Temporal Technologies Inc., a startup with a cloud platform that makes artificial intelligence agents more reliable, has closed a $300 million funding round.

Andreessen Horowitz led the Series D investment. Temporal stated in its announcement of the raise today that Lightspeed Venture Partners, Sapphire Ventures and several returning backers chipped in as well. The company is now valued at $5 billion.

An AI application can malfunction for a variety of reasons, including internal bugs and external factors such as an outage in a third-party service. As a result, developers must write code that enables the application to automatically recover in the event of a technical issue. The process can take a significant amount of time and effort.

San Francisco-based Temporal offers a cloud platform that eases the task. It’s based on an open-source project also called Temporal that is downloaded more than 20 million times per month. The company says its installed base includes OpenAI Group PBC, Nordstrom Inc. and other major brands.

The platform creates an activity log that contains information on every single task performed by an application. After an outage, the application can check the log to find the task that it was working on immediately before the malfunction. It can then pick up where it left off.

It also mitigates the impact of technical issues in other ways. If an external service that an AI agent uses to automate tasks experiences downtime, the platform can route the AI agent’s requests to a different service with similar features. Furthermore, the application activity log doubles as a debugging tool that developers can use to diagnose malfunctions.

Temporal makes money from its eponymous open-source project with a paid offering called Temporal Cloud. It’s a cloud-based version of Temporal that doesn’t require developers to manually set up the software or manage the underlying infrastructure. According to the company, its service also offers better performance than self-hosted environments.

Temporal Cloud includes an autoscaling feature that adjusts the amount of hardware available to a workload based on usage. According to the company, the mechanism can process more than 300,000 AI agent actions per second. If the data center that hosts a Temporal Cloud environment goes offline, the platform reroutes requests to a standalone environment in a different facility.

“Rather than create new problems, agentic AI tends to expose old ones such as managing state and failures,” said Temporal co-founder and Chief Executive Officer Samar Abbas. “We’ve been solving these same problems for years. Temporal exists to make agentic AI work in production as well as any other class of application, reliably, predictably and at scale.”

Temporal will use the proceeds from its funding round to enhance Temporal Cloud and the upstream open-source project. 

Image: Temporal