Durable runtimes for production AI agents - persistent state and recovery

Durable runtimes for production AI agents - persistent state and recovery targets a $3.6B = 60,000 target SaaS and platform engineering teams x $60K ACV. Rationale: mid-market and enterprise SaaS that will embed AI agents and need a production runtime, each allocating platform or infrastructure budgets (~$30k-$100k) for mission-critical agent infra. total addressable market with low saturation and a year-over-year growth rate of 30-50% adoption growth for agent platforms as LLM costs drop and product integrations rise.

Key trends driving demand: Agentization of SaaS -- more products embed multi-step LLM agents to automate workflows, increasing demand for durable runtime primitives; Separation of compute and state -- serverless compute remains ephemeral while apps need persistent state for long-running agents; Orchestration and checkpoints -- growing use of workflow engines and desire for checkpoint/retry semantics for reliability.