Quick answer
The tech layoffs in Europe are a reshuffle, not a collapse, and AI is the main reason for them, though not the only one. IT services will get cheaper, but unevenly: routine work by 10–25% over the next few years, while expert work holds its price. A steady trickle of the people let go will set up as contractors and small companies; a start-up wave like Finland’s after Nokia is possible, but less likely. The real casualty is the junior developer’s first job, along with hourly billing and per-seat software pricing.
Delfi’s editorial called the Pipedrive layoffs the start of a shock, and the figures justify the word. Pipedrive, one of Estonia’s unicorns, is letting go of 267 people in Estonia, about 70% of its staff here. Swedbank is cutting close to 500 jobs across the group, around a hundred of them in the Baltics. Telia is removing about 900 positions, roughly a hundred of them in Estonia. As Delfi summed up the explanations from the boardrooms, they come down to one sentence: AI has changed the work, so the structure has to change.
The pattern is global. Microsoft has offered a voluntary exit package to about 60,000 employees, more than a quarter of its workforce, according to press reports. Amazon has removed some 30,000 corporate roles in a year, months after its chief executive said generative AI would shrink its office workforce.
Are the Tech Layoffs Really Because of AI?
Not entirely. Several of these companies over-hired in the cheap-money years and are now trimming, and AI reads better in an announcement than costs; critics call it AI washing. Telia’s own plan, for one, talks about localisation and savings, not about algorithms. But I think the explanation is mostly true this time, for reasons set out below: the money is going into servers rather than salaries, and the jobs going first are the routine ones AI already does well. This is the first wave of AI layoffs, and I take the companies at their word, with a discount.
What Happens Next to Europe’s Tech Sector
Delfi asked what Estonia should do to adapt. I want to answer a different question, the one that matters to anyone who buys or sells IT work: what happens next to Europe’s tech sector, to the price of IT services and to the people who have just lost their jobs. I see three scenarios, and the table shows how I weigh them.
| 1. Hollowing out | 2. Cheaper IT services, fewer juniors | 3. The Nokia effect | |
|---|---|---|---|
| IT jobs | Cuts continue, hiring freezes at every level, talent leaves Europe | Senior demand recovers, junior hiring stays weak | Jobs move from a few large employers to many small companies |
| Price of IT services | Falls slowly: large vendors keep the gain as margin | Routine work 10–25% cheaper, expert work holds its price | Falls faster as small teams compete on price |
| New companies | Few: little capital, little confidence | A steady trickle: contractors, agencies, niche software | A wave, as in Finland after Nokia |
| How likely, in my view | 25% | 50% | 25% |
Scenario One: Hollowing Out
Will AI Replace Junior Developers?
At large companies, yes, and for the most part quietly: junior jobs are not so much cut as never created. The best data on AI and IT jobs is American, where payroll records are tracked at scale, and Europe tends to follow with a lag. Stanford’s tracking of US payroll records shows that employment of 22–25-year-olds in the jobs most exposed to AI stands 19% below where it would be had it kept pace with their less exposed peers, and the gap comes from hiring freezes, not from firing. Software development postings in the United States are still about a fifth below their pre-pandemic level, and roughly seven in ten of them are for senior people. A Harvard and INSEAD study of start-ups found that companies built around AI run teams 25% smaller, with 15% fewer entry-level staff.
Where That Line Leads
The pessimistic reading extends that line: the hiring freeze spreads up the ladder; Europe keeps the senior engineers it has for a decade, then runs out of them, because nobody trained the next cohort. IT services get cheaper only slowly, because the large vendors keep the productivity gain as margin, and the small suppliers who would undercut them lose the discretionary projects they live on. Capital for new companies stays thin, and the most ambitious founders leave for the United States. I give this one chance in four. The reason it is not higher is in the next scenario.
Scenario Two: Cheaper IT Services, Fewer Juniors
Will AI Make IT Services Cheaper?
Yes, and I put half the probability on this scenario because it is what the money is already doing. Gartner expects worldwide IT spending to grow by 14% in 2026 but spending on IT services by only 5%; the difference is going into data centres and software. Budgets are not shrinking. They are moving from people to machines.
The outsourcers feel it first, and they are saying so. Hexaware’s chief executive told Bloomberg that clients may demand fees 20–25% lower for routine work as AI speeds it up, and the large Indian IT groups report clients asking for the productivity gain back whenever a contract is renegotiated. That is how IT services become cheaper: a discount at each renewal, rather than a sudden crash. It is why I expect routine work to cost 10–25% less within a few years.
The discount is uneven, and this is the part the headlines miss. In METR’s trial, the best-known so far, experienced developers working on large codebases they knew well took 19% longer with early-2025 AI tools, and the same researchers expect the biggest gains for less experienced developers and in unfamiliar code. The tools have improved since and METR itself treats the figure as dated, but the direction holds: expert work holds its price, routine work does not, and the junior who used to do that work is the one not hired.
