From one of Europe’s biggest fintech exits to bootstrapping an AI startup: ‘There’s no limitation’

The Pivot from Open Banking to Autonomous Engineering

When you sell a company for two billion dollars, the tech world generally expects you to do one of two things: become a professional golfer or start a massive venture fund. Daniel Kjellén and Fredrik Hedberg, the minds behind the Swedish fintech giant Tink, decided to do neither. Instead, they are going back to the garage to build Freda, an AI startup that aims to tackle the specific, often painful friction points of software engineering.

After their exit to Visa in 2021, the duo spent time inside the corporate machine. While Tink was a success story in the open banking space—standardizing how financial data flows across Europe—their new venture is focused on the internal plumbing of development teams. They are moving from moving money to moving code, specifically looking at how large language models can handle the tasks that currently bog down human developers.

The Bootstrapping Paradox

The most interesting part of this story isn't the AI—it is the funding. Or rather, the lack of it. Despite having the kind of pedigree that makes VCs throw blank checks at their front door, Kjellén and Hedberg are bootstrapping Freda. For a founder who has already won the game, this is a strategic move that more builders should pay attention to.

When you have your own capital, you don't need to spend six months of the year pitching. You don't have to manufacture growth metrics to satisfy a board of directors who are looking for a 10x return on a specific timeline. Bootstrapping at this level allows for a pure focus on product-market fit. In the current AI hype cycle, where everyone is chasing the next massive round, the Tink founders are choosing silence and control. They are betting that their own experience and capital can get them further than a pressurized injection of outside cash.

Solving the Developer Bottleneck

We have all heard the pitch that AI will replace programmers. I don't buy that, and it seems the Freda team doesn't either. The goal isn't to fire the engineering team; it is to remove the limitations that prevent a small team from doing big things. In the traditional startup model, scaling usually means hiring more people. But hiring more people increases communication overhead, creates management layers, and slows down the actual shipping of code.

Freda is looking at the autonomous side of development. If an AI agent can handle the documentation, the initial scaffolding, or the basic debugging, a team of three can theoretically operate like a team of thirty. For builders, this is the holy grail. The bottleneck in most tech companies isn't a lack of ideas; it is the physical capacity to turn those ideas into stable code. If Freda can actually deliver on reducing that friction, they aren't just building a tool—they are changing the economics of how startups are built.

Why Experience Matters in the AI Rush

The AI space is currently flooded with twenty-somethings who have never seen a full market cycle. There is a lot of "wrapper" software being built—thin layers on top of OpenAI that don't solve deep problems. Kjellén and Hedberg bring a different perspective. They spent a decade building Tink, dealing with heavy regulations, complex integrations, and the realities of enterprise-grade security.

This is where the advantage lies. Successful founders know that the "magic" of a demo wears off quickly. What matters is reliability. By moving into AI with a builder-first mindset, they are likely looking at the plumbing rather than the paint. They understand that for AI to be useful in a professional engineering environment, it can't just be clever; it has to be predictable and integrated into existing workflows.

The Challenge of Autonomy

Despite their track record, the road ahead isn't easy. The AI engineering space is incredibly crowded. You have incumbents like GitHub Copilot and well-funded challengers like Cognition or Poolside. Freda is entering a fight against giants, and they are doing it with a lean team and their own bank accounts.

The skepticism here is obvious: Can a small, self-funded Swedish team out-engineer the massive compute budgets of San Francisco? The answer depends on their niche. If they try to build a general-purpose AI coder, they might get crushed. But if they focus on the specific structural pain points they felt while scaling Tink, they might find a wedge that the generalists missed. They are betting that their "no limitation" philosophy will allow them to move faster than the corporate-backed competition.

What Builders Should Take Away

There are three lessons here for anyone building in the crypto or AI space right now. First, your previous success is only a tool, not a guarantee. The fact that these guys are starting small and staying lean shows they respect the difficulty of the task.

Second, bootstrapping is becoming a power move again. In an era of high interest rates and cautious VC sentiment, being able to fund your own R&D is a massive competitive advantage. It allows you to ignore the noise and build for the long term.

Third, the real value in AI isn't in the models themselves, but in how they are applied to specific, high-value problems. Engineering capacity is the highest-value problem in the tech world. If you solve that, you don't need a marketing budget.

The goal for most founders is to get big enough to stop worrying about the details. The best founders realize the details are the only thing that actually matters.

Final Thoughts

I like this move because it lacks the typical ego of a second-act startup. There was no massive press release about a $50 million seed round. There was just a quiet launch and a focus on the work. For those of us in the trenches, watching how these seasoned veterans navigate the AI landscape will be a masterclass in disciplined building.

They aren't trying to change the world overnight. They are trying to make it easier to build things. In a world of hype, that is a mission worth watching. Whether Freda becomes the next Tink or stays a small, high-impact tool remains to be seen, but the approach is exactly what the industry needs more of: less talking, more shipping.

Read the original at Sifted →