AWS and Google Cloud have jointly built a managed way to connect private networks in the two clouds. The collaboration is specifically about networking—not a combined cloud platform, shared billing system, or portable applications. Announced on November 30, 2025, the service moved from preview to general availability on April 14, 2026; AWS added a free AWS-side 500 Mbps local tier on May 29, 2026. Availability and charges still depend on the regions and services involved.
What AWS and Google Cloud announced
The November 30, 2025 announcement paired AWS Interconnect–multicloud with Google Cloud Cross-Cloud Interconnect and introduced an open specification for cloud-network interoperability. The goal is to make private, high-speed connectivity between the providers easier to provision and operate. Google Cloud’s announcement and AWS’s announcement describe the jointly engineered networking work.
The product was announced in public preview on November 30, 2025, and AWS announced general availability with Google Cloud as its first launch partner on April 14, 2026. AWS’s preview notice and GA announcement document those milestones.
This is cooperation at the network-interoperability layer. It does not make AWS workloads run natively on Google Cloud, combine billing or consoles, or make identity systems, databases, Kubernetes environments, and application APIs interchangeable. Each provider continues to operate its own cloud services and controls.
#1 Best Overall
How the connection works
In a conventional cloud-to-cloud setup, customers may need to arrange connectivity through facilities or an exchange, configure routers and BGP, coordinate provisioning with multiple parties, and manage capacity and troubleshooting across separate networks. AWS says Interconnect–multicloud replaces much of that basic connection plumbing with provider-managed infrastructure; it does not remove the need to design the networks and applications using the connection. See the AWS Interconnect overview.
- Select an AWS region and a supported Google Cloud region.
- Request the AWS interconnect and provide the Google Cloud project details.
- Use the activation key to complete the corresponding Google Cloud-side activation.
- Connect the resulting interconnect to an AWS networking service and configure the routes and security policies your workloads require.
At a high level, traffic travels from an AWS VPC through an AWS networking service and Direct Connect gateway to AWS Interconnect–multicloud, then across the provider-managed connection to Google Cloud Cross-Cloud Interconnect and a Google Cloud VPC. AWS describes a managed Layer 3 connection with dedicated bandwidth; its product page outlines the service.
Supported AWS attachment patterns include Virtual Private Gateway, Transit Gateway, and Cloud WAN. Virtual Private Gateway and Transit Gateway connections are regional; Cloud WAN can provide broader reach through its core network, with corresponding architecture and pricing considerations. The AWS getting-started guide explains the attachment workflow.
Rank #2
Supported AWS–Google Cloud regions
AWS lists these eight supported region pairs. The service is not available in every region merely because both providers have a presence there; check the AWS regional availability page before designing around a particular location.
Free tools Windows power users keep installed
One-click scans. No signup required.
| AWS region | Google Cloud region |
|---|---|
us-east-1 — N. Virginia |
us-east4 — N. Virginia |
us-west-1 — N. California |
us-west2 — Los Angeles |
us-west-2 — Oregon |
us-west1 — Oregon |
eu-west-2 — London |
europe-west2 — London |
eu-central-1 — Frankfurt |
europe-west3 — Frankfurt |
eu-north-1 — Stockholm |
europe-north2 — Stockholm |
ap-southeast-1 — Singapore |
asia-southeast1 — Singapore |
ap-southeast-2 — Sydney |
australia-southeast1 — Sydney |
Provisioning and practical prerequisites
The AWS console workflow begins in the AWS Direct Connect Console. Choose AWS Interconnect in the navigation pane, then select Create new multicloud Interconnect. The setup guide walks through provider, region, bandwidth, gateway, and activation choices.
- Select Google Cloud, then choose the source AWS region and destination Google Cloud region.
- Choose a bandwidth option available for that provider and region pair.
- Select or create a Direct Connect gateway.
- Enter the Google Cloud project ID. AWS specifies a unique string of 6–30 letters, numbers, and hyphens.
- Submit the request, then use the activation key to finish setup on the Google Cloud side and confirm the connection is attached to the intended gateway.
Creating the AWS-side connection alone does not complete the setup. Also review address allocations and routes: overlapping IP ranges, unintended route exports, or asymmetric routing can stop applications communicating even when the interconnect itself is active. AWS lists a default quota of two multicloud connections per provider per account and ten AWS Interconnect connections total per account; check the quota documentation and request any needed increases before planning a larger topology.
