Build vs Buy Is Dead. It's Build vs Rent vs Own Your Moat

I read two headlines this morning that should not exist in the same news cycle. Computer Sweden says Swedish companies are spending more on cloud subscriptions than ever. Same week, a report lands showing offshore dev rates in India and Eastern Europe are not collapsing the way everyone promised once AI coding tools got good. And then, quietly, CapMan sells a Swedish software company to Bridgepoint and Polaris. Real money, real exit, real proof that custom-built software is an asset, not a cost line you outsource away.

Three signals. One conclusion. Everyone is still fighting the wrong war.

The 2015 Debate Is Over and Nobody Noticed

Build vs buy used to mean something simple. Buy off-the-shelf SaaS because it is cheaper and faster. Build custom because you need something nobody sells. That framing made sense when building software meant hiring a team, waiting six months, and hoping the spec survived contact with reality. That world is gone. What I see at HEIMLANDR every week is founders and CTOs who can spin up a working MVP in days, not quarters, using rapid MVP development workflows built around AI agents doing the grunt work a junior dev used to do. The build side of the equation got radically cheaper. The buy side did not get cheaper at all. It got more expensive, because SaaS vendors know your data is locked in and your team is scared to migrate. So the old debate, build vs buy, is a false binary now. The real question is: what are you renting forever, and what should you own outright because it is your actual competitive edge?

Why Swedish Hourly Rates Stopped Being an Excuse

For a decade, the standard move for a Swedish company was: our dev rates in Jönköping or Stockholm are too high, so we buy SaaS or we offshore to Poland, Ukraine, India. That logic held when a Swedish senior engineer cost three times what a team in Bangalore cost for roughly the same output speed. It does not hold anymore, and the offshoring numbers prove it in a strange way. The piece making the rounds this week, "Everyone Said AI Would Kill Offshoring, The Numbers Disagree," shows offshore rates staying roughly flat even as AI coding assistants got dramatically better. Why? Because offshoring firms are not passing productivity gains to clients. They are pocketing the margin. You are paying 2023 prices for work that should take half the hours in 2026. Meanwhile a Swedish team using the same AI tooling, Claude Code with proper agent skills, Cursor, or Codex running against a well-structured codebase, closes that hourly rate gap fast. If a Swedish senior engineer with good AI tooling does the work of three offshore juniors, the math flips. Sweden's brutal hourly rate was never really the problem. Lazy scoping and lazy tooling were the problem, and offshoring just hid that laziness behind a lower invoice.

What CapMan Selling a Swedish Software Company Actually Tells You

Private equity does not buy sentimental stories. CapMan selling a Swedish developer to Bridgepoint and Polaris is a cold signal: custom, well-built Swedish software is a real asset class with a real exit multiple. Nobody is buying a company because it has a nice Salesforce integration. They are buying because it owns proprietary technology that generates cash and cannot be trivially replaced by a subscription. That is the whole argument in one transaction. Rented software is a monthly expense forever. Owned software, built right, is a balance sheet asset someone will pay a premium for later. If you are a founder deciding whether to build your core product internally or bolt together five SaaS tools and call it a platform, look at what actually gets acquired. It is never the SaaS stack. It is the thing underneath it that nobody else has.

The Real Framework: Build, Rent, or Own Your Moat

Forget build vs buy. Use three buckets. Rent: anything that is not your competitive advantage. Payroll, email, project management, CRM. Renting these is smart. Nobody wins by building their own Slack. Build: things you need custom but that are not your moat. Internal tools, integrations, automation glue. This is where fullstack development with AI-assisted teams crushes the old offshore-or-suffer choice. You get speed without giving away your differentiation to a vendor who serves your competitors too. Own your moat: the thing that actually makes your business different. If your edge is proprietary logic, a unique data pipeline, an AI agent that does something competitors cannot replicate, that cannot live on rented infrastructure controlled by someone else's roadmap. You need custom SaaS development here, full stop, because the day your moat depends on a vendor's pricing decision is the day you stop having a moat. Most Swedish companies I talk to have this completely backwards. They rent their moat through some vertical SaaS tool that every competitor also uses, and they build custom internal ticketing systems nobody needed custom. Flip it.

