Infrastructure-as-code company Pulumi Corp. says it wants to accelerate cloud infrastructure management automation with the use of agentic artificial intelligence.

To do this, it has announced the launch of what it says is the cloud industry’s first “platform engineering AI agent,” which is designed to manage infrastructure on any public, private or hybrid cloud platform.

Pulumi Neo, as the new agent is called, is an AI cloud engineering assistant, built on the company’s flagship IaC technology, which is used by companies to simplify the provisioning and management of cloud computing infrastructure. Pulumi’s tools enable cloud engineering teams to use code to provision their cloud-based servers automatically, eliminating the need to adjust hundreds of different settings manually.

IaC has a lot of merit because it dramatically accelerates the process of preparing the infrastructure for cloud applications. Doing so manually is complicated and tedious, with engineers required to provision and configure each resource that the application will use. They also need to define security rules and set up the individual components of each app. All of these tasks can be simplified by using code.

With Pulumi Neo, the company is taking cloud infrastructure automation to the next level, providing teams with an AI agent that fully understands cloud dependencies, can execute changes, monitor outcomes and maintain compliance across the infrastructure lifecycle. It appears as an AI teammate in the Pulumi Cloud platform, and engineers can ask it to perform various different management tasks, with full confidence it will respect all security policies and governance settings, thanks to its enterprise-grade guardrails.

The company said it transforms IaC into a fully agentic workflow, enabling engineers to launch short- and long-running tasks with approvals, interactive guidance and a complete history of tasks performed. Users will be able to generate previews for everything Neo does, before it does it, to ensure that its work won’t create any unforeseen problems. It automatically understands each customer’s multicloud context, across providers including Amazon Web Services, Google Cloud and Microsoft Azure, as well as private cloud-based Kubernetes environments.

Cloud infrastructure bottlenecks

Pulumi founder and Chief Executive Joe Duffy said AI has almost completely transformed the way applications are built in less than a year. Thanks to generative AI coding tools such as GitHub Copilot, Cursor and Windsurf, developer productivity has accelerated rapidly. Given that, the platform engineering side needs a way to catch up.

Duffy points to a recent forecast from Gartner Inc., which says that by 2026, more than 80% of large software engineering organizations will employ specialized platform engineering teams to help keep up with application delivery, up from just 45% in 2022. But he says this is not enough, and that infrastructure bottlenecks are often the primary constraint limiting the velocity of new application deployments.

Holger Mueller of Constellation Research Inc. told SiliconANGLE that infrastructure platforms are getting more complicated by the month, while it’s becoming more difficult than ever for organizations to hire the necessary talent to manage them.

“Anything that can help platform engineers to automate infrastructure management is not only welcome, but nowadays likely essential for organizations to keep operating successfully,” the analyst said. “Pulumi is leading this change to platform engineering with Neo’s agentic capabilities, which are based on its strong infrastructure-as-code assets.”

“Pulumi Neo builds on Pulumi’s proven product foundation to ensure reliability and safety, and combines the best frontier models with cloud context, tools, and workflows, enabling it to automate entire mission-critical tasks,” Duffy said. “Things that used to take weeks can now confidently be done in minutes.”

Pulumi said Neo has made a huge splash with early adopters while in beta test, with customers reporting an ability to deliver 10 times as much infrastructure than before and deploy applications up to 75% faster. Moreover, it claims that Pulumi Neo’s automated governance has reduced policy violations by an average of 90%.

The logistics company Werner Enterprises Inc. was one of the first to get to grips with Pulumi Neo, and says it was able to reduce infrastructure provisioning times from an average of three days per application to just four hours.

“We have a high tolerance for AI capabilities and we’re actively looking for where we can integrate it without compromising our infrastructure governance and reliability requirements,” said Werner’s cloud architect lead Jason Harris.

Pulumi said Neo is being made available in preview to all of its customers today, and it will be free to use prior to general availability. Customers can access it via Pulumi Cloud or alternatively through the Pulumi Model Context Protocol Server, which is available in the AWS Marketplace and through AI coding assistants such asCursor and Cognition Inc.’s Devin.

Image: SiliconANGLE/Dreamina AI