Sweden's 2030 Digital Strategy vs Its Collapsing Digital Present

Last week the Swedish government published Digitaliseringsstrategin 2025–2030, a five-year roadmap for how Sweden will lead in digital transformation. The same week, Dagens Medicin ran a piece called "Sverige backar in i den digitala framtiden," Sweden is backing into the digital future. Dala-Demokraten went further: "Sverige håller på att radera sitt demokratiska minne." Sweden is erasing its own democratic memory. Nobody in government connected these dots out loud. I'm going to.

I run an AI development company in Europe out of Jönköping. Not Stockholm, not a lab, not a ministry. We build things that ship, for clients who need them to work on Tuesday, not in a five-year vision document. So when I read a strategy paper projecting Sweden's digital leadership through 2030, written by the same institutional machine whose healthcare digitization is collapsing right now, I don't see contradiction. I see cause and effect.

The Strategy Document Is a Confession

Here's the pattern I've watched for two decades of building software in Sweden: the further a government or organization drifts from actually building something, the more comfortable it gets writing about the future. Strategy documents are what you produce when you've stopped shipping. They cost nothing to write, nothing to be wrong about, and by the time anyone checks whether 2030 arrived on schedule, the people who signed off will be in different jobs. Meanwhile Dagens Samhälle just ran a 60-voice pre-election roundup on healthcare's "digital crossroads." Sixty voices. Not sixty solutions. Sixty people describing the same crossroads from sixty angles, right as an election approaches and everyone needs to sound serious about digitalization without actually having fixed anything.

Dagens industri covers Social Democrats now demanding tougher build requirements on new datacenters, which is a real policy conversation Sweden needs to have. But notice the sequencing: infrastructure demands arrive as opposition positioning during an election cycle, not as an operational plan tied to a government's own digital delivery record. You cannot credibly regulate the infrastructure of the future while your regional healthcare systems are running on record-keeping so fragile that journalists are calling it a democratic memory problem.

What "Erasing Democratic Memory" Actually Means

The Dala-Demokraten piece isn't hyperbole. When patient records, case histories, and administrative documentation degrade or become unrecoverable across siloed regional systems, that's not an IT inconvenience. That's the erosion of the paper trail democracy runs on: who decided what, when, based on what information, and who is accountable. Sweden built a welfare state on the promise that the state remembers you. If the digital infrastructure underneath that promise is rotting while a strategy document promises leadership five years out, the strategy is not describing reality. It's covering for it.

How This Looks From Jönköping, Not Stockholm or Brussels

From where I sit, running software development in Sweden for clients across Europe, the gap between Stockholm policy and regional execution is not new. It's structural. Regeringen writes the strategy. Regions and municipalities own the execution, the budget shortfalls, and the legacy systems nobody wants to touch because ripping them out risks patient data during the transition. That's the actual bottleneck, and no five-year vision document fixes a procurement process still running on 15-year-old vendor contracts.

Compare that to what I see when I look at where AI and infrastructure investment is actually happening. The US is building compute capacity at a pace that makes European datacenter debates look quaint, whatever you think of the wisdom of it. Asia, particularly Singapore and increasingly Gulf state-backed initiatives, moves from policy to procurement in months, not electoral cycles. The EU's answer has been the AI Act: a compliance framework that is genuinely useful for governance, but that arrives as a checklist for companies to pass, not as a forcing function that makes any government actually rebuild its own broken healthcare IT.

I'm not arguing Sweden should deregulate everything and copy Silicon Valley. I've built here my whole career because Swedish engineering culture, when it's actually building, is excellent. Spotify, Klarna, Truecaller, these companies came out of Swedish technical talent solving real problems fast. The issue isn't Swedish capability. It's that government digitalization strategy and government digitalization execution have become two entirely separate professions, and only one of them has consequences.

The EU AI Act Won't Save Anyone From This

Every conversation I have with founders and CTOs this year eventually turns to AI Act compliance: risk classification, documentation burden, conformity assessment. Fine. Necessary. But compliance checkboxes don't fix an execution gap. A hospital system that can't reliably retrieve a patient's medication history isn't failing an AI Act audit, it's failing at the basic digital plumbing the AI Act assumes already works. You cannot bolt responsible AI governance onto infrastructure that can't do reliable record-keeping in 2026. That's the uncomfortable order of operations nobody in the strategy document wants to say out loud.

