Stop selling talent. Sell product.

The claim

For two decades Southeast Europe has exported hours. AI is making hours cheaper. The next decade belongs to whoever owns the product, so position the region on what it builds, not who it hires.

The consensus

Vestbee’s Q1 2026 CEE report says it plainly: ‘CEE in 2026 is best framed as a talent-arbitrage market.’ Engineering, founder ambition and execution capacity are world-class on a global benchmark, it argues, but the financing infrastructure to convert them into liquid value isn’t yet built locally.

The evidence

• Hours: Serbia exported €4.55B of ICT services in 2025, ten times its 2012 level. Its startups raised $43.9M, which is under €1 for every €100 exported.
• Product: 76.2% of the 170 Serbian startups in the Digital Serbia Initiative’s Startup Scanner 2026 are building an IP-rich product designed for international markets from the outset.
• Who wins: three of Vestbee’s 15 largest CEE rounds of 2026 went to SEE-founded product companies: Fonoa ($110M Series C), CyberSwarm ($50M Series A) and Alcatraz AI ($50M Series B). None sells hours.
• Pressure: a Polish software services executive told Tech.eu that ‘writing code is now relatively cheap’; the real work is verification, architecture, governance and validation.
• Category: vertical AI is ‘the one category in which the region’s smaller markets consistently produce fundable companies’ (Vestbee, Q2 2026).

Serbia exports hours. The largest SEE-founded rounds of 2026 went to companies that sell product.

The dissent: Danijela Vasiljev

"Services revenue is the cheapest runway a SEE founder can get, and it’s how many product teams here paid for version one. Sell product, yes, but don’t burn the bridge you’re standing on."

Corridor cross-read

The Monday move

Open your deck. Count the slides about what your engineers do, then the slides about what your customers pay for. Rewrite the first three around the second, and put one product metric (usage, retention or revenue per customer) on slide two.