AWS and Google Cloud Bury the Lock-In War — Their Joint Multicloud Framework Resets the Rules for CIOs
On August 12, 2026, AWS and Google Cloud unveiled an open-source multicloud interoperability framework that eliminates egress fees and standardizes identity across both platforms. Azure will join before the end of the year. For CIOs, this marks the end of forced infrastructure duplication — and the beginning of genuinely agnostic cloud architecture.
August 12, 2026. AWS and Google Cloud officially announced a multicloud interoperability framework that will restructure how enterprises design their infrastructure. This is not a lukewarm press release about “strategic collaboration” — it is an open-source networking protocol, jointly developed over eighteen months by engineers from both hyperscalers, enabling Kubernetes clusters and containerized applications to communicate natively and securely across both platforms, without egress fees and without third-party routing.
Microsoft Azure has confirmed it will join the framework before the end of 2026. Oracle Cloud Infrastructure is also in discussions. For the first time since the public cloud was born, hyperscalers are deliberately building bridges instead of walls.
What the Framework Actually Changes
Until now, an enterprise wanting to run a workload on AWS and another on Google Cloud had to solve three problems, none of them trivial.
Networking. Connecting an AWS VPC to a Google Cloud VPC required a VPN or a dedicated interconnect, introducing additional latency, often-prohibitive data transfer costs, and a fragile manual configuration that broke every time someone renamed a subnet.
Identity. AWS IAM and Google Cloud IAM didn’t talk to each other. Each platform had its own logic of roles, policies, and federation, forcing teams to maintain two separate permission directories — an audit nightmare and a security incident waiting to happen.
Egress fees. Moving data out of AWS or Google Cloud was deliberately expensive. Egress fees were the economic weapon of vendor lock-in: the larger your data footprint, the more it cost to leave — or even to peek across the fence.
The new framework tackles all three head-on:
- A native, end-to-end encrypted networking protocol that connects Kubernetes clusters across both platforms without VPNs, manual routing, or data transfer fees for inter-cloud traffic.
- A federated IAM layer that synchronizes identities and security policies between AWS and Google Cloud, allowing a developer authenticated on one platform to access resources on the other without duplicating credentials.
- Unified cost governance, with a console that normalizes billing across both providers — a problem that until now was keeping FinOps consulting firms very well fed.
Why Now — and Not Five Years Ago
The answer comes down to two letters: EU.
The European Commission and the UK’s Competition and Markets Authority (CMA) have been investigating anti-competitive hyperscaler practices since 2023, with an explicit threat: impose legally mandated interoperability standards if the market doesn’t build them voluntarily. The EU Data Act, which came into force in 2025, gave this threat a timeline.
AWS and Google Cloud aren’t practicing philanthropy. They’re getting ahead of a regulation that would have been far more prescriptive — and almost certainly less favorable to their interests — than the framework they’ve just designed collaboratively. By taking the initiative, they retain control over the technical specification and avoid a Brussels-imposed standard.
There’s also business pressure: enterprise customers are demanding multicloud. According to figures released this summer, AWS posted $42.2 billion in quarterly revenue (+37%), while Google Cloud jumped 63% year-over-year. Both providers know their next growth tier will come from organizations that refuse to put all their eggs in one basket — organizations that today are either building expensive workarounds or simply declining to migrate certain workloads at all.
The Trap That Remains: Multicloud Isn’t Free
The framework eliminates network egress fees, but it doesn’t solve every multicloud problem. Operational complexity remains: orchestrating workloads across two platforms requires scarce skills, and teams that are fluent in both AWS SDKs and Google Cloud APIs are hard to find and expensive to keep.
More importantly, the framework does not cover — at least not yet — the proprietary managed services that constitute the bulk of each hyperscaler’s value proposition. You’ll be able to connect your Kubernetes clusters, but Aurora (AWS) won’t natively talk to BigQuery (Google Cloud), and Lambda won’t trigger a Cloud Function. The abstraction layer stops at the network and identity level. Everything above that remains provider-specific — and that’s probably intentional.
AWS and Google Cloud are opening the door, not tearing down the building.
The Geopolitics of Multicloud
This framework isn’t just a networking story. It carries a geopolitical dimension that European CIOs need to factor in.
First, the EU Data Act and the UK CMA antitrust investigations created the window that the hyperscalers have now stepped through. But European regulation also has a digital sovereignty objective: ensuring that critical data and workloads can be repatriated or transferred without obstruction. The AWS-Google Cloud framework partially addresses this — it eases mobility between American clouds but says nothing about interoperability with European providers (OVHcloud, Scaleway, Hetzner).
Second, the framework lands amid an explosion in AI infrastructure spending. With AWS alone planning $200 billion in annual capex and combined hyperscaler spending approaching $600 billion, the question is no longer “should we be multicloud” but “how do we do it without bleeding cash on transfer fees.” The framework answers the economics question before it even fully addresses the technical one. CIOs who were previously hesitant to deploy AI workloads across two clouds simultaneously will now be able to reconsider their architecture.
Third, Azure’s announced entry into the framework will create a cascade effect: once the three major providers are natively interconnected, regional clouds and niche providers (DigitalOcean, Vultr, Linode) will either adopt the protocol or risk isolation. This is market-driven standardization, not regulation-driven — and it may prove more durable.
What to Watch Next
The framework’s governance model is the next critical milestone. Will the protocol be submitted to an independent open-source foundation — the CNCF would be the natural home — or will it remain under joint AWS/Google stewardship? If it stays with the two incumbents, the framework is a tactical concession, not a structural shift. If it moves to a foundation, it becomes infrastructure.
Azure’s integration timeline is another key signal. Microsoft joining the framework would cover the three major providers and effectively make the protocol a de facto standard. Until then, the framework is a powerful bilateral agreement — but not yet an industry norm.
Verdict
If your organization already uses both AWS and Google Cloud, the multicloud framework is unambiguously good news: you’ll save on transfer costs and simplify identity management. Plan a phased adoption, starting with staging environments and non-critical workloads.
If you’re single-cloud, don’t change anything right now. The framework enables multicloud — it doesn’t require it — and the operational complexity of a bi-cloud architecture remains significant. Wait until Azure joins the initiative (covering all three major providers) before designing a native multicloud strategy.
The real signal is regulatory. This framework is proof that hyperscalers take antitrust threats seriously and would rather build their own standard than endure one imposed by Brussels. The next milestone to watch: does the protocol land at an independent foundation? The answer will determine whether this is a durable opening or a tactical concession. Either way, the era of egress fees as a lock-in weapon is ending — and that’s good for everyone except the hyperscalers’ finance departments.