AI Newsway

Temporal Raises $550M as AI Agents Start Running for Weeks

The durable execution company processed 1.9 trillion billable actions in August alone, up more than 350% year over year.

|5 min read0
AI Summary
Temporal Technologies raised a $550 million Series E at a $12.55 billion valuation on September 14, co-led by Lightspeed, more than doubling the $5 billion price it set in February. The durable execution platform processed 1.9 trillion billable actions in August, up over 350% year over year, with 4,300-plus paying customers including OpenAI, Snap and JPMorgan Chase. The raise signals that reliability infrastructure, not model access, is becoming the bottleneck for long-running AI agents.
Temporal's Durable Execution platform keeps long-running workflows alive across server failures in modern data centers.
Temporal's Durable Execution platform keeps long-running workflows alive across server failures in modern data centers.

Temporal Technologies said on September 14 that it raised a $550 million Series E at a $12.55 billion valuation, co-led by Lightspeed, in a round the company frames as a bet on software that has to keep running long after the request that started it is gone.

The Seattle company sells what it calls Durable Execution: developers write ordinary application code, and Temporal persists the state of that code across crashes, restarts and outages, so a process can pause for days waiting on a human approval and resume exactly where it stopped. That guarantee was originally pitched at payments, onboarding and order fulfillment. In its funding announcement, Temporal says customers now ask it to hold agent workloads that run for days, weeks or months.

Key takeaways

  • Temporal raised a $550 million Series E at a $12.55 billion valuation, co-led by Lightspeed with Wellington Management, Goldman Sachs Alternatives and Tiger Global.
  • The platform processed 1.9 trillion billable actions in August alone, up more than 350% year over year, across more than 4,300 paying customers.
  • The new valuation is 2.5 times the $5 billion mark set by February's $300 million Series D, taking total capital raised to roughly $1.2 billion.

Who is in the round

Lightspeed co-led the Series E alongside Wellington Management, Growth Equity at Goldman Sachs Alternatives and Tiger Global, with participation from T. Rowe Price and SV Angel. Returning investors include Andreessen Horowitz, Sequoia, Index, GIC, Sapphire Ventures and Amplify.

Andreessen Horowitz led the $300 million Series D that valued Temporal at $5 billion in February. Doubling that figure in roughly seven months puts the company in the same repricing band as the agent-infrastructure startups investors have been chasing all year, including River AI's $1.1 billion round.

The numbers behind the raise

Temporal's disclosures are unusually specific for a private company. Annualized revenue run rate is up more than 200% year over year, and net dollar retention has stayed above 200% since February — a level that implies existing accounts alone are more than doubling their spend. Paying customers passed 4,300, up 139% year over year, while open-source installations crossed 43 million, a 134% increase since December 2025.

Consumption grew faster than either figure. The platform handled 1.9 trillion billable actions in August, more than 350% above the same month a year earlier. Temporal points to Snap moving 414 million Stories a day through the system and JPMorgan Chase running it in regulated production. Its sharpest data point is OpenAI, whose usage the company says has grown 60-fold in under a year.

Why agents changed the pitch

Temporal's argument is that a working agent demo is now cheap and a trustworthy agent is not. Once a model starts touching the systems that move money rather than sitting on top of them, every retry, timeout and partial failure becomes a correctness problem rather than a logging problem — the same workflow orchestration issue the founders have chased for a decade.

Co-founders Samar Abbas and Maxim Fateev built Amazon's Simple Workflow Service, co-created Microsoft's open-source Durable Task Framework that became Azure Durable Functions, and later built Cadence at Uber before leaving in 2019 to start Temporal. OpenAI's VP of infrastructure, Venkat Venkataramani, said durable execution has become a core requirement for modern AI systems and cited it as a reason OpenAI built its own durable orchestration framework on Temporal.

Reliability has never been optional, but AI has quickly raised the cost of skipping it, Abbas said, describing the round as a bet that far more of the world's most demanding software will be built on that foundation.

What the money buys

The capital goes toward global operations on top of a team that doubled to 570 people over the past year, plus deeper investment in the platform's core primitives and the reliability and security work enterprise customers keep requesting. The open question is whether durable execution stays a distinct category or gets absorbed into the agent frameworks shipping from every major cloud — a race in which Temporal's answer is that it already runs the plumbing underneath several of them.

FAQ

What is Durable Execution?

Durable Execution is a model where a platform persists the full state of a running program, including its position in the code, so the work survives process crashes, machine failures and long waits. Developers write normal code instead of hand-building retry queues and state machines. Temporal popularized the term and ships it as an open-source platform plus a managed cloud.

How much has Temporal raised in total?

The $550 million Series E brings Temporal's total capital raised to roughly $1.2 billion. That follows a $300 million Series D led by Andreessen Horowitz in February 2026, which valued the company at $5 billion.

Who uses Temporal in production?

Temporal reports more than 4,300 paying customers, including OpenAI, Snap, NVIDIA and JPMorgan Chase. Snap runs 414 million Stories a day on the platform, and JPMorgan Chase uses it in regulated production environments.

How do you feel about this article?

SJ

Discussion

Sign in to post
Loading...

Related articles

Perplexity's New Agent Runs on Your Own GPU and Bills You Nothing to Do It
Tech & Business

Perplexity's New Agent Runs on Your Own GPU and Bills You Nothing to Do It

Perplexity and Nvidia shipped Portable Computer, an agent running Qwen or PPLX 27B on local RTX hardware with no metered tokens until the user allows a cloud step.

Seung Jung22 days ago
Huang Defends Nvidia's 70% Growth Call: 'We Put In $1 and $100 Comes Back'
Tech & Business

Huang Defends Nvidia's 70% Growth Call: 'We Put In $1 and $100 Comes Back'

Jensen Huang reaffirmed Nvidia's 70% growth guidance, implying $680 billion in revenue, and answered circular-deal critics with '$1 in, $100 back.'

Seung Jung6 days ago
NVIDIA Makes Tokens Per Megawatt the Metric, and Lambda Puts a Number on It
Tech & Business

NVIDIA Makes Tokens Per Megawatt the Metric, and Lambda Puts a Number on It

At AI Infra Summit, NVIDIA reframed AI infrastructure around tokens per megawatt, backed by Lambda's 24% throughput gain and new agentic benchmark results.

Seung Jung17 hours ago
OpenAI Says Its Own Models Helped Tape Out Jalapeño in Nine Months
Tech & Business

OpenAI Says Its Own Models Helped Tape Out Jalapeño in Nine Months

AI-generated kernels beat OpenAI expert-written versions by up to 1.8x, as Jalapeño posts its first InferenceX benchmark results.

Seung Jung22 days ago
Starcloud Raised $250M for Orbital AI Compute. Its Hardest Problem Is Booking a Rocket.
Tech & Business

Starcloud Raised $250M for Orbital AI Compute. Its Hardest Problem Is Booking a Rocket.

The Series A extension doubles Starcloud to a $2.3B valuation, but Falcon 9 winds down in 2028 and Starship has yet to prove rapid reuse.

Seung Jung25 days ago
Feds Warn AI-Written Exploits Are Already Hitting Critical Infrastructure Controllers
Tech & Business

Feds Warn AI-Written Exploits Are Already Hitting Critical Infrastructure Controllers

Five US agencies say attackers are using AI coding assistants and the open source snap7 library to build tools that read and write Siemens S7 controller memory.

Seung Jung28 days ago