Infrastructure Roadmap

Platform Roadmap

Transparent view of what we're building for enterprise AI infrastructure. Updated every quarter.

Q2 2025 — Shipped

shipped

Run agents across US-East, US-West, EU-West, and AP-Southeast with automatic failover.

Agents retain context across sessions with vector-backed long-term memory.

Native connectors for Kafka, Pub/Sub, and SNS/SQS to trigger agents from streams.

Per-agent cost, latency, and throughput dashboards.

active

Formal BAAs for healthcare enterprise customers. Dedicated PHI-safe execution zones.

Bring your own Docker image for specialized agent workloads.

Policy engine to enforce data residency, cost caps, and approval gates per agent.

Deploy and manage AgentCloud infrastructure as code via the official Terraform provider.

planned

FedRAMP Moderate authorization for US federal agency customers.

Dedicated GPU instances for vision, speech, and model-heavy agent workloads.

Deploy new agent versions with traffic splitting and one-click rollback.

Tag execution costs by department, project, or environment for chargeback.

exploring

Deploy lightweight agents to 200+ edge PoPs for ultra-low latency tasks.

Automatically burst to AWS/GCP/Azure capacity during peak demand.

Dedicated single-tenant environments for regulated industries.

Certified AgentCloud agents listed on AgentDesk for cross-product discovery.

Enterprise customers directly shape our priorities. Tell us what you need.