Seattle-built durable-execution & workflow-orchestration platform — spun out of Uber's Cadence project in 2019 by Samar Abbas (now CEO; previously architected Amazon Simple Workflow Service and Microsoft's Durable Task Framework) and Maxim Fateev (now CTO); $300M Series D at $5B valuation closed 17 February 2026 led by Andreessen Horowitz; the durable-execution backbone for OpenAI, Replit, Lovable, Stripe, HashiCorp, Snap and 2,500+ Temporal Cloud customers
Temporal is not a Zapier-style automation tool — it is a durable-execution platform for engineers who need to write long-running, stateful, retry-on-failure workflows in real code (Go / Java / TypeScript / Python / .NET / PHP / Ruby). It belongs in the conversation alongside AWS Step Functions, Cadence (which the founders also created), Airflow and Argo Workflows — not Zapier, Make, or n8n. The company was founded in 2019 in Seattle by Samar Abbas and Maxim Fateev, both veterans of the distributed-systems lineage that runs from Amazon Simple Workflow Service (Samar architected it from inception) and Microsoft's Durable Task Framework through Uber's Cadence project (the two of them co-created the open-source Cadence at Uber, which is the direct technical predecessor of Temporal). In 2024 the founders swapped roles — Samar Abbas became CEO and Maxim Fateev moved to CTO. Funding arc: $103M Series B in February 2022 at $1.5B+ valuation → $146M Series C on 31 March 2025 at $1.72B led by Tiger Global (total then $350M) → secondary round at $2.5B in October 2025 → $300M Series D on 17 February 2026 at a $5 billion valuation led by Andreessen Horowitz with Lightspeed, Sapphire, Sequoia, Index, Tiger, GIC, Madrona and Amplify participating. Total funding now approximately $650M. The Series D is explicitly framed around "making agentic AI real for companies" — OpenAI, Replit, Lovable, Abridge, Block and Washington Post are listed as Temporal customers using it as the durable-execution layer for AI agents in production. For Indian buyers the right framing is: Temporal is the right call for Series B+ engineering teams at Indian fintechs, payment platforms, e-commerce, AI startups and SaaS companies where workflows must survive crashes, restarts, deploys and weeks of duration without losing state. It is the wrong call for product managers, marketers, or anyone looking for a no-code Zapier replacement.
Temporal is a developer-facing platform that lets engineers write code (in Go, Java, TypeScript, Python, .NET, PHP or Ruby) where every function call, every retry, every wait, every signal, and every long-running state transition is durably executed — persisted to a database such that a server crash, a deploy, an AZ failure, or a multi-week sleep does not corrupt the workflow. The mental model: an order-fulfillment workflow that calls payment service → fraud check → inventory hold → shipment → tracking → refund-window, with a 30-day reconciliation timer at the end, is written as a regular function that just looks like normal code — the Temporal runtime handles the rest. The internal Cadence project at Uber (also created by the Temporal founders) is the predecessor that proved this works at production scale across thousands of services and billions of executions.
The company was founded in 2019 in Seattle by Samar Abbas and Maxim Fateev, who together have over 30 years of combined experience architecting durable workflow systems. Samar Abbas (CEO since the 2024 role-swap) worked on Amazon Simple Workflow Service from its inception, then led development of Microsoft Azure's Durable Task Framework, and finally co-created the open-source Cadence project at Uber — three of the most influential durable-workflow systems ever built. Maxim Fateev (now CTO) was the original Temporal CEO from 2019 to 2024 and brings the same pedigree from AWS and Uber. The role-swap reflects the company's transition from open-source-led growth to enterprise-product-led scale.
The funding trajectory has been one of the most striking in developer-infrastructure SaaS. Temporal raised seed-stage capital in 2020, then a $18.75M Series A from Amplify Partners and others (total then $25.5M). In February 2022 Index Ventures led a $103M Series B at over $1.5B valuation. On 31 March 2025, Tiger Global led a $146M Series C at a $1.72B valuation with Sequoia, Index, Amplify, MongoDB Ventures and others participating — bringing total raised to $350M. A secondary round in October 2025 valued the company at $2.5B. Then on 17 February 2026, Andreessen Horowitz led a $300M Series D at a $5 billion valuation, with Lightspeed, Sapphire, Sequoia, Index, Tiger, GIC, Madrona and Amplify all participating. Total funding stands at approximately $650M. The valuation has nearly tripled in under 12 months — driven primarily by Temporal's positioning as the durable-execution layer for production AI agents.
Customer base in 2026 reads like a survey of who is building serious infrastructure: Stripe, HashiCorp, Datadog, Doordash, Netflix, Snap, Coinbase, Airbyte on the foundational-infra side; OpenAI, Replit, Lovable, Abridge, Block on the AI-agent side; Nordstrom, ADP, The Washington Post on the enterprise side. Adoption metrics from the Series C disclosure: 183,000+ weekly active open-source developers, 7M+ unique Temporal clusters deployed globally, 2,500+ Temporal Cloud customers, revenue growth 4.4x in 18 months, and Net Dollar Retention of 184% — exceptional even for top-quartile infrastructure SaaS.
Write workflows as plain Go / Java / TypeScript / Python / .NET / PHP / Ruby functions. The Temporal runtime persists every function call, every retry, every wait, every signal. The workflow survives crashes, deploys, AZ failures, and weeks of sleep — without you writing checkpointing, retry logic, or state machines by hand.
Built-in exponential backoff, configurable retry policies per activity, automatic compensation patterns for sagas. The deepest production-grade retry surface in the workflow-orchestration category.
Workflows can sleep for days, weeks, months or years — useful for subscription renewals, refund windows, compliance hold periods, KYC re-verification cycles. The wait does not consume runtime; only resources when the workflow advances.
Send signals into a running workflow (e.g. "approval received from compliance team"), query its current state from outside (e.g. "what's the status of this refund?"), spawn child workflows for parallelisation. The primitives that make multi-step business logic readable.
Production agentic AI workflows need durable state for context windows, retry-on-failure for LLM calls, human-in-the-loop pauses, and tool-call orchestration. Temporal is now the durable-execution layer behind OpenAI, Replit, Lovable, Abridge and Block's agentic systems — the strategic narrative driving the $5B valuation.
Same engine, two deployment surfaces. Temporal Cloud is the managed SaaS (AWS / GCP / Azure regions). The open-source server is fully self-hostable (Postgres / MySQL / Cassandra backends), which matters for Indian fintechs / banks / NBFCs under RBI data-residency mandates.
Pricing correction: earlier versions of this page listed "from $0.00025/action" — that figure is roughly 5x the current rate after the 2025 pricing update. Live rates from docs.temporal.io/cloud/pricing and temporal.io/pricing:
For Indian fintech / payments / e-commerce teams running typical workflow volumes (100K-1M actions/month), Temporal Cloud Growth at $200/month converts to roughly ₹18,500/month all-in with 18% IGST — significantly cheaper than building equivalent durable-execution infrastructure in-house. At 50M+ actions/month, the volume-discounted Business / Enterprise tiers typically cluster around $30K-$120K/year for Indian Series B+ engineering organisations.
Temporal is the wrong call when: you're a product manager / marketer / non-engineer wanting Zapier-style point-and-click automation (use Zapier, Bardeen, or n8n); you're an enterprise needing screen-scraping RPA on legacy mainframes (use Automation Anywhere or UiPath); you're a single-developer side project that won't see real production traffic (Temporal is overkill, build it natively); or your team does not have at least one senior engineer comfortable with distributed-systems concepts (Temporal has a real learning curve — workflows are deterministic, idempotency matters, and the developer needs to actually understand why).