AI coding startup Emergent has shared new details about the technology running behind its platform as usage continues to climb.
In a recent blog post, the company said its system is now capable of running more than 30,000 development environments at the same time, allowing AI agents to build and deploy software for users automatically.
The platform works by creating temporary development environments whenever a user asks the system to build something. These environments run on Kubernetes, and each one operates inside a full Linux setup so the AI agent can write code, install packages, run tests, and launch applications.
From the user’s perspective, the process is straightforward. A person simply describes the type of product they want to create — for example a SaaS dashboard, a REST API, or a data pipeline. After that, the AI agent takes over most of the development work.
The system creates a dedicated workspace for the request. Inside that environment, the agent plans the steps, writes the code, installs dependencies, runs builds, and starts application services before deploying a working version of the product.
According to the company, the aim is to shorten the path from idea to working software. Tasks that might normally take days can potentially be completed in minutes.
As more developers started using the platform, the underlying infrastructure had to scale quickly. Emergent said it initially handled only a few hundred environments at once, but demand soon pushed that number past 30,000 active workspaces.
Managing that scale required careful design. The system needed to start environments quickly, keep user data safe, and free up computing resources smoothly once a session ends.
While building the platform, the company explored several options for running development environments, including virtual machines and lightweight sandbox systems.
Virtual machines provide full operating systems but often take 30 to 120 seconds to start. Sandboxes launch faster but may not support system-level packages, databases, or background services that many applications rely on.
Because of these limitations, Emergent decided to use Kubernetes pods as the base for its environments. This setup allows each workspace to function like a full Linux system where the AI agent can install software, run services, and preview applications through secure links.
The company also designed a lifecycle system for these environments. When a session begins, the platform restores the user’s previous state from backups. During the session, the AI agent continues building and testing the application. Before the environment shuts down, incremental changes are saved so work can be restored later if needed.
Emergent said running its own infrastructure also helps maintain a quick feedback loop between the AI agents and the development environment. At the same time, Kubernetes features such as scheduling, bin-packing, and autoscaling allow the platform to manage thousands of environments efficiently while keeping them isolated from each other.
The company plans to continue improving the system, with future work focused on features such as pre-warmed development environments, multi-cluster orchestration, and improved methods for handling persistent data.
Emergent has also grown quickly as a business. The startup recently reached $100 million in annual recurring revenue (ARR) just eight months after launching.
Earlier in January 2026, the company raised $70 million in a Series B funding round led by Khosla Ventures and SoftBank Vision Fund 2. Other investors included Prosus, Lightspeed Venture Partners, Together AI, and Y Combinator, bringing the company’s total funding so far to $100 million.
Also Read: Anthropic Accuses Chinese AI Firms of Mining Claude Amid U.S. AI Chip Export Debate
