Hetzner raised prices on its dedicated server lineup effective June 15, 2026, the third pricing change the German host has made this year, and the underlying cause is a sharp rise in memory and storage component costs tied to AI demand. Unlike the broad April round that hit existing customers, this change applies to new orders and rescales only, with existing rented servers keeping their current terms.

What exactly is changing at Hetzner?

The 2026 timeline reads as steady escalation. Hetzner announced a price adjustment in February that took effect April 1, raising cloud server prices in Germany and Finland by roughly 30 to 37 percent for both new and existing customers. In April it raised setup fees for dedicated servers, citing rising prices for memory and fast storage. The June 15 round standardizes the dedicated lineup into numbered tiers, introduces a new limited hardware tier, and increases monthly prices, but only for new orders, rescales, and future products. To give a sense of scale from the earlier round, the popular AX42 dedicated server moved from 47.30 euros to 57.30 euros per month. One softening detail is that monthly rates rose while one time setup fees fell, which rewards longer tenancy over short lived provisioning.

Why are server prices rising across the board?

The real driver is memory pricing. According to TrendForce, server DRAM contract prices rose 43 to 48 percent in the fourth quarter of 2025 alone, with quarterly increases projected above 60 percent in the first quarter of 2026 and 58 to 63 percent in the second. AI demand for memory, fast storage, and GPUs has disrupted supply chains for components that used to be predictable and cheap. This is, in effect, a hidden tax of the AI boom, and it is landing on everyone who runs a server, not just the companies training frontier models. Hetzner is simply the most visible host passing it through, because it competes on price and has the least room to absorb the hit quietly.

What should you do if you host on Hetzner?

Do not panic migrate. The smart move is to get precise about exposure rather than reacting to the headline. Map which of your servers are protected under existing terms, and identify the workloads that actually create or rescale capacity often, such as continuous integration runner fleets that spin up new machines every week. Those are the ones exposed to new pricing. Competitors at the lower end like Contabo have not matched the increases as of this writing, so there is room to shop, but a rushed migration can turn a pricing update into an outage. Treat this as a budgeting exercise first and a migration decision second.

Check out what else is trending at Hacker News Trending