Nearly a decade ago, a major financial services firm faced an AWS account manager during an aggressive cloud migration. AWS advised to keep it simple, run everything on AWS, and avoid other clouds. This seemed wise, as AWS promised ease of use, top-tier services, and little interoperability trouble.
Over the years, that decision proved costly. While trying to adapt to new markets, evolving compliance standards, and rapidly changing technology, the company found itself cornered, spending far more than necessary, battling vendor lock-in, and missing out on innovations AWS didn’t offer. The dream of simplicity turned into a nightmare of missed opportunities.
Last week, AWS previewed Interconnect-multicloud, a multicloud service that provides direct, high-speed connections to Google Cloud and will soon connect to Microsoft Azure. This dramatic shift in AWS’s multicloud strategy declares that the world’s original cloud innovator has finally recognized what its most forward-thinking customers have known for years: The future is multicloud.
AWS’s decade-long resistance
Since cloud computing became mainstream in IT, AWS has consistently promoted a clear message: Going all-in on AWS is the best choice. The company has emphasized technical simplicity and reduced operational risk. Deeper research reveals AWS white papers, blogs, and field advice all warning about the “risks” of multicloud, such as higher costs, added complexity, and potential security vulnerabilities. Every customer presentation I’ve seen from them highlights the dangers of inconsistency.
But here’s the truth: These warnings rarely reflected reality. The driving logic behind AWS’s stance was market control, not customer value. When you run everything on AWS, the flywheel effect kicks in for their bottom line rather than your innovation. The assertion that multicloud was risky, too expensive, or inherently unmanageable was a narrative driven by self-preservation.
For those of us working with real-world enterprises—those whose architectures refused to fit neatly into a single branded silo—this dogma rang hollow. For years, we watched businesses force square pegs into round holes, all because the wisdom of the world’s largest cloud was: “Stick with us. Don’t worry about the rest.”
The innovation of best-of-breed
A common myth is that multicloud adds overhead without benefits, but this is false. Cloud computing’s promise has always been the ability to provision, experiment, and scale across diverse resources. Modern enterprise architecture utilizes the best available tools: AWS for elastic compute, Google for artificial intelligence and machine learning, and Azure for data analytics. If a CTO ignores two-thirds of market innovation, the lost opportunity costs can be huge.
Best-of-breed philosophies thrive in multicloud environments. Businesses need the flexibility to combine services: storage from one provider, data lakes from another, networking from various sources. This optimizes performance, compliance, cost, and even proximity to partners and users. The practitioners who embrace this model despite vendor pressure build architectures that are more resilient, cost-effective, and better aligned with business goals. These aren’t theoretical musings. The world’s most successful digital-native enterprises are all intentionally multicloud.
Single-cloud tunnel vision
The opening story isn’t unique. Over the years, I’ve encountered countless organizations whose loyalty to AWS’s vision cost them dearly: extra millions spent on suboptimal services, years wasted on migration projects that never delivered the promised ROI, and competitive initiatives lost to those unburdened by single-vendor dogma.
Do-it-yourself multicloud strategies have been intimidating mainly because hyperscalers made cross-cloud networking and management overly complicated. The need for patchwork overlays, third-party SD-WANs, and complex security setups caused real frustration. The irony? Most of that frustration came from cloud providers’ hesitation to build native interoperability. Instead of uniting customers, they built walls high enough to turn climbing them into an enterprise challenge.
The end result? AWS likely left billions on the table. As customers grew more sophisticated, the world shifted inexorably toward the best tool for the job, regardless of the badge. AWS’s insistence on all-in strategies looked less like thought leadership and more like stubbornness.
AWS Interconnect-multicloud recognizes reality
What makes last week’s unveiling of AWS Interconnect-multicloud so significant is its recognition that AWS is now adopting an open, resilient, high-speed private connection model for other cloud service providers, starting with Google Cloud (some poetic justice there) as the launch partner and Microsoft Azure soon to follow.
What’s striking is how AWS Interconnect-multicloud doesn’t just close a connectivity gap; it acknowledges that customers were right all along. This new solution makes it easy to link AWS’s flagship services—VPCs, Transit Gateway, Cloud WAN—to other clouds, reducing what once took weeks or months to a single click in the AWS Management Console. By leveraging dedicated bandwidth, built-in resiliency, and the simplicity of open APIs, AWS is making connections that are physically and philosophically smoother.
Enterprises will keep demanding flexibility, performance, and innovation in their cloud journeys. Multicloud isn’t a trend; it’s a necessary capability. The future can’t be controlled by a single brand but by how effectively cloud services meet business needs. AWS’s sudden change of course is evidence of the wisdom of removing artificial barriers that keep users from true innovation. The message could not be clearer: Use the right tool for the job, wherever it lives, however it needs to connect.
For the company I mentioned at the start—and for many like them—today’s landscape has finally changed for the better. No single provider owns the cloud. The best architectures are multicloud.
And, at long last, even AWS is on board.