For Europe the twist is timing. Only 20% of EU companies with ten or more employees used AI in 2025; among large companies the figure was 55%. Pipedrive and Swedbank are the large companies moving first. The small and mid-sized firms that make up most of the European economy have barely started, and for them cheaper development means more software: the internal tool or integration that was never worth the quote now gets built. In this scenario the sector employs fewer people than at its peak and produces more, and that demand from smaller firms is where a steady trickle of those let go will find their first clients, as contractors, agencies and small software companies.
Scenario Three: The Nokia Effect
Do Laid-Off Developers Start Companies?
Some do, and Europe has seen it at scale before. When Nokia fell as a phone maker, its Bridge programme helped former employees set up about 400 companies in Finland, and a survey by Aalto University found some 90% of them still active; worldwide the count passed 1,000. I do not think it is a coincidence that Finland’s start-up boom followed.
Two conditions are better today than they were then. First, the cost of software development has collapsed. Y Combinator’s chief executive says companies in its recent batches reach millions in revenue with fewer than ten people, and in a quarter of one batch 95% of the code was written by AI. A team of three ex-Pipedrive engineers can now ship what used to need a department, and when many such teams compete, prices fall faster than in my base case. Second, Europe is, for once, building the legal vehicle. The Commission’s EU Inc proposal would let founders register a company recognised across the EU within 48 hours, for under €100 and with no minimum capital, and Brussels wants it agreed before the end of 2026. Estonia offers most of that already: a company registered online within a day, and no tax on the profit you reinvest.
What this scenario needs is for the people let go to be hungry in the productive sense: to treat severance as seed capital rather than a cushion. I give it one chance in four, because the Finnish evidence also shows the limits. Most Bridge companies stayed small: the Aalto survey expected the 400 to employ about 2,500 people between them. An engineer with a mortgage is also more likely to take a contract than to found a start-up. Even so, it is the only scenario that produces the next Pipedrive, and layoffs at the old one are exactly how such things begin.
Seven Predictions for IT Jobs and Prices, With Dates
Scenarios are a way of being right whatever happens, so here are seven predictions specific enough to be wrong. Each comes with a horizon and the sign I will be watching.
| What I expect | When | The sign to watch |
|---|---|---|
| The junior developer disappears from large companies; a graduate’s first job is in a team of under ten, on a contract or in their own company. | Three to five years | Entry-level roles stay below 5% of software postings, where Indeed already has them. |
| Pay splits in two. Senior engineers get more expensive, because nobody is training their replacements, while the median developer salary in Europe falls in real terms. | Five years | Senior offers rising faster than inflation, junior offers below it. |
| The cuts run for another 18–24 months, then hiring resumes, but the sector settles with 10–20% fewer people than at its peak and a different mix: more AI engineers and generalists, fewer managers. | Two years | Software postings climb back towards their pre-pandemic level without reaching it. |
| Billing by the hour dies: most IT service contracts are priced per result, and agencies that sell hours lose to fixed prices. | Five years | The large outsourcers reporting most of their revenue from outcome-based contracts. |
| Software gets cheaper, not just the work. Per-seat pricing collapses as AI agents replace the users who held the seats, and CRM-type subscriptions lose 20–30% of revenue per customer, a discount of the same order as the one the outsourcers are now giving. | Three to five years | The big subscription-software vendors moving to usage-based prices and reporting falling revenue per customer. |
| The offshore model of bodies by the hour ends. India and the nearshore hubs of Eastern Europe lose their price advantage on routine work; Europe buys expertise from abroad, not capacity. | Five years | IT services growth stuck in low single digits (Allianz already has it at 3.6%) while total IT spending grows by double digits. |
| Estonia’s ICT sector has passed its headcount peak: 10–20% fewer people, the same share of GDP, and the tens of thousands of new tech jobs Delfi mentions appearing in other sectors rather than this one. | Five years | Statistics Estonia showing ICT employment down while ICT value added holds. |
The fifth is the one I would underline. Delfi’s editorial opens with the image of AI lightning striking Pipedrive’s business model, and that is the right diagnosis: a sales tool sold per user is in trouble once the user is an agent. The companies that cut staff because AI changed their market are next in line for the discount: their customers will ask them the question the outsourcers are already being asked, why pay the old price for work a machine now does. That is why I expect the fall in prices to spread from services to software itself.
Two practical conclusions follow for those who buy and sell IT work. If you buy it, sign shorter contracts and ask for the AI discount at every renewal, because your competitors will. If you sell it, move to fixed prices before your clients move you, and sell the judgement that AI cannot yet replace rather than the hours it can.
Laid Off From a Tech Job: What to Do Next
Whichever scenario plays out, my advice to anyone who has just been laid off is the same. Clients who no longer hire juniors still buy work from contractors, and the simplest way to invoice that work is to start a company of your own. For a one-person set-up, our guide to working as a freelancer through an Estonian company covers registration, taxes and banking. If you are building a product, the guide to setting up an IT or SaaS company in Europe explains when Estonia is the right base and when it is not, and the overview of starting an AI company in Estonia covers the ecosystem and the rules. We register Estonian companies for founders from across Europe, including people who run them from abroad as e-residents; the company formation page lists what each package includes.
