AWS Interconnect multicloud with Azure is GA: pricing, regions and how to set it up
AWS Interconnect multicloud with Azure is GA in 3 regions. Hourly pricing, no per-GB fee, free 500 Mbps tier, up to 100 Gbps, setup steps and limits.

On 17 November 2026, AWS made AWS Interconnect – multicloud with Microsoft Azure generally available. You can now order a managed, private, MACsec-encrypted link between your Amazon VPCs and Azure virtual networks from the AWS console or CLI, with no cross-connects, VLANs or BGP to configure. At launch it works in three AWS Regions: us-east-1 (N. Virginia), eu-central-1 (Frankfurt) and ap-southeast-2 (Sydney).
On the AWS side you pay by the hour, by bandwidth and pricing tier, with no per-GB data transfer charge. Microsoft bills its side separately. Here's how it works, what it costs, where it's available and the limits worth knowing before you build on it.
What is AWS Interconnect – multicloud?
AWS Interconnect is a family of managed private connectivity services. It has two parts: multicloud, which connects AWS VPCs directly to other public clouds, and last mile, which connects offices and data centres to AWS through partner networks. (We covered the last-mile option in our AWS Interconnect last mile explainer.)
AWS calls multicloud "the first purpose-built product of its kind." Here's what changes for you in practice:
- No physical networking work. AWS and the other cloud have pre-built capacity pools between their networks. You pick a provider, a region and a bandwidth, and both sides provision it.
- One object to manage. AWS presents the result as a single interconnect. Behind it there are at least four connections spread across two or more physically separate facilities, load-balanced with ECMP.
- Encrypted by default. The physical links between AWS routers and the partner's routers use IEEE 802.1AE MACsec. AWS's docs say the devices only carry customer traffic while the encryption session is active.
- Traffic stays on the AWS backbone until it's handed to the other provider. It doesn't hairpin through your on-premises routers.
Which clouds and regions are supported?
| Provider | Status on AWS | AWS Regions / locations (Azure) | Notes |
|---|---|---|---|
| Google Cloud | GA since 14 April 2026 | Varies; AWS's GA demo used eu-central-1 ↔ europe-west3 | First provider at launch |
| Microsoft Azure | GA, 17 November 2026 (preview from August) | us-east-1, eu-central-1, ap-southeast-2 | Azure peering locations: East US, Germany West Central, Australia East |
| Oracle Cloud (OCI) | Listed in AWS's GA announcement | Check the console | AWS's product page still describes Oracle's support as planned |
Two details are worth flagging:
- West US has dropped out. During preview, the Azure side also supported a Silicon Valley / West US pairing, according to a hands-on write-up by Gergo Vadasz. Microsoft's docs now say AWS connectivity in West US "isn't supported at general availability."
- OCI status is ambiguous. AWS's announcement says GA customers "can interconnect workloads across Microsoft Azure, Google Cloud, and Oracle Cloud Infrastructure," but the multicloud product page says Oracle "plans to expand" its capabilities. Check the console in your Region before planning around OCI.
How much does AWS Interconnect – multicloud cost?
AWS side. According to the official pricing page:
- You're charged hourly per interconnect, based on the bandwidth you choose and an automatically assigned pricing tier. Billing is rounded up to the nearest hour.
- There are no per-gigabyte data transfer charges on the interconnect.
- Billing starts when you create an interconnect and stops when you delete it.
- There are five tiers, from Tier 1 (cheapest) to Tier 5. Your tier depends on the distance between the AWS Region(s) where your VPC traffic originates and the interconnect's local Region. Each interconnect is billed at the single highest tier of any path it carries.
- Free tier: one 500 Mbps interconnect per cloud provider, per Region.
AWS's own worked examples:
| Scenario | Bandwidth | Tier | Hourly | Monthly (730 h) |
|---|---|---|---|---|
| us-east-1 VPC → interconnect in us-east-1 (to Google Cloud us-east4) | 10 Gbps | Tier 1 | $12.33 | $9,000.90 |
| Same interconnect, but Cloud WAN also has a Core Network Edge in Singapore | 10 Gbps | Tier 4 | $51.78 | $37,799.40 |
AWS uses Google Cloud in its examples, not Azure, and those are AWS-side charges only. For a smaller reference point, Gergo Vadasz reported during the Azure preview that his 1 Gbps Frankfurt interconnect cost $1.37/hour, "around $1000 per month" on the AWS side. That's one person's bill during the preview, not an official rate card.
