What would have sounded like science fiction a few years ago is now happening: Amazon Web Services (AWS) and Google Cloud, two direct competitors in the cloud space, have launched a joint solution that allows applications to move seamlessly between clouds. The goal: ensure that the next major Internet outage doesn’t leave half the planet without services.
The new proposal combines AWS Interconnect – multicloud with Google Cloud Cross-Cloud Interconnect to offer private, high-speed connections with MACsec encryption and a redundant architecture between both providers.
The Catalyst: an outage that cost up to $650 million
This alliance doesn’t come out of nowhere. In October 2025, an AWS outage took down services like ChatGPT, Reddit, and Disney+ for hours. It’s estimated that incident caused losses between $500 million and $650 million for US-based companies.
That episode served as a stark reality check for everyone:
- One provider’s failure can take down half of the Internet.
- Many “critical” architectures practically depended on a single cloud.
- Multicloud was more of a PowerPoint concept than a real infrastructure strategy.
The response from AWS and Google has been clear: serious multicloud is on the table.
What this new multicloud “interconnect” really brings to the table
Until now, establishing private, resilient connectivity between two clouds usually involved weeks of:
- Manual configuration of VLANs, routers, and BGP
- Coordination between network teams on both sides
- Risks of human error and bottlenecks
With the new solution:
- Private connections between AWS VPCs and Google Cloud projects are provisioned in minutes, managed automatically.
- Traffic travels over dedicated links, encrypted with MACsec, protected against eavesdropping and tampering.
- The architecture incorporates multiple layers of redundancy, reducing the risk of a total outage due to network, software, or physical hardware failures.
Additionally, Google introduces the concept of “transport resources”, a logical layer that abstracts physical connectivity: users no longer wrestle with virtual cabling, but with resources managed and automatically scaled by the cloud.
And in 2026, Microsoft Azure joins the equation
Currently, AWS’s first partner in this scheme is Google Cloud, but Amazon has already confirmed that Microsoft Azure will join in 2026.
This opens the door to architectures where:
- The frontend of an online store can reside in one cloud.
- The payment or anti-fraud engine in another.
- And the generative AI logic in a third.
All connected via a managed, standardized private network mesh. Less reliance on a single provider and more flexibility to play with pricing, latency, or specific services.
For eCommerce and critical apps: pure gold
For those managing eCommerce, SaaS, online banking, gaming, or media, the core message is straightforward:
“Your business doesn’t have to go down just because one cloud fails.”
With this new multicloud layer:
- You can design systems where, if a region or provider fails, traffic is automatically redirected to the other.
- The end user continues shopping, watching videos, or chatting… unaware of what’s happening behind the scenes.
- Infrastructure teams can plan for multicloud resilience without building a nightmare of VPNs, tunnels, and handcrafted scripts.
It’s not magic: architecture, data, and consistency remain complex. But the most challenging part of networking — the pipelines between clouds — is beginning to be addressed as a managed service.
Open standards: not just AWS and Google
Perhaps the most interesting move isn’t technical but strategic: AWS and Google have published an open network interoperability specification on GitHub, allowing other providers to implement it.
This means:
- The future inclusion of other hyperscalers and regional clouds.
- Fewer “walled gardens” and more de facto standards for multicloud connectivity.
- Increased regulatory and market pressure to prevent providers from blocking exits or interconnections.
Practically, it’s an implicit acknowledgment that the future of cloud will be multicloud by necessity, driven by resilience and competitiveness.
What’s next?
For many infrastructure teams, this announcement is a clear invitation to reconsider their strategy:
- If your critical app relies entirely on a single provider, it’s time to rethink that approach.
- If you already talk about multicloud, but only on a business level, now there’s a key technical piece to make it real.
- And if you work in frontend or product, perhaps it’s time to talk with the systems team: “What happens if our main cloud goes down for 8 hours tomorrow?”
Given the stakes, trusting everything to one provider no longer looks like a good idea. The fact that AWS and Google are partnering to say this openly makes it even clearer.
Source: AWS Interconnect API