They're cutting while they're winning

DeepL's CEO told his staff the right time to restructure is "before you have to." Then he cut a quarter of them. Bloomberg reported that the AI translation company, profitable and eyeing an IPO, is simultaneously acquiring a team to build real-time voice translation, opening a San Francisco office, and eliminating 250 jobs. The hiring and the firing are happening in the same sentence because they're solving the same problem: reshaping the company for a future where AI handles the routine work and fewer humans steer the rest.

He's not alone. In a single week, four companies announced double-digit percentage workforce cuts while posting revenue growth. CNBC reported that Cloudflare cut 1,100 jobs, 20% of staff, despite Q1 revenue of $639.8M (up 34% year-on-year). CEO Matthew Prince said internal AI usage had surged 600% in three months and that smaller teams can now handle work that previously required entire departments. Benzinga reported Freshworks cut 500 while growing revenue 16%, with its CEO stating that over half the company's code is now AI-written. And in the week's sharpest irony, Upwork cut 24% of its own staff while AI-related work on the platform surged 40% to over $300M in gross services volume. A company built to connect employers to human talent, shrinking because the talent its customers want is increasingly artificial.

The conventional reading is that AI is eating jobs. That's true but insufficient. What makes this wave distinct is that it arrives during growth, not decline. These aren't companies in trouble. Cloudflare beat forecasts. Freshworks grew 16%. The cuts aren't emergency responses to falling demand. They're deliberate bets that the old headcount was solving a problem that no longer exists at the same scale.

Profitable and shrinking

There's a historical echo worth hearing. When lean manufacturing spread through Western factories in the 1980s and 1990s, the companies that cut deepest weren't failing. They were the ones disciplined enough to recognise that production lines carried structural overstaffing built up during decades of expansion. The profitable restructuring was harder to criticise than the crisis layoff, and harder to reverse.

The same logic is appearing now, and the language is identical across all four announcements: "AI-native," "high-agency teams," "smaller teams doing more." Every CEO is telling the same story: we're not cutting because things are bad; we're cutting because the work changed.

Markets aren't sure they believe it. Cloudflare dropped 18% on the announcement. Upwork fell 19.3%. Investors appear to be pricing in execution risk, not rewarding efficiency gains. They want to see the smaller teams actually deliver before they'll pay for the restructuring thesis.

For Upwork, the tension runs deeper than share price. The company sells access to human freelancers. Its fastest-growing category is AI work that replaces human freelancers. This is its third round of cuts in three years. At some point the platform either becomes an AI-work marketplace or it becomes the intermediary that automated itself out of relevance.

The pattern to watch isn't whether AI can replace the work. Four earnings reports just answered that. It's whether the companies that cut first will outperform those that didn't, or whether "AI-native" turns out to be this cycle's "pivot to video."

---