Temporal brings durable execution platform to AWS Marketplace

Temporal Technologies Inc. is making its open-source durable execution platform available through AWS Marketplace, extending its reach to enterprises trying to run AI agents with fewer production failures. The company also recently introduced Serverless Workers, a service that lets users run Temporal tooling on AWS Lambda, as agentic applications push more orchestration work into cloud infrastructure.

Temporal, founded in 2019, sells infrastructure for durable execution: keeping workflows running reliably across retries, failures, state changes and long-running processes. Preeti Somal, Temporal’s senior vice president of engineering, said in an interview with SiliconANGLE’s theCUBE that the platform is meant to absorb the operational work developers otherwise have to build around agents and AI applications.

Somal described Temporal’s role as handling error recovery, retries, state management, queue management, task dispatching and flow control. She said the company’s infrastructure is designed for applications that may see sudden demand spikes, with Temporal ensuring tasks are not lost as worker capacity catches up.

Temporal has not disclosed new revenue, valuation, customer count or pricing details tied to the AWS Marketplace availability. The business case is distribution: AWS Marketplace can shorten procurement for enterprises that already buy software through AWS accounts, which is material for infrastructure vendors selling into large engineering organizations.

Agent reliability becomes the selling point

The pitch lands in a market where enterprises are experimenting with multiple models, agent frameworks and workflow patterns, while still having to meet production expectations. Somal said companies can choose among many models and agents, but maintaining the supporting software over time remains difficult.

She also said that, depending on the study, more than 50% of developer time is spent on error-handling work. Temporal did not identify the study in the remarks, so that figure should be read as a general framing claim rather than a disclosed internal metric.

Temporal’s technical argument is that agentic applications need more than a model call and a prompt. Long-running agents have to track state, recover from interruptions, avoid losing tasks and continue workflows when dependencies fail. Those requirements are familiar to distributed systems teams, but they are becoming more visible as AI agents move from demos into internal and customer-facing software.

Somal said Temporal has scaled to 150,000 actions per second. The company did not provide further context in the interview on benchmark conditions, customer deployment specifics or whether that figure refers to sustained production usage or a platform capability test.

AWS integrations anchor the latest push

Temporal demonstrated its platform with AWS Strands, an open-source software development kit for building and running AI agents. Its Serverless Workers service adds support for running Temporal tools on AWS Lambda, AWS’s serverless compute service.

The AWS Marketplace listing gives Temporal a more direct route to AWS customers evaluating agent infrastructure. Somal said customers including Stripe, Netflix, Datadog and Snap use Temporal for agents and AI applications. Temporal did not disclose the size of those deployments or the commercial terms.

The move is routine in one sense: infrastructure vendors often use cloud marketplaces to reduce procurement friction. It is more pointed in the current AI cycle because agent frameworks are multiplying faster than many companies’ operational controls. Temporal is positioning durable execution as the layer enterprises need once agent workflows run long enough, and fail in enough ways, to require production-grade orchestration.

This story draws on original reporting from SiliconANGLE.