DeepL Layoffs: Why Germany’s Translation Darling Just Cut a Quarter of Its Staff

On a Thursday, when most people are already half-checked out and scrolling through their phones in between meetings, the memo appeared on LinkedIn. The founder of DeepL, Jarek Kutylowski, described it as the hardest choice of his professional life. A quarter of the Cologne-based company, or about 250 employees, were being let go. The post had been shared, screenshotted, analyzed, and subtly contrasted with six other posts that had appeared in the same week by Friday morning. Reading these things now gives me the impression that they are beginning to rhyme.

DeepL used to be Google Translate’s courteous German counterpart. The company was founded by linguists and engineers who genuinely seemed to care about how a Polish sentence should sound when translated into French. Translators, attorneys, academics, and anybody else who ran the same paragraph through both engines and saw that DeepL simply got the rhythm right had a small but devoted following. It was founded in 2017, raised $300 million in 2024, reached a $2 billion valuation, and was openly considering an IPO in the United States. Then ChatGPT appeared. Claude as well. Gemini as well. Suddenly, the moat appeared shallower than anyone had publicly acknowledged.

Naturally, that isn’t exactly stated in Kutylowski’s memo. According to the report, AI is changing “what work exists, who does it, and how many people it takes to do it well,” and the company is going through a “massive structural shift.” According to the report, DeepL requires smaller teams, fewer levels of management, and a stronger dedication to becoming what the tech industry now refers to as AI-native. It claims that he is entering “founder mode,” a term that Paul Graham popularized a few years ago and that has since solidified into a corporate genre. The announcement even mentions a new San Francisco office, which seems more like a passport stamp than a calculated move.

The odd thing is that all of this sounds so familiar. Block performed a rendition of it. A version was created by Atlassian. Brian Armstrong of Coinbase has been discussing “one-person teams” and limiting management tiers beneath the C-suite. Similar remarks were made by Mark Zuckerberg during a Meta earnings call in January, almost as a shrug. The layoff letter is evolving into a literary genre, and DeepL’s contribution fits right in.

It brings up a somewhat cheeky question that Sam Altman raised earlier this year regarding “AI washing.” The idea is that because AI sounds forward-thinking rather than distressed, some businesses blame it for cuts they were going to make anyhow. It’s really unclear if that applies to DeepL. The business has been making significant investments in voice translation; in April, it launched DeepL Voice-to-Voice, partnered with AWS to handle data processing, and recently acquired a startup named Mixhalo. These are not the actions of a company that is winding down. These are the actions of a company attempting to run while switching shoes.

The DeepL thread on Reddit quickly filled up in the usual forums where laid-off workers congregate to compare notes. One commenter said they were in the middle of an interview with the company right now. Another engineer with thirty years of experience in the field wrote about seeing executives automate themselves into corners by chasing prototypes they didn’t understand. It’s the type of texture that is left out of the press release. Before the news, Glassdoor reviews had already noted “layoffs disguised as performance issues” and falling engagement scores.” It’s difficult to tell from the outside whether that’s true or just the bitter exhaust of any large workplace.

The larger pattern is more difficult to overlook. Over 1,100 were cut by Cloudflare in the same week. Recent examples include Shopify, PayPal, Spotify, and Amazon’s AWS retail division. The script continues to tighten. smaller groups. More AI. fewer layers. mode of the founder. Kutylowski may be correct in saying that businesses that take action now will shape the decade to come. It’s also possible that, once the prototype phase concludes and the actual engineering bills are due, we are witnessing the start of a lengthy and costly lesson in what AI can and cannot actually do. That distinction is unlikely to have much of an impact on the 250 individuals departing DeepL this month.