The two have launched a jointly developed multicloud networking standard and managed interconnect service, aiming to eliminate barriers
AWS and Google Cloud have introduced a unified multicloud networking solution designed to simplify private connectivity between the two hyperscalers, replacing the manual, hardware-centric models that have traditionally slowed cross-cloud deployments.
The collaboration establishes an open interoperability specification intended for adoption across the wider cloud and service provider ecosystem. Given recent cloud outages, the service could be described as “much anticipated” – AWS finally joins the fold. The new integration connects AWS Interconnect – multicloud with Google Cloud’s Cross-Cloud Interconnect and forms part of Google Cloud’s broader Cross-Cloud Network framework.
At its core, the system abstracts physical connectivity, link-local addressing, routing configuration and security policy management, offering automated provisioning through standard cloud consoles and APIs. Both companies say that connectivity workflows that previously required weeks of procurement, installation and co-ordination can now be established within minutes.
AWS VP of network services Robert Kennedy described the move as “a fundamental shift in multicloud connectivity”, noting that the open specification removes the “heavy lifting” associated with physical builds and allows customers to activate high-availability, private circuits on demand. Google Cloud VP/GM of cloud networking Rob Enns said the collaboration “delivers on Google Cloud’s Cross-Cloud Network solution” by enabling customers to move applications and data between providers with enhanced operational efficiency.
Expansions planned
The solution is launching initially in Northern Virginia, Oregon, London and Frankfurt, with further regional expansions planned. Bandwidth begins at 1 Gbps during preview and is expected to scale to 100 Gbps at general availability, a detail aimed at organisations designing latency-sensitive applications or distributing AI workloads across heterogeneous infrastructure.
Security and resiliency form critical elements of the architecture according to the companies. Traffic between the providers’ edge routers is encrypted using MACsec with managed key rotation, and the physical underlay incorporates quad-redundant paths across distinct facilities and routers. Both cloud providers have integrated proactive monitoring and co-ordinated maintenance processes to reduce the risk of service-impacting failures.
The managed design is also intended to reduce customer exposure to multi-step networking builds, which previously required dedicated circuits, VLAN provisioning, BGP session configuration, Autonomous System numbering and bespoke encryption schemes. By replacing these tasks with a single Google Cloud “transport” construct and an AWS acceptance step, engineers can treat the connection similarly to a native VPC peering workflow.
Uses cases
In a bunch of blog posts the two highlighted applications such as active-active and active-standby disaster recovery for distributed databases and AI platforms; high-throughput pipelines between BigQuery and AWS data-stores; and private, inbound API access between services running across the two environments. Jim Ostrognai, SVP of software engineering at Salesforce, said the AWS Interconnect – multicloud capability allows the firm to connect workloads to Google Cloud “with the same ease as deploying internal AWS resources”, helping customers anchor AI and analytics models in consistent datasets regardless of location.
Both AWS and Google Cloud emphasise that the specification is open and intended for wider industry adoption, inviting other cloud and service providers to support the same model of on-demand, private interconnects. Google Cloud said the openness of the framework allows additional suppliers to contribute implementations, strengthening multicloud portability at a time when AI-driven infrastructure sprawl is accelerating.
For operators, the development offers a route to more tightly integrated hybrid cloud environments without the operational overhead of managing physical circuits or multivendor routing topologies. As networks evolve to support distributed AI inference, edge workloads and cross-domain orchestration, automated multicloud transport may help reduce latency exposure and simplify high-bandwidth workloads that span multiple hyperscalers.