Rank #3
Security, resiliency, and what remains your responsibility
The service uses private connectivity rather than the public internet for the interconnect path. AWS describes MACsec encryption on physical connections between AWS and adjacent provider devices, provider-managed redundant infrastructure, and up to four-way resiliency. Those are AWS-described service properties, not a guarantee that an application will remain available through every outage or configuration error. The AWS overview also describes CloudWatch Network Synthetic Monitor and bandwidth-utilization metrics.
Link protection does not replace cloud security design. You still need to manage AWS and Google Cloud firewall rules, identity and authorization, segmentation, data-at-rest protection, and application-layer encryption where appropriate. Route isolation and least-privilege access remain important because a private network path can still carry traffic to an improperly exposed service.
Do these 3 things before closing this tab:
1Clear out junk files and repair common Windows errors2Scan for outdated or missing drivers - takes under a minute3Repair Windows errors before they cause bigger problemsRedundant network infrastructure is only one part of resilience. Regional failures, control-plane issues, route-policy mistakes, or application dependencies can still interrupt service. Disaster-recovery plans should test the application and data path, not just verify that the connection has redundant components.
Pricing: free on one side is not free end to end
AWS pricing for the interconnect is an hourly charge based on bandwidth and geographic scope, with the applicable region and route affecting the price. AWS says it does not add a separate Interconnect data-transfer charge for traffic over the interconnect itself, but other network charges can apply. Google Cloud sets and bills its side independently. The AWS pricing documentation directs customers to price a specific route rather than rely on one universal rate.
On May 29, 2026, AWS announced one free local 500 Mbps interconnect per AWS region, per generally available cloud provider. That is an AWS-side allowance; Google Cloud charges may still apply, as can cross-region transfer, Transit Gateway data processing, Cloud WAN, and other workload-related costs. Review the free-tier terms and route-specific AWS pricing scenarios before estimating total cost. AWS documentation gives a supported-speed example of up to 10 Gbps for Google Cloud, but actual options vary by provider and region pair.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Where the collaboration may help
A managed private link may be useful when an organization already runs production systems in both clouds and needs predictable network connectivity without building the physical cross-connect and routing arrangement itself. Potential workloads include application tiers split between providers, replication, disaster recovery, or using analytics and AI services in one cloud with data or applications in the other. These are architecture possibilities, not built-in application portability.
Outdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware matchPC Slower Than It Used to Be?
A free scan shows the junk files, broken settings and background clutter dragging Windows down - then fixes them in one click.Free scan · Windows 10 & 11Best Value
- Good fit: Both clouds are strategic platforms, the needed region pair is supported, and reducing network provisioning work is valuable.
- Consider another approach: Workloads sit in unsupported regions, you need several clouds or sites, or centralized policy and traffic steering matter more than a direct AWS–Google link.
- Check the economics: Transfer, transit-processing, and provider-side charges may outweigh the interconnect fee or free AWS tier.
- Separate the goals: If the primary need is application portability, identity federation, or unified cloud operations, this network service does not provide those capabilities.
How it compares with provider-neutral connectivity
AWS–Google native interconnect is most compelling for a supported two-cloud connection in an AWS-centered network architecture. Network-as-a-service providers, cloud exchanges, SD-WAN, and virtual-router platforms can be worth evaluating when an organization needs several clouds, colocation facilities, broader regional reach, or a more centralized policy layer. Examples include Megaport Cloud Router, Equinix Fabric, Aviatrix, and Alkira.
Those alternatives add their own provider, billing, and support boundaries, and may be unnecessary for a simple supported AWS–Google connection. Compare region coverage, bandwidth, redundancy design, route control, monitoring, incident ownership, and all transfer and processing costs. Google-side teams may also assess Google Cloud Cross-Cloud Interconnect and Network Connectivity Center; AWS-centric teams can compare the broader options in AWS Direct Connect.
What the open specification does—and does not—mean
The open specification is intended to give other cloud providers and partners a model to adopt for network interoperability. AWS describes it in its multicloud architecture post. “Open” does not mean the service is open-source software, that all providers have adopted it, or that a universal multicloud control plane now exists. Each provider can still determine its own availability, pricing, quotas, and operational processes.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
