Mews cut 15% of its staff and called the roles obsolete.

In July 2026 Mews cut around 15% of its staff, saying AI had made the roles obsolete.

Last Tuesday Mews cut around 15% of its staff, roughly 170 roles, six months after raising $300 million partly to build agentic AI. Founder Richard Valtr didn’t blame performance or the market. He said AI has made the roles obsolete, jobs “built for an era that is ceasing to exist.”

The layoffs are the headline but the strategy underneath is the interesting part. Mews isn’t just trimming to run leaner. Valtr says the plan is to move beyond selling software and become a service provider, with AI absorbing work hotels currently do themselves. Revenue management. Procurement. Financial services. The restructuring is the demonstration of that. If one person with AI can do what used to take a team inside Mews, the pitch is that the same applies inside your hotel, with Mews running the system.

In May I wrote about the Sequoia claim that the next trillion-dollar company won’t sell software, it will sell the work. At the time it was a thesis. It’s now the published strategy of a $2.5 billion vendor serving more than 15,000 hotels.

It’s also the same consolidation pattern I’ve been writing about for the last few months, running in a different direction. DoorDash consolidated the channels: delivery, reservations and guest data in one stack. Mews is consolidating the functions: more of the hotel’s actual work inside one vendor’s platform. Either way the result for the operator is the same. More of your business flows through fewer companies.

That changes what dependence means. Switching software is painful but it passes. You migrate, you retrain, you move on. If your vendor is running your revenue management and your procurement, there’s less of your own operation left to migrate. The knowledge of how the work gets done lives with them, not with you.

Mews isn’t alone. Expedia, American Express Global Business Travel and Oracle have made similar cuts on the same logic. And nothing about this is hotel specific. Restaurant tech vendors will follow the same playbook, and the ones that consolidated first are best placed to run it.

None of this means the model is bad for operators. Handing off work that AI can now do, and that was never the reason guests chose you, can be a good trade. But it’s a different trade than buying software, and it deserves different questions. Not just what does it cost and what does it do, but what do we stop knowing how to do ourselves, and what does leaving look like in five years. Better to ask those before the pitch arrives. Because it’s coming.