AWS has announced the general availability of AWS Interconnect, a managed private connectivity service with two capabilities. Interconnect multicloud provides Layer 3 private connections between AWS VPCs and other cloud providers, starting with Google Cloud at GA and Microsoft Azure and Oracle Cloud Infrastructure coming later in 2026.
Interconnect last mile simplifies high-speed private connections from branch offices, data centers, and remote locations to AWS through participating network providers. Robert Kennedy, VP of Network Services at AWS, described the collaboration with Google Cloud:
By defining and publishing a standard that removes the complexity of any physical components for customers, with high availability and security fused into that standard, customers no longer need to worry about any heavy lifting to create their desired connectivity.
Previously, connecting workloads across cloud providers required managing VPN tunnels, working with colocation facilities, and configuring third-party network fabrics, a process that could take weeks or months. With Interconnect multicloud, provisioning happens through the AWS Direct Connect console in minutes: select the cloud provider, choose source and destination regions, specify bandwidth, and provide a Google Cloud project ID. AWS generates an activation key that completes the connection on the Google Cloud side. Routes propagate automatically in both directions.
(Source: AWS News Blog post)
The AWS-Google Cloud launch is a jointly engineered solution. On the Google Cloud side, the same capability is available through Cross-Cloud Interconnect, which Google originally launched in 2023 with support for AWS, Microsoft Azure, Oracle Cloud Infrastructure, and Alibaba Cloud. The earlier version required customers to manage physical cross-connects through colocation facilities. The new jointly engineered approach abstracts the physical layer entirely, allowing on-demand provisioning from either cloud's console. Google Cloud's Cross-Cloud Interconnect also supports a Partner variant for AWS and OCI that enables connectivity without dedicated physical infrastructure.
Microsoft Azure takes a different approach. ExpressRoute provides private connectivity to the Microsoft cloud, but cross-cloud connections to AWS or Google Cloud still require customer-managed routing via colocation facilities or third-party multicloud connectivity providers such as Megaport or Equinix. Azure does offer a native managed interconnect with Oracle Cloud, but an equivalent managed service for AWS or Google Cloud does not yet exist. This makes the AWS-Google collaboration notable: it is the first jointly engineered, fully managed multicloud interconnect between two of the three major hyperscalers.
Traffic flows entirely over the AWS global backbone and Google Cloud's private network, never traversing the public internet. Every connection uses IEEE 802.1AE MACsec encryption on the physical links between the two providers' edge routers. Resiliency is built in, with each connection spanning multiple logical links distributed across at least two physical facilities. CloudWatch integration provides a Network Synthetic Monitor for tracking round-trip latency, packet loss, and bandwidth utilization.
The open specification angle is architecturally significant. AWS has published the underlying specification on GitHub under the Apache 2.0 license, allowing any cloud provider to become an Interconnect partner by implementing the technical specification and meeting AWS operational requirements. Forrester Principal Analyst Lee Sustar offered a more strategic reading of the move. In comments to CIO Dive when the preview was announced at re:Invent 2025, Sustar noted that AWS is "leveraging its market share to muscle the industry into making its approach a de facto standard," adding that the partnership puts pressure on Oracle and other cloud providers to adopt the open API.
Salesforce is among the early adopters. Jim Ostrognai, SVP of Software Engineering at Salesforce, noted that:
AWS Interconnect - multicloud allows us to establish these critical bridges to Google Cloud with the same ease as deploying internal AWS resources
Practitioners should be aware that IP address ranges on both sides cannot overlap. The Maximum Transmission Unit (MTU) must match across peered VPCs, and the default values for AWS and Google Cloud differ, which can cause silent data loss, degraded throughput, and broken connections if not addressed. At the time of GA, Google Cloud does not yet offer a web console for provisioning, so provisioning requires CLI commands on the Google Cloud side.
For organizations scaling beyond a single VPC, AWS Transit Gateway provides regional aggregation through a centralized routing hub with a single Interconnect attachment. At a global scale, AWS Cloud WAN extends the model across regions with centralized policy management and segment-based routing. These reference architectures are documented in a separate blog post by the AWS networking team.
Interconnect last mile, the second capability to launch with GA, connects on-premises locations to AWS via participating network providers. It automatically provisions four redundant connections across two physical locations, configures BGP routing, and activates MACsec encryption and Jumbo Frames by default. Bandwidth ranges from 1 Gbps to 100 Gbps and can be adjusted from the console without reprovisioning. The service includes a 99.99% availability SLA.
Pricing for both capabilities is based on a flat hourly rate for the selected bandwidth, varying by region pair for multicloud connections. Tobias Schmidt mentioned in a LinkedIn post:
Starting May 2026, each account receives one free 500 Mbps local interconnect per region.
Interconnect multicloud is available in five region pairs between AWS and Google Cloud: US East (N. Virginia), US West (N. California), US West (Oregon), Europe (London), and Europe (Frankfurt). Interconnect last mile launches in US East (N. Virginia) with Lumen, with AT&T and Megaport in progress, and additional regions planned.