Temporal raises $550M at a $12.55B valuation as AI workloads get longer

Why it matters

Temporal's $12.55 billion valuation and company-reported growth show how much capital is moving toward the infrastructure beneath AI agents. Long-running systems still need to preserve state, recover from failures and avoid restarting work when a model, tool or external service breaks.

Temporal, the durable-execution software company founded by Samar Abbas and Maxim Fateev, announced a $550 million Series E at a $12.55 billion valuation on September 14th. Lightspeed co-led the round, which Temporal says will fund global operations, core platform work, and additional reliability and security efforts.

Wellington Management, Growth Equity at Goldman Sachs Alternatives and Tiger Global joined the financing. T. Rowe Price and SV Angel participated, while returning investors included a16z, Sequoia, Index Ventures, GIC, Sapphire Ventures and Amplify. Temporal did not disclose the financing's legal closing date, investor ownership stakes or other terms.

The round follows Temporal's previous $300 million Series D, announced on February 17th at a $5 billion valuation. That financing was led by a16z and included Lightspeed and Sapphire Ventures, with participation from Sequoia, Index, Tiger Global, GIC, Madrona and Amplify.

Temporal reports growth across customers, usage and revenue

Temporal's Series E announcement supplied an unusually broad set of operating metrics alongside the new round. Temporal reported more than 4,300 paying customers, up 139% year over year, and said its annualized revenue run rate had increased more than 200%. Net dollar retention has remained above 200% since February, according to Temporal.

Temporal also said its platform processed 1.9 trillion billable actions in August, an increase of more than 350% from a year earlier. Open-source installs passed 43 million, up 134% since December 2025. Temporal's team doubled over the previous year to 570 people.

Those figures were reported by Temporal and were not independently audited in the supplied material. Temporal did not disclose revenue dollars, profitability or customer concentration, leaving investors with growth rates rather than the financial statements behind them.

Temporal named OpenAI, Snap, NVIDIA and JPMorgan Chase among its customers. In its Series E announcement, Temporal said Snap moves 414 million Stories per day on the platform, JPMorgan Chase runs Temporal in regulated production and OpenAI's usage grew 60-fold in less than a year. Those customer figures also come from Temporal.

Temporal said it will use the proceeds to expand its global operations, develop the core primitives beneath the platform, and increase reliability and security work requested by enterprise customers. The spending plan points to a familiar challenge for infrastructure companies selling into large organizations: usage can grow quickly, while security requirements and support obligations grow with it.

Durable execution gets an AI budget

Temporal's software preserves an application's state, retries failed operations and resumes long-running work after outages. Developers write workflows using language-specific software development kits, while Temporal coordinates execution across services and records enough state to continue after a process, dependency or network connection fails.

That model predates the current agent push. A payment process might pause while awaiting confirmation, and an onboarding workflow might wait days for a human approval. AI agents add model calls, tool use and third-party services that can time out, produce unexpected output or fail midway through a job. Temporal's pitch is that applications running for days, weeks or months need the same recovery machinery without each engineering team rebuilding it.

Temporal distributes an MIT-licensed server that customers can operate themselves and sells the managed Temporal Cloud. Temporal Cloud's signup offer includes $1,000 in credits, although the supplied material does not establish standard rates for production usage. The self-hosted option gives engineering teams control over deployment, while the managed service turns the open-source project into a commercial infrastructure business.

Temporal competes with established workflow systems including AWS Step Functions, AWS Simple Workflow Service and Azure Durable Functions. Uber Cadence, the open-source project that preceded Temporal, remains another option. Newer orchestration vendors include Restate, Trigger.dev, Hatchet, Inngest and Orkes, which is extending the Netflix-originated Conductor project into agent workflows.

Temporal's differentiators are its code-first workflow model, broad language support, MIT-licensed core and choice between self-hosting and a managed cloud. Temporal can support payment processing, order fulfillment, onboarding, durable ledgers and continuous-integration pipelines alongside AI agents, giving Temporal a market beyond projects tied directly to model adoption.

Abbas and Fateev built the predecessor inside Uber

Abbas and Fateev founded Temporal in 2019 after working on workflow and messaging infrastructure at Amazon, Microsoft and Uber. According to Temporal's account of their careers, Fateev led messaging infrastructure that helped lay the groundwork for Amazon Simple Queue Service and later led Simple Workflow Service. Abbas joined the team that brought Simple Workflow Service to the public in 2012.

At Microsoft, Abbas co-created the open-source Durable Task Framework, which became a foundation for Azure Durable Functions. The pair reunited at Uber in 2015 to build Cadence. That workflow engine expanded to more than 100 internal use cases within three years and was released as open-source software in 2017.

They left Uber two years later to build an independent platform around the same premise: application developers should not have to reconstruct state management, retries, timers and failure recovery for every distributed service. Abbas has described their work as a career-long effort to solve distributed-systems reliability, an origin that now gives Temporal a ready-made answer to one of the less glamorous problems in deploying AI agents.

The $12.55 billion valuation assigns a steep price to that answer. Temporal's reported customer growth and usage suggest durable execution is finding buyers across conventional backend systems and AI workloads. The Series E gives Abbas and Fateev another $550 million to establish their software as the reliability layer underneath both.