Existing services are not affected — Zeabur is focusing on being your AI DevOps Engineer
Today we're announcing an important decision: Zeabur's shared cluster will no longer accept new projects. If your services are already running on the shared cluster, they will not be affected in any way.
This is not a reactive decision. It's not driven by cost pressure or technical debt. It's a deliberate choice we made after seriously asking ourselves: "What should Zeabur actually be?"
For the past few years, Zeabur has been doing two things at once: providing compute resources (the shared cluster) and helping users manage and deploy services. These two things seem closely related, but they are fundamentally different businesses.
Providing compute resources means we are an infrastructure provider. We compete with AWS, GCP, Linode, and Hetzner on price, availability zones, and hardware specs. In that competition, we have no structural advantage — our servers run on top of these providers, so our pricing can never be lower than theirs.
But helping users manage and deploy services — that's what we're actually good at. Push your code and it deploys automatically. Set up databases and domains in a few clicks. Let AI handle the DevOps busywork. That's where Zeabur's real value lies.
So we made a choice: stop playing both roles, and go all in on what we do best.
Zeabur's positioning is Your AI DevOps Engineer. Our job is to be the abstraction layer between you and your infrastructure — helping you select the right compute resources, deploy and manage your services, and handle the tedious but critical DevOps work. The compute resources themselves should be left to specialized infrastructure providers.
Your instinct might be: "Moving from a shared cluster to my own server — won't that cost more?"
It's actually the opposite.
To maintain acceptable performance and stability on the shared cluster, we had to use AWS as the underlying infrastructure. Frankly, the current pay-as-you-go pricing on the shared cluster is below our cost — we've been subsidizing this service. Yet many users have noticed that even this subsidized price is still higher than VPS providers like Linode, DigitalOcean, or Hetzner.
This structural contradiction has no solution: when you provide infrastructure on top of someone else's infrastructure, you can never be price-competitive.
With Dedicated Servers, you choose your own provider. Want the best value? Go with Hetzner. Want top-tier performance and reliability? Choose AWS or GCP. The choice is entirely yours. And Zeabur focuses on what we're actually good at: giving you a consistent, simple deployment and management experience on any provider's machine.
This is not a zero-sum game. Users get lower prices and better service. Zeabur gets a sustainable business model. For the first time, our incentives are truly aligned.
You might worry that things will work differently after migrating. They won't.
On Zeabur, using a Dedicated Server is identical to using the shared cluster. Same interface, same deployment flow, same operations, zero learning curve. That's the whole point of Zeabur as an abstraction layer — no matter whose machine is underneath, you always interact with the same Zeabur.
And Dedicated Servers actually give you something the shared cluster never did: root SSH access to the entire machine. Full control over the underlying system, whenever you need it.
We're rolling this out in three phases:
| Date | Change |
|---|---|
| Feb 21, 2026 | Project creation defaults to Dedicated Server options; shared cluster still accessible via a toggle |
| Mar 15, 2026 | New project creation on the shared cluster is disabled |
| Apr 1, 2026 | New service creation in existing shared cluster projects is disabled |
Services already running on the shared cluster will not be affected. Even if you choose not to migrate, we will ensure these services continue to run safely and stably. This is our commitment to existing users, no exceptions.
We've built a "One-Click Copy" feature in the project settings page. It automatically transfers your services, environment variables, and disk data from the shared cluster to a Dedicated Server — no manual work required.
If you don't have a Dedicated Server yet, check out our latest pricing page.
A Dedicated Server is a single VPS, which covers most use cases. But if your business requires high availability and cluster-level architecture, we're also planning to make Dedicated Clusters available for self-serve purchase, expected to launch in mid to late March.
Dedicated Clusters were previously only delivered to enterprise clients through our sales team. We've decided to open this architecture to all users.
Technically, a Dedicated Cluster runs on managed Kubernetes services from major cloud providers (AWS EKS, GCP GKE, etc.), with Zeabur's full DevOps system — battle-tested in large-scale production — running on top. It's the exact same architecture as our current shared cluster, except the entire cluster is dedicated to you. Zeabur's shared cluster currently powers hundreds of nodes and tens of thousands of containers; its stability and security have been thoroughly proven. Soon, you'll be able to have this entire system exclusively for your enterprise workloads.
Going forward, plan subscriptions are primarily for accessing Zeabur services beyond the server itself — such as AI Agent, data backups, monitoring, CI/CD, and other additional services. You don't need a Developer Plan subscription to deploy services on a Dedicated Server, but subscribing gives you better tools to manage and operate your server through Zeabur.
Yes, services already running on the shared cluster still require an active Developer Plan subscription to continue operating. Additionally, we may adjust the shared cluster's usage-based pricing in the near future, and we are considering removing the $5 shared cluster usage credit currently included in the Developer Plan.
The two existing free clusters (Tencent Cloud Silicon Valley and BytePlus Jakarta) will be converted to paid shared clusters starting March 15, requiring a Developer Plan subscription to use. Their pricing and rules will be aligned with the other existing clusters.