Samar Abbas (@SamarAtTemporal) and Maxim Fateev (@mfateev)'s Temporal Technologies is in talks to raise around $500 million at a pre-money valuation of at least $12 billion, Bloomberg reported on August 18th. The financing has not closed, and its amount and terms could change. (bloomberg.com)
At the reported floor, $500 million would buy roughly 4% of Temporal after the financing. The proposed mark is at least 2.4 times the $5 billion valuation Temporal announced with its $300 million Series D on February 17th. That round was led by Andreessen Horowitz, with Lightspeed Venture Partners and Sapphire Ventures joining existing backers including Sequoia Capital, Index Ventures, Tiger Global, GIC, Madrona and Amplify Partners. (temporal.io)
The speed of that repricing is the story. Abbas and Fateev spent close to two decades working on the unglamorous problem of keeping distributed software running through crashes, failed services and network interruptions. AI agents have turned that old infrastructure problem into one of venture capital's more expensive new categories.
A two-decade systems bet meets the agent boom
Abbas and Fateev worked together on Amazon Simple Workflow Service before taking separate paths through Microsoft, Google and Uber. Abbas led development of Microsoft's Durable Task Framework. Fateev had led the architecture of Amazon Simple Workflow Service and the storage backend for Amazon Simple Queue Service. The pair reunited at Uber, where they co-created Cadence, the open-source workflow engine that became Temporal's technical predecessor. They founded Temporal in 2019 to build a commercial platform around the same durable-execution model. (temporal.io)
Temporal lets developers write multi-step workflows in ordinary programming languages while its infrastructure records execution history and application state. When a service, process or network connection fails, the workflow can retry work or resume from its last recorded point instead of starting again or relying on developers to stitch together queues, timers and recovery logic. Temporal releases its core platform under the MIT license and sells Temporal Cloud as a managed service. (docs.temporal.io)
That architecture predates the current agent cycle. Its AI relevance comes from the way agents operate: they make model calls, use tools, wait for external systems or human approval, and may run for minutes, hours or days. Each added step creates another place where state can disappear, an API can fail or an unsafe retry can repeat an action.
Abbas summarized Temporal's pitch in the February financing announcement: "It fails because the systems around them can't handle real-world execution." Temporal says its customers use the platform to maintain agent state, trace failures and keep long-running tasks moving through infrastructure interruptions. (temporal.io)
Temporal does not train AI models. It is betting that companies will pay for a common execution layer underneath models and agents, much as they pay for databases, observability tools and cloud infrastructure. That gives Temporal exposure to AI spending without requiring the capital costs of a frontier model lab.
The valuation is moving faster than the disclosed economics
Temporal's proposed round follows a rapid sequence of financings. On March 31st, 2025, Temporal announced a $146 million Series C at a $1.72 billion post-money valuation, led by Tiger Global. Temporal said that round brought total funding to $350 million. The $300 million Series D in February 2026 lifted publicly announced primary financing to at least $650 million. A completed $500 million round would take that total to roughly $1.15 billion. (temporal.io)
Temporal supplied strong growth percentages alongside those rounds. In March 2025, Temporal reported more than 2,500 Cloud customers, 183,000 weekly active open-source developers and 7 million unique cluster deployments. Temporal also reported 4.4 times revenue growth over 18 months and net dollar retention of 184%. (temporal.io)
By February 2026, Temporal said revenue had grown more than 380% year over year, monthly installations had passed 20 million and Temporal Cloud had processed 9.1 trillion lifetime actions. Those figures are Temporal's own and measure a mix of commercial growth and product usage. The percentage growth does not reveal the starting revenue base, while installations and action executions do not translate directly into recurring revenue. (temporal.io)
That distinction matters at a proposed $12 billion pre-money valuation. Investors considering the round are being asked to price the durability layer before Temporal has publicly attached a current annual revenue figure to it. The wager rests on Temporal becoming embedded deeply enough in production systems that usage expands with every new workflow and agent its customers deploy.
Durable execution is becoming a platform fight
Temporal's fundraising comes as durable workflows move from a specialist infrastructure category into major developer platforms. Cloudflare made Workflows generally available on April 7th, 2025, offering long-running execution inside its Workers platform. Restate opened its managed cloud on September 30th, 2025, targeting workflows and agents with usage-based pricing. Vercel made Workflows generally available on April 16th, 2026, tying durable TypeScript and Python functions directly to its deployment platform. (blog.cloudflare.com)
Those products compete from different starting points. Cloudflare and Vercel can bundle orchestration with compute, hosting and existing developer distribution. Restate and other focused vendors can compete on programming models, operating costs and deployment choices. Temporal's defense is the founders' long record in distributed systems, an open-source project that can be self-hosted, and a managed cloud business built around workloads that are costly to move once they become part of critical application logic.
The proposed financing would give Abbas and Fateev substantially more capital to defend that position. It would also set a demanding benchmark. A $12 billion pre-money valuation assumes durable execution becomes a standard component of the AI application stack and that Temporal captures a large share of the resulting commercial spend.
Abbas and Fateev began building workflow systems long before investors started putting "agentic" into funding announcements. Six months after Temporal's $5 billion round, that history is being priced as an AI infrastructure advantage. The round under discussion would turn the founders' patient systems bet into a $12 billion test of whether reliability becomes the layer every production agent eventually has to buy.