Azure side. Microsoft says pricing "depends on the location, SKU, and bandwidth" and that "the participating provider can charge separately." Azure offers two SKUs: Local, which reaches the Azure regions tied to that interconnect location, and Premium, which reaches all Azure regions. You also need an ExpressRoute virtual network gateway, which is its own billable resource. Vadasz noted that Azure's interconnect itself was free during preview, but the gateway wasn't.
The Cloud WAN trap. If you attach the interconnect's Direct Connect gateway to an AWS Cloud WAN core network, any Core Network Edge in your topology can send traffic through it, and your tier is set by the farthest CNE, not the local Region. AWS's own example shows a us-east-1 interconnect jumping from Tier 1 to Tier 4 because the core network also has a CNE in Singapore. Check your topology before attaching.
What bandwidths can you choose?
Microsoft's docs list the bandwidths for AWS connectivity at GA: 1, 2, 5, 10, 20, 50 and 100 Gbps. (During preview, Azure was limited to 1 Gbps.) AWS's docs say you can change bandwidth on an existing interconnect without recreating it, so you can scale up for a migration and back down afterwards.
How do you set up AWS Interconnect with Azure?
At a high level, based on AWS's and Microsoft's documentation:
- Check the prerequisites. You need an Azure subscription, an ExpressRoute virtual network gateway (or a Virtual WAN hub with an ExpressRoute gateway), and non-overlapping address spaces between your AWS VPCs and Azure VNets. On AWS, the attach point is always a Direct Connect gateway.
- Avoid Microsoft's reserved ranges. Don't use ASNs 65515–65520, IPv4 169.254.255.0/24 or IPv6 fc00:ffff:ffff:ffff/120.
- Create the interconnect. In the AWS console (under Direct Connect → AWS Interconnect), the CLI or the API, choose Azure as the provider, choose the destination region, pick a bandwidth and select your Direct Connect gateway. This generates an activation key.
- Accept on the other side. AWS calls this the Create/Accept flow: one party creates and generates the key, and the other uses it to accept. After that, both providers provision automatically.
- Attach your networks. On AWS, use a virtual private gateway or Transit Gateway in the same Region, or Cloud WAN. On Azure, connect up to 10 virtual network gateway connections per interconnect. For Virtual WAN, Microsoft says you currently have to use the authorization key workflow.
- Route and monitor. Update route tables. Every interconnect comes with one CloudWatch Network Synthetic Monitor at no extra cost for latency and packet loss, plus a bandwidth-utilisation metric you can alarm on.
Vadasz reported that his whole preview setup took about 60 minutes, with a measured round-trip time of about 2.5 ms between VMs in Frankfurt. He also warned that with Terraform, "you may need to run terraform apply (and destroy) twice, because each side has to wait for the other," and that the ExpressRoute gateway is slow to deploy.
What limits and gotchas should you know?
Use this as a pre-flight checklist:
- ☐ 250-prefix limit per Azure connection. If a connection advertises more than 250 prefixes, Microsoft says "the Azure Multicloud Interconnect connection goes down and connectivity is interrupted." Summarise your routes.
- ☐ Region pairing. A virtual private gateway or Transit Gateway can only reach an interconnect local to its own Region. Cross-Region reach needs Cloud WAN, which can raise your pricing tier.
- ☐ Private connectivity only. Microsoft peering doesn't apply.
- ☐ No overlapping CIDRs. AWS's Google Cloud walkthrough also requires matching IP families (IPv4/IPv6) and MTU on both sides. Check the same on Azure.
- ☐ Two bills. AWS charges for the interconnect. Microsoft charges for its interconnect SKU and the ExpressRoute gateway.
- ☐ Delete what you don't use. AWS billing runs from creation until deletion, even when the link is idle.
- ☐ SLA. Microsoft points GA customers to its Azure Multicloud Interconnect SLA page. Preview connections had no SLA.
