Sweden Outsources Its Digital Future and Calls It Innovation

Sweden's most iconic companies are handing the keys to someone else

SKF, one of the most recognizable industrial brands Sweden has ever produced, just signed a multi-year digital transformation deal with Tata Consultancy Services. Let that sink in. A company born in Gothenburg in 1907, a symbol of Swedish engineering excellence, is now outsourcing its core digital capability to India. Not a side project. Not a peripheral system. The digital backbone.

The same week, Implema and New Minds announce an SAP Academy because Swedish companies cannot find people to run the enterprise software they already own. And Karolinska Institutet is still holding events about "deploying trustworthy AI in healthcare" while Egypt is actively expanding healthcare digitalization with Swedish partners and India is convening bilateral talks with Sweden on telecom and digital innovation. Sweden's partners are building faster than Sweden itself.

I run a tech company from Jönköping. Not Stockholm. Not San Francisco. Jönköping. And from here, the view is painfully clear: Sweden is not building the future. It is purchasing it off the shelf, slapping a Swedish flag on it, and calling it innovation.

The SKF deal tells you everything

Let me be specific about what the TCS deal means. When a manufacturer like SKF decides that its digital transformation should be led by an external partner headquartered in Mumbai, it is making a statement. That statement is: we do not have the internal capability to do this ourselves, and we have decided that building it is not worth the effort.

That is a rational business decision for SKF. TCS is excellent at what it does. They have scale, process maturity, and deep enterprise integration expertise. I respect TCS. The problem is not TCS. The problem is that Sweden produced SKF but cannot produce the digital infrastructure SKF needs to survive the next twenty years.

This is not an isolated case. Look across Swedish industry. Volvo, Ericsson, H&M. Every major Swedish company now relies heavily on external technology partners for core digital work. The in-house capability has been hollowed out over a decade of cost-cutting, offshoring, and the comforting lie that "we're a business, not a tech company."

Every company is a tech company now. If you outsource that, you outsource your future.

The SAP Academy problem is an admission of failure

Implema launching an SAP Academy sounds like a positive story. Skills development. Education. Workforce investment. Read it again. They are launching this because Swedish companies cannot staff the platforms they already own and pay millions of kronor per year to license.

This is not a skills gap. This is a skills vacuum. And it tells you something uncomfortable about how Sweden has treated technical talent for the past fifteen years.

We pushed everyone toward product management, UX design, and "digital strategy." We told engineers they should become team leads by year three. We created a culture where the person who makes the PowerPoint is valued more than the person who writes the code. And now we are shocked that nobody wants to configure SAP modules for a living.

The situation is worse than it looks. SAP S/4HANA migrations are hitting deadlines across Europe. Companies that delay will lose support for legacy systems. And Sweden does not have the people to execute. So what happens? More TCS deals. More Infosys contracts. More Wipro engagements. The digital backbone of Swedish industry, running on talent pools that Sweden did not invest in building.

At HEIMLANDR, we see this from the builder's side every day. Companies come to us after burning through a year with a large consultancy that delivered a strategy document and a roadmap but no working software. They need something built. Actually built. And they are surprised that a small Swedish company can do it faster and cheaper than a 500-person offshore team.

Karolinska and the Swedish disease of perfectionism

Karolinska Institutet is one of the finest medical research institutions on the planet. It awards the Nobel Prize in Medicine. When they talk about AI in healthcare, people listen. And what are they saying right now? They are hosting discussions about "deploying trustworthy AI" focused on validation and adoption frameworks.

Meanwhile, Egypt is deploying healthcare digitalization at scale with Swedish partners. India is having bilateral discussions about telecom and digital innovation implementation. Not workshops. Not frameworks. Implementation.

Sweden has a disease. I call it consensual paralysis. We study the problem. We form a working group. We publish a report. We discuss the report. We hold a seminar about the discussion. And while we do all of this, someone else ships the product.

Trustworthy AI is important. Validation matters. But at some point you have to actually deploy something and learn from real-world data. The irony is devastating: the countries buying Swedish medtech are deploying AI faster than the country that invented the medtech.

If you are a hospital system, a research institution, or a healthtech company in Sweden, you need to stop workshopping and start building. Custom AI solutions are not theoretical anymore. They are production-ready. The risk is not deploying too early. The risk is deploying too late and finding that the standard of care has moved on without you.

How Sweden compares to everyone else

Let me put this in global context because the picture is not pretty.

United States: Pouring hundreds of billions into AI infrastructure. Every major company is building internal AI capability. Even mid-market firms are hiring AI engineers. The default posture is build, not buy.

India: Not just an outsourcing destination anymore. India is building its own digital public infrastructure. UPI processes billions of transactions. The India Stack is a model that other countries study. And now India is convening bilateral talks with Sweden about digital innovation. Think about that power dynamic for a moment.

