Freepik Becomes Magnific, Lovable Hits $100M ARR, and the Week AI Creative Tools Got Serious

Freepik sheds its name and calls everything Magnific

Freepik rebranded as Magnific this week, and honestly it's about time. For years the company operated a confusing portfolio: Freepik was a stock asset library, Magnific was the AI upscaler everyone actually talked about, and a handful of other products sat somewhere in between. Now it's all Magnific, one name for the whole stack.

The company is bootstrapped, profitable, and worth $230 million. It's been quietly printing money in Málaga while everyone else chased VC rounds. The rebrand doesn't change the product much—it just stops pretending the various tools are unrelated. If you've used Magnific's upscaler, you know it works. The new name is the company admitting that's the center of gravity.

This is smart. Brand confusion kills momentum, especially when you're selling to creators who just want the thing that works. Freepik was the legacy brand; Magnific was the future. Now there's no split.

Lovable hits $100M ARR in eight months

Lovable crossed $100 million in annual recurring revenue eight months after launch. That's fast enough to make it the fastest-growing software startup in history, if you trust the press release math. The Swedish company sells a no-code AI platform that lets non-technical users build apps and websites from text prompts.

This is the "vibe coding" category—tools that let you describe what you want and get a working prototype. Lovable isn't the first, but it's the first to scale this aggressively. The growth suggests two things: demand for no-code tools is real, and Lovable's execution is better than the pack.

The platform competes with everything from Webflow to Replit's Ghostwriter to every "chat-to-code" experiment running on GPT-4. What's different here is revenue proof. A lot of these tools get buzz but never convert. Lovable apparently does both.

Worth watching: whether this ARR is sticky or inflated by early adopter churn. Eight months is too short to declare victory, but it's long enough to suggest the product isn't vaporware.

Synaps raises $3.6M for AI architecture design

Synaps, an AI design canvas for architects, closed a $3.6 million pre-seed round. The Vienna-based startup launched in beta five months ago in Tirana, which has become one of Europe's most active architectural scenes. It now has 60,000 users and hundreds of paying customers.

The product is a collaborative canvas where architects can sketch, refine, and iterate on building designs with AI assistance. It's not generative design in the "let the machine do everything" sense—it's more like a smart co-pilot that suggests options and handles tedious layout work.

Three Albanian-born founders built it. They picked Tirana for the beta launch deliberately, betting that architects in a fast-growing city would actually use the tool under time pressure. Apparently it worked.

This is part of a broader trend: vertical AI tools that target specific professions (architects, lawyers, doctors) instead of trying to be everything to everyone. The narrow focus makes the product better and the go-to-market easier. Synaps benefits from both.

Bond launches a social network with no feed

Bond, a "post-feed" social network, went live this week. There's no infinite scroll, no algorithmic feed, no timeline. Instead, the app uses AI trained on your photos, videos, and audio to recommend real-world activities. The idea is to fight doomscrolling by getting you off the app.

The founders are Dino Becirovic (ex-Index Ventures) and Arthur Brazinskas (ex-Google DeepMind). They're betting that people are tired of feeds and want something that actually suggests things to do instead of things to watch.

It's an interesting premise, but the execution risk is high. If the recommendations are bad, the app has no fallback content layer to keep you around. And if they're good, the app trains you to leave—which is the point, but also a brutal retention challenge.

Social networks that position themselves as anti-social (less screen time, more real life) tend to struggle because the incentives are backwards. Users say they want less engagement, but their behavior says otherwise. Bond's bet is that the behavior is starting to shift. We'll see.

The rest: trading platforms, IP tools, and Google expanding Gemini

B2BROKER Group shipped an AI assistant for its B2TRADER platform, targeting brokers who want AI-first infrastructure. Genki AI launched globally with an autonomous engine for IP management that automates the path from creative input to global fulfillment. And Google expanded Gemini to Android Automotive, replacing Assistant in cars.

None of these are going to dominate headlines, but they're all part of the same pattern: AI tools moving from prototype to production, from generic to vertical, from buzz to actual revenue. The creative and design space is especially hot right now—every category from architecture to stock photos to social networks is getting rebuilt with an AI layer.

The question is whether any of these tools stick, or whether we're just in the hype window before consolidation. Lovable's ARR suggests at least one is sticking. The rest are still proving it.