Restate raises $20M to build durable infrastructure for AI agents

Restate, a Berlin-based startup founded in 2022, has secured $20 million in Series A funding to expand its durable workflow infrastructure. The round was led by Singular, with participation from Redpoint Ventures and Capital One Ventures, marking a significant step as the company positions itself against larger competitors in the rapidly evolving AI infrastructure market.

What happened

Stephan Ewen co-founded Restate with the initial goal of creating an execution engine that could make multistep software workflows resilient to crashes and network interruptions. While the technology was not originally designed specifically for artificial intelligence, the rise of AI agents has created a sudden and powerful demand for such capabilities. AI agents often execute long-running, unpredictable tasks that require strict consistency and reproducibility, making durable execution critical. Ewen noted that the platform inadvertently became a perfect match for the complex problems surfaced by modern agentic systems.

The startup has seen rapid traction over the last few months, closing multiple customer contracts worth six and seven figures. This momentum attracted high-profile clients, including the vibe-coding platform Replit, as well as Fortune 500 companies in the financial sector. The new capital will allow Restate to compete more directly with Temporal, the current heavyweight in the durable infrastructure space. Temporal recently announced a massive $550 million Series E round at a valuation of $12.55 billion, highlighting the scale of the market Restate is entering.

Restate’s founding team brings significant experience from the data streaming world. Ewen previously co-created Apache Flink, a popular open-source stream-processing framework, and served as CTO of Data Artisans, which was acquired by Alibaba in 2019 and rebranded as Ververica. He co-founded Restate with former colleagues Igal Shilman and Till Rohrmann. At the time of this Series A announcement, Ahmed Farghal, a former engineer at Stripe and Meta, joined as the fourth co-founder. The company plans to use the funds to hire a go-to-market team, recruit more engineers, and expand its office presence in the San Francisco Bay Area to be closer to its growing customer base.

How it works

Restate differentiates itself through its underlying architecture. Unlike many competitors that build their durable execution engines on top of external databases, Restate developed its own storage, replication, and redundancy layers from the ground up. This integrated approach allows the system to be exceptionally fast and lightweight compared to heavier alternatives. By controlling the entire stack, the company aims to reduce the overhead typically associated with ensuring durability in distributed systems.

The core function of the platform is to track every action taken during a workflow to ensure reproducibility. If a process fails mid-execution due to a crash or network issue, the system automatically detects the failure and restores the workflow to its last consistent state. This mechanism ensures that the outcome remains consistent regardless of interruptions. Ewen emphasized that this capability is essential for agent workflows, which are inherently longer and less predictable than traditional software processes. The goal is to provide a system that puts itself back together without manual intervention, maintaining data integrity throughout complex operations.

Key details

• Restate raised $20 million in Series A funding led by Singular, with participation from Redpoint Ventures and Capital One Ventures.
• The company was founded in 2022 by Stephan Ewen, Igal Shilman, and Till Rohrmann, with Ahmed Farghal joining as a co-founder during this round.
• Restate competes with Temporal, which recently raised $550 million at a $12.55 billion valuation, but offers a lighter, self-contained architecture.
• High-profile customers include Replit and various Fortune 500 companies in the financial sector, with recent contracts reaching six and seven figures.
• The platform uses custom-built storage, replication, and redundancy layers rather than relying on external databases for durability.
• Funds will be used to expand the engineering and go-to-market teams, as well as to grow the company’s office in the San Francisco Bay Area.

Why it matters

For software engineers and technical leaders, the shift toward AI agents introduces new reliability challenges. Traditional workflows are often short and deterministic, but agent workflows can span hours or days, involving multiple external API calls and decision points. Without durable execution, a single network timeout or server restart can cause an entire agent session to fail, leading to inconsistent states and lost work. Restate’s approach offers a way to mitigate these risks without the heavy operational burden of managing complex state machines manually.

The competition between Restate and Temporal highlights a broader trend in infrastructure: the need for specialized tools that balance power with efficiency. While Temporal dominates the enterprise space with a robust, feature-rich platform, Restate’s lightweight architecture may appeal to teams looking for lower latency and reduced cost. As AI agents become more prevalent across industries, the ability to guarantee consistent outcomes will transition from a nice-to-have feature to a fundamental requirement, similar to how databases became universal tools for data persistence.

What you can do

• Evaluate your current agent workflows for points of failure, such as long-running API calls or state-dependent steps that lack retry logic.
• Compare durable execution frameworks like Restate and Temporal based on your specific needs for latency, cost, and architectural complexity.
• Implement idempotency keys in your external service calls to ensure that retries do not result in duplicate actions or side effects.
• Monitor the performance of your AI agents to identify patterns where crashes or timeouts frequently occur, indicating a need for better durability infrastructure.
• Consider piloting a lightweight durable execution engine for new agent projects to assess the impact on development speed and system reliability.
• Stay informed about advancements in AI infrastructure, as tools for managing agent state and consistency are likely to become standard components of the tech stack.