VPC deployment
person concept tool org talk claim — click a node to jump to its page; hover an arrow for the relation
VPC (Virtual Private Cloud) deployment is a hosting model in which software is installed and run within a customer's own cloud network environment, rather than accessed as a shared external SaaS service. In the context of AI agent infrastructure, it is a key requirement for enterprises that need to connect sensitive production systems to agent-based tooling without routing data through third-party platforms.
Role in Enterprise Agent Workflows
Jason Lopatecki of Arize identifies VPC deployment as a practical necessity for large enterprise customers adopting agent-based debugging and observability workflows. The core driver is data security and trust: enterprises with production systems are unwilling to connect those systems directly to external providers such as Anthropic. As a result, platforms like Arize's AX SaaS offering support VPC installation — Lopatecki notes that Arize "installs in the VPC of a lot of, you know, big-name companies" as a standard deployment path for this customer segment. →
Relationship to Sandboxes
VPC deployment intersects with the question of sandboxes for agent execution environments. Lopatecki positions VPC-deployed sandboxes as one option alongside hosted sandbox providers: customers can use Arize sandboxes, or alternatives such as Daytona, depending on their comfort level and existing infrastructure. → This framing treats VPC deployment not as a monolithic requirement but as a flexible axis — the underlying sandbox tooling can vary, while the VPC boundary addresses the enterprise data-residency and connectivity concern.
Platform Availability
The Signal feature within the Arize AX SaaS platform is the specific context in which VPC deployment is discussed; Lopatecki notes Signal is currently available on that platform, which itself supports VPC deployment as a configuration option. →