had was based on relevance. As Cuenca put it, people had been seeing the company in fragments, a stock bank on one side and an upscaler on the other, when in reality it had become one creative platform. Freepik as a name wasn't saying anything interesting anymore, especially not about artificial intelligence. Magnific, from magnify, described what the tool did to your work: it made it bigger and better. The new identity, developed with the agency AREA 17, follows the same idea, with two squares expanding upwards.

The change had a cost. For sixteen years, millions of designers typed "freepik" into Google whenever they needed an icon or a background, and that habit now ends at a redirect. The company also ended up with two Magnifics: magnific.com, the new home for everything, and magnific.ai, the original upscaler, with separate accounts. Its own help page has to explain what to do if you have one of each. Cuenca and his team decided, in effect, that a few years of confusion was a smaller risk than spending the next ten as a stock image company in people's heads.

Cuenca frames what comes next as a third category: the industrial revolution created blue-collar work, the digital revolution created white-collar work, and AI is creating something with no collar at all. Work that once required a studio, a crew and real capital, now doable by one person with a vision and a browser tab.

The numbers back the framing. Seventy-two percent of new creators joining Magnific describe themselves as beginners, people the old Freepik model was never built to serve directly. Video now accounts for roughly half of the company's revenue, a category that barely existed for them three years ago. And the work is real. Magnific's tools were used in the Amazon Prime Video series House of David, in campaigns for Puma and Carl's Jr., and by the BBC.

Magnific does not even build its own AI model. It openly says it doesn't claim to have the best one. Users choose between models from Google, ByteDance and others, and Magnific provides everything around them, from upscaling and editing to keeping characters and scenes consistent across a video. The bet is that creative teams care more about the workflow than about which model sits underneath it.

Cuenca was changing more than the brand and the culture. He was changing the category. Freepik competed in the image bank sector. Magnific competes in creative production, with a speed the old Freepik could never match.

The old image banks faced the same moment and made very different choices.

Shutterstock decided its library was raw material for AI. In 2023 it signed a six-year deal allowing OpenAI to train on its images, followed by deals with Meta and others. AI licensing brought in $104 million in 2023, and its CEO expected $138 million in 2024. It is real money, although it comes from helping to build the tools that are replacing the stock photo.

Getty chose the courts. In January 2023 it sued Stability AI in London over the images used to train Stable Diffusion. In November 2025, the High Court largely ruled against Getty. Its one small win was on trademark, because some of the images Stable Diffusion created carried copies of Getty's watermark. I couldn't help thinking of those designers in 2007, removing the same watermark by hand. Almost twenty years later, a machine was putting it back.

Then the two companies tried to merge, in a $3.7 billion deal. After a long regulatory review, the UK's competition authority asked for Shutterstock's editorial business to be sold as a condition. Getty refused, and in July 2026 the merger was terminated.

Shutterstock sold access to its archive and Getty defended its own in court, and then they tried to put the two together. Each move rested on the same idea: that the archive was the company. Magnific also has an archive, more than 200 million assets, and it still sells them, but they are no longer the main thing it offers.

Most brands treat their name as their identity. The name carries the accumulated meaning of everything the company has done: the positioning, the community, the trust, the recognition.

Freepik spent more than fifteen years delivering high-quality images for free. Magnific completed that sequence by becoming an AI platform that speeds up the creative process itself. Two very different business ideas, with the same contribution underneath. In 2010, it meant free images for people who couldn't pay for a Getty licence. In 2026, it means video tools for people who could never pay for a film crew.

There was also something in the old name the company needed to leave behind. In its own telling of the story, Magnific admits that early decisions, like living on advertising and competing on price, slowed its path to becoming a premium platform. A name that starts with the word "Free" was never going to help with that. Dropping it was a pricing decision as much as a brand one.

The transformation took three years, and the company did it at its own pace. Freepik was never part of Silicon Valley. It started with almost no capital, and although the Swedish fund EQT bought a majority stake in 2020, it has never raised a venture capital round. Competitors like Runway are VC-backed, and Adobe Firefly belongs to one of the biggest software companies in the world. A different pressure level.

Patiently, they became experts. Today the company generates $230 million in annual recurring revenue, with more than one million paying subscribers and over 290 enterprise teams, including the BBC, Guess and R/GA. And it is profitable.

Freepik became Magnific because the reality of the market needed it. After working with creatives for more than a decade, the company wanted to give them the tools to make the images themselves, and to make their creative work faster. Some see it as an evolution. I see it as a sacrifice of the old for the sake of the new.

The image banks are still here, still protecting their archives. Freepik had one too, and a name millions of people knew. It kept the archive and gave up the name.

I still think about those designers in 2007, carefully removing a watermark for a pitch that would probably never be approved. The company that saved them all that work has now done the same thing to itself. It removed its own name, because it was standing in the way of what came next.