Temporal in talks for $500M round at $12B, six months after $300M raise
What's the deal? Temporal Technologies is in talks to raise about $500 million at a pre-money valuation of at least $12 billion, Bloomberg reported on August 18. The financing has not closed, and its amount and terms could change.
Why now? The proposed mark is at least 2.4 times the $5 billion valuation Temporal set in February, when it raised a $300 million Series D led by Andreessen Horowitz. Lightspeed Venture Partners and Sapphire Ventures joined that round, alongside existing backers including Sequoia CapitalDealroom has a profile for this one. Try Dealroom →, Index Ventures, Tiger Global, GIC, Madrona, and Amplify Partners.
What's the endgame? Founded in 2019 by Samar Abbas and Maxim Fateev, Temporal sells a durable-execution platform that lets developers write multi-step workflows in ordinary programming languages. When a service or network fails, the workflow can resume from its last recorded point instead of starting over.
The company releases its core platform under the MIT license and sells Temporal Cloud as a managed service. It does not train AI models; it is betting companies will pay for a common execution layer beneath agents, much as they pay for databases and cloud infrastructure.
By the numbers: Temporal's proposed round follows a rapid sequence. A $146 million Series C in March 2025 valued it at $1.72 billion post-money; the February Series D lifted publicly announced primary financing to at least $650 million. A completed $500 million round would take that total to roughly $1.15 billion. In its March 2025 update, Temporal reported more than 2,500 Cloud customers, 183,000 weekly active open-source developers, and 4.4 times revenue growth over 18 months.
The customers: Temporal's biggest customer is OpenAI, according to people familiar with the matter. It also works with startups including AI coding company Replit and health-tech maker Abridge.
The signal: Orchestration platforms have grown more valuable as agentic AI takes off, since agents make model calls, use tools, and can run for hours or days — each step a place where state can vanish or an unsafe retry can repeat an action. Abbas and Fateev spent close to two decades on the unglamorous problem of keeping distributed software running; the agent boom has turned that old infrastructure work into one of venture's more expensive new categories.
Read more: runtimewire.com, Bloomberg
Image credit: Temporal