Is AWS Interconnect – multicloud worth it vs VPN or DIY?
| Your situation | Likely best fit |
|---|---|
| Low, bursty traffic; tolerant of internet paths | Site-to-site IPsec VPN between clouds |
| Steady multi-Gbps traffic between AWS and Azure in N. Virginia, Frankfurt or Sydney | Interconnect – multicloud (no per-GB transfer on the AWS side) |
| You already run Direct Connect + ExpressRoute through a colocation provider | Compare your current port, cross-connect and router costs against Interconnect's hourly rate |
| Workloads in Regions Azure GA doesn't cover yet | DIY or partner-based connectivity for now |
| Many AWS Regions sending through one interconnect via Cloud WAN | Model the tier first. Long-haul tiers cost several times more |
This table is our guidance based on the published features and pricing model, not a benchmark.
What are engineers saying?
Public discussion of the GA is still thin; we found no substantial Reddit threads on it yet. The most detailed hands-on account we found is Gergo Vadasz's preview write-up on DEV Community. He said "the best part is how little there is to do," praised the ~2.5 ms round trip, and called the smallest 1 Gbps AWS-side circuit "around $1000 per month." His open question, how Azure-side pricing would look at GA, is the one most teams will now need to answer for their own regions.
Editor's take
This is the most useful piece of the "multicloud is real" story AWS has shipped. Azure is where most enterprise identity and Microsoft 365 workloads live, and until now, linking it privately to AWS meant colocation contracts, routers and BGP. GA with up to 100 Gbps and no AWS per-GB charge makes it a serious option for data-heavy setups, such as analytics pipelines and AI training data, that span both clouds. The catches are real, though: only three Regions, two separate bills, a 250-prefix limit that can drop your link, and a Cloud WAN pricing rule that can quietly quadruple your hourly rate. Start with the free 500 Mbps interconnect in one Region, and model the tier before you attach a global network. If you're mapping cross-cloud identity at the same time, our guide to API Gateway's new JWT authorizer for REST APIs covers accepting Entra ID tokens without a Lambda. (Analysis, not hands-on testing.)
FAQ
What is AWS Interconnect multicloud?
It's a managed AWS service that creates private, encrypted, high-bandwidth connections between your Amazon VPCs and other clouds, including Google Cloud and Microsoft Azure. You don't manage routers, cross-connects or BGP.
Is AWS Interconnect multicloud with Azure generally available?
Yes. AWS announced general availability with Microsoft Azure on 17 November 2026, after a public preview that started in August.
What is AWS Interconnect multicloud pricing?
On AWS you pay hourly per interconnect based on bandwidth and one of five distance-based tiers, with no per-GB data transfer charge. AWS's example is $12.33/hour for 10 Gbps at Tier 1. Azure bills its side separately.
Is there a free tier for AWS Interconnect?
Yes. AWS allows one free 500 Mbps interconnect per cloud provider, per Region.
Which regions support AWS Interconnect with Azure?
AWS us-east-1 (N. Virginia), eu-central-1 (Frankfurt) and ap-southeast-2 (Sydney). On Azure, these are East US, Germany West Central and Australia East.
What bandwidth does AWS Interconnect multicloud support?
For AWS connectivity at GA, Microsoft lists 1, 2, 5, 10, 20, 50 and 100 Gbps. Bandwidth can be changed without recreating the interconnect.
Does AWS Interconnect multicloud support Google Cloud and Oracle?
Google Cloud has been generally available since April 2026. AWS's Azure GA announcement also lists Oracle Cloud Infrastructure, but its product page describes Oracle's support as planned, so check availability in the console.
How is AWS Interconnect different from Direct Connect?
Direct Connect links your own premises or colocation to AWS over ports you order. Interconnect – multicloud uses pre-built capacity between AWS and another cloud, and you get it as a single managed object attached to a Direct Connect gateway.
Sources
- AWS What's New: AWS announces general availability of AWS Interconnect – multicloud connectivity with Microsoft Azure (17 November 2026)
- AWS: AWS Interconnect – multicloud product page
- AWS: AWS Interconnect – multicloud pricing
- AWS Docs: What is AWS Interconnect?
- AWS News Blog: AWS Interconnect is now generally available, with a new option to simplify last-mile connectivity (14 April 2026)
- Microsoft Learn: Azure Multicloud Interconnect availability and limits (updated 16–17 November 2026)
- DEV Community, Gergo Vadasz: AWS and Azure Multicloud Interconnect: Hands-On First Impressions
Featured image: AWS data centres in the us-west-2 (Oregon) Region, not an Interconnect facility. Photo: Tedder, CC BY-SA 4.0, via Wikimedia Commons. Source