We are excited to announce that Zeabur has closed a $2M Seed Round to build the world’s first AI DevOps Engineer—a platform that helps developers automate deployment, scaling, and infrastructure management.

This round was led by 500 Global in Silicon Valley, with participation from Baidu Ventures, Taiwan’s AVA Angels, and our early backer MiraclePlus.

Coincidentally, exactly one year ago on November 11, 2024, Zeabur celebrated its first milestone: surpassing 1,000 individual paying users.

Today, on this symbolic date, Zeabur now serves over 5,000 individual paying customers, with a 500% year-over-year growth rate and a steady 10%–15% month-over-month growth that continues to accelerate.

Our team is currently in Silicon Valley, learning how AI startups validate products, markets, and user needs at breakneck speed.

In the following sections, we’ll share photos and stories from this journey.

How Did We Get Here?

Two years ago, I was still at Zhejiang University, struggling with my thesis while juggling outsourced development projects that all needed deployment. Managing backend infrastructure felt frustrating and unintuitive.

On the frontend, platforms like Vercel offer one-click deployment, but for the backend you still have to wrestle with AWS, GCP, networking, DNS, firewalls, endless YAML configs, and CLI scripts.

So I told my thesis advisor that I wanted to build a “Vercel for the backend.” After shipping the MVP, we joined the MiraclePlus accelerator.

Over time, I noticed a recurring debate online: “Should I use a PaaS platform like Vercel, or rent a VPS from providers such as Hetzner or Linode to self-host?”

That question stuck with me. I began to wonder: could there be a simple, low-friction way for developers to deploy no matter which infrastructure they use, with the same ease as Vercel?

So a year ago, we launched the Dedicated Server feature. It transformed Zeabur from a pure PaaS platform into a DevOps service that fully supports BYOS (Bring Your Own Server).

But it’s more than just “bringing” a server—it’s about “buying your own server.” Developers can purchase infrastructure from major cloud and VPS providers—AWS, GCP, Linode, Hetzner, DigitalOcean—directly through Zeabur.

We also observed how developers’ workflows are changing. More people are using Cursor or Claude Code to write software. That begged the question: when coding ends, shouldn’t an AI partner also handle “deployment and operations”?

So earlier this year, we launched the Zeabur Agent—letting developers use natural language to collaborate with AI to accomplish tasks that used to be tedious and complex.

From now on, you can manage your infrastructure with Zeabur the same way you code with Cursor.

Whether you’re running on GCP, AWS, DigitalOcean, Linode, Hetzner, or even your company’s internal servers, Zeabur has you covered.

Where We’re Headed Next

I’d like to share our roadmap in two parts: infrastructure and product.

Infrastructure

We deeply understand how critical DevOps is to any product, team, or company. As an AI DevOps Engineer, our top priority from day one has been ensuring every deployment plan is better than the last, and that every incident can be diagnosed and resolved with minimal effort and downtime.

Over the past two years, we’ve weathered several incidents, which impacted some of our earliest supporters. I want to extend my sincerest apologies and thank those who gave this young startup a chance.

Our next step is to honor that trust. We are expanding our engineering and customer support teams to elevate system stability and security. Even when problems occur, you’ll be able to reach experts immediately.

Product

Zeabur has pivoted several times, and our latest—and most significant—shift is the Zeabur Agent. Our goal is to refine the experience until managing infrastructure feels as smooth as building software in Cursor.

We’re not just focused on the Agent interface; we’re also expanding the Agent’s capabilities. Today it can deploy services, analyze logs, and monitor resource utilization. In the future, we want it to configure and optimize databases and cloud servers. Designing a process that is fully controlled, safe, and trustworthy is non-negotiable for us.

Beyond the Agent, we’re constantly exploring other places in the DevOps lifecycle where Zeabur can help builders streamline their work—like the upcoming Zeabur AI Hub, which we believe will make AI-era product development even more accessible. Stay tuned!

Coming Soon: Our First Office

After this trip to Silicon Valley, we’ll open Zeabur’s first office in Taipei by the end of the year. We want our team to collaborate more closely and iterate faster on our product and services. Once the office is ready, we’ll publish another article (or video) to show you around!

We also plan to host workshops and hackathons of all sizes at our Taipei office. We hope you’ll come hang out with us soon.