Every major paradigm shift creates the need for new, foundational infrastructure. The move to the cloud gave rise to companies like Snowflake and Datadog. The shift to microservices spawned Kubernetes for orchestration. Now, agentic AI is triggering the next rupture of foundational technology. These autonomous systems that can plan, reason and execute complex tasks on our behalf present a new infrastructure challenge: how do you make these systems actually work in production?

Enter Temporal. Founded in 2019, Samar and the founding team saw early on that there would be a need for a durable, reliable execution layer as AI systems become more complex, autonomous and embedded in how businesses operate. Agentic AI’s rise happened what seems like overnight and is evolving fast, which means backend technology has to move even faster. Temporal is providing the backbone powering today’s AI agents, an area we see continuing to explode. That’s why we couldn’t be more excited to back Temporal’s $300M Series D.

The Reliability Problem Holding Back AI Agents

We believe the promise of agentic AI is enormous – as is the resulting reliability problem. BCG estimates the global AI agents market is roughly $12B today, projected to eclipse $50B by 2030. Every day, agents are becoming increasingly complex and longer-running. Instead of simply answering questions, AI agents are now able to plan multi-step workflows, call external tools and APIs, make decisions and take action without constant human supervision.

The catch here is most agentic AI efforts stall at the prototype stage. Impressive demos crumble when they encounter the messy realities of production, such as network failures, cloud outages, API timeouts and long running processes that need to run for hours, days, or even weeks. A single failure at any step in an agent’s workflow can wipe out the entire execution chain, wasting time and compute resources, disrupting the customer’s experience and ultimately impacting the bottom line. The models themselves are increasingly capable, but what’s been missing is the infrastructure to make them reliable.

The Durable Execution Layer for AI Agents

Temporal is an open-source durable execution platform that was built to solve reliability head-on. At its core, Temporal lets developers write workflows in the programming language of their choice, and then guarantees that those workflows will run to completion, regardless of what goes wrong along the way and why. If a cloud instance goes down, an API call times out, or a process needs to pause and resume days later, Temporal picks up exactly where it left off. Every decision, step and result is durably recorded, so execution can resume precisely from the point of failure rather than starting over from scratch.

We feel the results speak for themselves. Over the past year, Temporal has reported over 400% YoY revenue growth, a 350% increase in weekly active usage and a 500% increase in installations, which now exceeds 25 million installs per month and 9.1 trillion lifetime action executions on Temporal Cloud alone. 

What’s more is the platform has also proven its mettle in real-world stress tests. During a major AWS outage in October 2025, customers running on Temporal’s high-availability architecture continued operating without data loss or manual intervention. In another instance, Temporal handled sudden traffic spikes exceeding 150,000 actions per second with no advance notice.

Powering the Most Advanced AI Systems in the World

What also stands out to us is who is building on Temporal. The platform powers some of the world’s most advanced agentic applications, spanning AI research labs, AI-native startups and global enterprises alike. 

For example, OpenAI runs its agentic workflows (including complex, multi-step research and data retrieval pipelines) on Temporal. AI-native companies like Replit and Lovable rely on Temporal to build agents that are reliable in production at massive scale. And that adoption extends well beyond the AI-native world. Financial services companies like JPMC and Block use Temporal to accelerate developer productivity, Yum! Brands (Taco Bell, KFC) build on the platform across their restaurant tech stack, healthcare companies like Abridge use it to power ambient AI across over 200 health systems, and the Washington Post uses Temporal to run its AI-powered video pipeline. And these are only the highlights.

In the new paradigm we now live in, Agentic AI doesn’t fail to subsist because the models aren’t good enough, but rather breaks because the systems around them can’t handle real-world execution. Temporal’s growing ecosystem of partnerships and integrations, including companies like OpenAI, Pydantic and Vercel, makes it increasingly easy for teams to move from experimentation to production without re-architecting their systems. The company is also investing in a forward-looking R&D roadmap spanning serverless execution, durable application communication (Temporal Nexus), and other features designed to reduce the operational burden of running AI at scale.

A World-Class Founding Team

It was only after seeing the explosive pull from the broader market that they decided to fork Cadence and launch Temporal as a standalone company in 2019. We believe that depth of experience and conviction is rare, and it shows in the product, the community and the caliber of customers they’ve attracted.

At Sapphire Ventures, we pride ourselves on partnering with Companies of Consequence – category-defining businesses building foundational infrastructure at the heart of major technology shifts. We’ve had the privilege of backing companies across the AI / infrastructure stack like LangChain, Glean, Weights & Biases and many others at pivotal moments in their growth. We’re thrilled to add Temporal to that list.

If you’re interested in learning more about Temporal, you can find more information here. And if you’re interested in joining their growing team, check out openings here.