China: DeepSeek and dozens of other labs pushing open-weight AI models. Massive domestic AI agent development happening in manufacturing, logistics, and healthcare. No workshops. No consensus-building. Deployment at speed.

Estonia: A country of 1.3 million people with more advanced digital government infrastructure than Sweden. X-Road has been running for over two decades. They did not workshop it. They built it.

Sweden: Hosting events. Writing reports. Outsourcing to TCS. Launching academies to teach people how to use software we bought from Germany.

The Nordic myth is powerful. We tell ourselves we are innovative, technologically advanced, early adopters. And twenty years ago, that was true. Spotify, Skype, Klarna, King. Sweden punched way above its weight. But that generation of companies was built by founders who ignored the establishment, not by the establishment itself. The establishment is what we are looking at now. And it is not building anything.

Where this goes: 2027 to 2030

Here is what I see coming, and it is not comfortable.

AI automation in business will split companies into two categories within three years. Those that have internalized AI capability and those that rent it. The ones that rent it will pay more every year and control less every year. This is the same dynamic as cloud computing ten years ago, but faster and with higher stakes. If your AI automation business strategy is "we'll buy it from our vendor," you are already behind.

AI agents are about to restructure enterprise operations. Not in the way consultants describe with nice diagrams. In the way that actual agent systems handle workflows that today require 5-10 people coordinating across email and Slack. AI agent development is the highest-leverage investment most companies can make right now. And almost nobody in Swedish industry is doing it internally.

The path toward AGI changes the outsourcing math completely. If you outsource your digital capability today, you are outsourcing to humans. In three years, you will be outsourcing to organizations that wield increasingly autonomous AI systems. The knowledge gap between you and your vendor will become an intelligence gap. And intelligence gaps do not close. They compound.

EU regulation will not save you. The AI Act is well-intentioned but it is designed to regulate, not to accelerate. Sweden's position as a compliant, well-regulated market is not a competitive advantage when the competition is deploying at 10x speed. Regulation gives you guardrails. It does not give you a car.

Blockchain and decentralized infrastructure will matter more as AI scales. When AI systems handle more business logic, the question of verifiability becomes critical. Blockchain development is not hype anymore. It is becoming necessary infrastructure for proving that AI-generated outputs, transactions, and decisions are authentic and auditable. Sweden should be leading here. It is not even in the conversation.

What to look at

If you are a technical leader or founder trying to actually build instead of buy, here are concrete resources worth your time this week:

OpenHands (75K+ stars on GitHub). AI-driven development that lets you go from specification to working software with an AI agent. This is the kind of tool that makes a 5-person team as productive as a 50-person offshore engagement. If SKF's leadership looked at this, they might reconsider the TCS deal.

Agno (40K+ stars). Build, run, and manage agent platforms. If you want to understand what AI agent development looks like in practice, not in a strategy deck, start here. This is the infrastructure layer for enterprise agents.

rtk (58K+ stars). A CLI proxy written in Rust that reduces LLM token consumption by 60-90%. Single binary, zero dependencies. This is the kind of practical engineering that makes AI deployment affordable for teams that are not burning venture capital. If you are doing any LLM integration work, this saves real money immediately.

Daytona (72K+ stars). Secure infrastructure for running AI-generated code. As AI writes more of your software, the execution environment matters enormously. Daytona solves the "where does this code actually run safely" problem.

So what do you actually do?

If you are a CEO or CTO at a Swedish company reading this, here is what I would tell you if we were having coffee in Jönköping.

Stop treating digital capability as something you can buy. You cannot. You can buy tools. You can buy temporary labor. But capability is something you build, retain, and grow. If you outsource it entirely, you do not have it. You have a dependency.

Hire engineers and treat them like the strategic assets they are. Not as cost centers. Not as resources to be optimized. As the people who will determine whether your company exists in ten years.

Deploy AI now. Not after the workshop. Not after the framework is finalized. Not after the EU AI Act gives you a checklist. Now. Start small. Learn fast. Build internal muscle memory for working with AI systems before your competitors make it the default.

And for the love of everything, stop confusing procurement with innovation. Signing a contract with TCS is not digital transformation. It is digital delegation. There is a difference, and in five years, that difference will be existential.

I build things from Jönköping because I believe Sweden still has the talent, the discipline, and the engineering culture to compete globally. But that belief gets harder to defend every time I see another headline about another Swedish company handing its future to someone else. We have a choice. We can keep buying off the shelf. Or we can remember what made Sweden matter in the first place: we built things that the world wanted.

Time to build again.

Fredrik Brunnberg is the CEO of HEIMLANDR.IO, building AI and software solutions from Jönköping, Sweden. This is the daily HEIMLANDR briefing. If you found this valuable, share it with someone who builds things.