Sweden vs the World: Who Actually Gets This

From Jönköping, the contrast is sharp. In the US, venture-funded startups have been aggressive about owning core IP for years because investors demand defensibility, not a stack of Zapier workflows. Silicon Valley companies build their moat in-house almost by reflex, then rent everything commodity around it. Sweden has been slower here, and I think it comes down to risk aversion baked into how Swedish boards think about IT spend. Computer Sweden's reporting on rising cloud spend shows companies treating SaaS subscriptions as a safe, board-approvable line item, while building custom feels risky, expensive, and hard to budget. That was true in 2018. It is an outdated fear in 2026 when AI-assisted development timelines have collapsed. Asia tells a different story again. Chinese and Korean firms have been building proprietary software stacks aggressively, partly because of data sovereignty concerns and partly because labor cost advantages are shrinking there too. The EU, and Sweden specifically, sits in a strange middle position: strong engineering talent, strict data protection rules that actually favor owning your infrastructure rather than shipping data to a US SaaS vendor, and yet a corporate culture still defaulting to "buy the platform" out of habit. GDPR and the EU AI Act, if anything, are quietly pushing the calculus toward "own it." If you build your own AI agent stack, you control exactly what data goes where. If you rent a US SaaS platform with an AI feature bolted on, you are one terms-of-service update away from a compliance headache. Swedish companies should see this as an advantage, not a burden. Owning your stack is becoming the compliant option, not the expensive one.

Where This Goes: AGI and the Death of "Buy It Because It's Easier"

Push this forward two to five years. As AI agents get closer to handling full engineering workflows autonomously, the marginal cost of building custom software keeps dropping toward zero for anything well-specified. That does not mean SaaS dies. It means SaaS has to justify itself on something other than "building this yourself is too expensive," because that excuse stops being true. What survives on the rent side: genuinely hard infrastructure problems, network effects, and regulatory moats like payment rails or compliance certifications that take years to build regardless of how good your AI tooling is. What dies on the rent side: mid-tier vertical SaaS that exists purely because building custom used to be a six-month, six-figure project. That premise is already false today and gets more false every quarter. The regulatory gap nobody in Sweden is talking about enough: AI agents writing production code raise liability questions the EU AI Act barely touches. If your AI agent, running through something like OpenHands or a Claude Code harness, ships a bug that costs a customer money, who is liable? The company that owns the code, presumably, which is one more reason ownership matters more than renting. When you rent, liability gets fuzzy and contractual. When you own, you know exactly where the buck stops, and you built the guardrails yourself. I expect the next wave of PE activity in the Nordics to specifically hunt for companies with owned, defensible AI-native software, the way CapMan just proved out. Rented SaaS stacks with no proprietary layer will get valued like commodity businesses, because that is what they are.

What to Look At This Week

If you are rethinking your build vs rent vs own strategy right now, here is where I would actually spend time: Graphify: turns your existing codebase, docs, and schemas into a queryable knowledge graph with deterministic parsing, no vector store nonsense. If you are deciding whether to rebuild something in-house versus keep renting it, understanding what you already have matters more than most teams admit. awesome-claude-code: a curated list of the tooling ecosystem around Claude Code. If your team is still coding like it's 2023, this is the fastest way to see what "10x cheaper to build" actually looks like in practice. rtk: a Rust CLI proxy cutting LLM token consumption 60 to 90 percent on common dev commands. Cost control matters even when building is cheap, especially at Swedish scale where every krona of infra spend gets scrutinized. Run an honest audit against these three buckets. Every subscription your company pays for, ask: is this commodity infrastructure I should keep renting, or is this quietly become my competitive edge while I pretend it is just a tool? If it's the latter, that's your candidate for owning outright, and that's exactly the kind of build we take on through AI solutions and AI agent development at HEIMLANDR.

What You Should Actually Do

Stop asking "build or buy." Ask three sharper questions instead. One: what am I renting right now that is actually my moat? If the answer is anything close to "our core product logic," fix that immediately, because you do not own your own business. Two: what am I building custom that is not my moat? Kill that. Rent it, use an off-the-shelf tool, stop wasting engineering time on internal tooling that adds no differentiation. Three: for whatever remains, is offshoring actually cheaper anymore once you account for AI-assisted local development? Run the real numbers before renewing that contract. The offshoring rate data this week suggests you are probably paying old prices for work that should cost less now, wherever it gets done. This is not abstract for us. Every SaaS development project we take on at HEIMLANDR starts with this exact conversation before a single line of code gets written. Most clients come in thinking they need custom software. Half the time they need to rent smarter instead. The other half, they have been renting their moat for years and did not realize it.

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.