Where This Actually Goes: 2026 to 2030

Here's my honest read on the next five years, the same window the strategy document claims to own.

Sweden will hit some of its stated 2030 digitalization targets, mostly the ones that are already privately built infrastructure with a government label slapped on afterward. Klarna-adjacent fintech infrastructure, private sector AI adoption, and telecom will look fine because they were never the government's to build in the first place.

Public sector digital infrastructure, especially healthcare, will continue its slow-motion crisis through at least 2028 unless something forces a real capital and organizational commitment, not another strategy refresh. The 60 voices in that Dagens Samhälle roundup will still be talking about the crossroads in three years, because a crossroads with no decision-maker attached to a budget line is just a permanent feature, not a moment. The datacenter build-requirement debate the Social Democrats are pushing now will matter more than most people currently think. As AI compute demand keeps climbing toward genuinely capable models with real reasoning ability, energy and infrastructure policy becomes AI policy, whether or not anyone calls it that. Sweden has cheap, clean power and cold climate, genuine structural advantages for datacenter build-out. If the political fight over "tougher requirements" turns into years of delay instead of a fast, clear framework, Sweden hands that advantage to Norway, Finland, or Gulf-backed capital that moves faster.

And when the AGI conversation stops being theoretical, and I think that's a three-to-five year horizon for something that meaningfully changes how governments operate, the countries that win are the ones whose basic digital infrastructure already works. Not the ones with the best-written strategy about how it will work eventually. If your civil service can't keep a reliable medical record today, no amount of AI capability arriving in 2029 fixes that. AGI doesn't repair broken plumbing. It just runs faster through whatever pipes you already have.

What Builders Should Actually Do About This

If you're a CTO or founder reading a government strategy document and wondering what it means for your roadmap, here's my answer: almost nothing, directly. Government digital strategy in Sweden right now is a lagging indicator of political priorities, not a leading indicator of infrastructure you can build on. Plan around what actually exists, not what's promised for 2030.

Second: the execution gap is an opportunity, not just a scandal. Every region and municipality quietly patching together legacy healthcare and public records systems needs actual working software, not another strategy consultant. If you're building in the public sector adjacent space, this is where rapid MVP development and pragmatic AI solutions beat five-year vision decks every time. Ship something narrow that works, prove it, expand it. That's how Sweden's actual digital wins happened historically, not top-down.

Third, take security and compliance seriously regardless of what the government strategy says, because NIS2 and DORA enforcement is real even when public sector digitalization isn't. If you haven't run a proper posture check, Prowler gives you open-source cloud security and compliance automation that's actually being used at scale, not theorized about. Pair that with CISO Assistant for GRC across ISO 27001, NIS2, GDPR, and DORA in one place. Neither of these came out of a government ministry. Both came out of people who needed the thing to work.

What to Look At This Week

• Prowler (14.6k stars, Python): open-source cloud security and compliance automation, genuinely useful if you're navigating NIS2 or DORA requirements without a compliance department. github.com/prowler-cloud/prowler
• CISO Assistant (4.3k stars, Python): a one-stop GRC platform mapping 150+ frameworks including NIS2, DORA, GDPR, ISO 27001. Built by people solving their own compliance headache, which tells you something. github.com/intuitem/ciso-assistant-community
• Baserow (5.6k stars, Python): if your organization's actual problem is "we have no database and too many spreadsheets," this is a GDPR and SOC 2 compliant no-code database platform that ships faster than any government procurement process ever will. github.com/baserow/baserow

The Real Test

I'll believe Digitaliseringsstrategin 2025–2030 is more than a confessional document the day someone in government points to a specific, funded, accountable project that fixes healthcare record reliability in a named region by a named date, with a named person's job on the line if it slips. Not a target. A commitment with consequences. Until then, treat every five-year strategy document from any government as what it actually is: a description of what the current leadership wishes were true, written by people who will not be in the room when 2030 arrives to explain why it isn't. If you want proof Sweden can still build, don't look at the strategy. Look at the companies quietly shipping software that works, in Jönköping, in Malmö, in Umeå, without waiting for permission from a five-year plan. That's where I'd put my money, and where I put my company.

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.