AWS Outage Highlights Cloud Concentration and the Multi-Cloud Shift
A massive AWS outage on Oct. 20, 2025 disrupted services worldwide, underlining how heavily the internet relies on a few cloud giants. In North Virginia’s AWS “us-east-1” region, a misconfigured DNS service knocked out key systems for more than 15 hours, taking down apps from social media (Snapchat, Reddit) to fintech (Coinbase, Venmo). Downdetector logged issues for thousands of companies, roughly 2,000 firms globally, and millions of end users, on the order of 4 to 8 million problem reports. This outage wasn’t unique to AWS. It only grabbed headlines because AWS underpins so many services, but it exposed the fragility of today’s centralized cloud infrastructure.
The impact was felt across industries. Government services (UK HMRC), financial apps (Lloyds Bank, Nasdaq), streaming (Netflix), and even Amazon’s own retail and Ring IoT services were knocked offline. In effect, when a major hyperscaler hiccups, it can bring down apps and websites around the world. One expert summarized this as a massive concentration risk inherent in building the global internet on just a few providers. With AWS alone serving hundreds of thousands of customers, even users who had no idea they relied on AWS discovered the dependency during the outage. The fallout, millions of users unable to send Venmo payments, delayed flight check-ins, unresponsive web pages, was a blunt reminder of how much a single failure can cascade.
Hyperscaler Market: AWS, Azure, Google (and Oracle)
AWS remains the largest cloud provider, but it now shares the stage with two fast-growing rivals. As of 2025, AWS held roughly 30 to 38 percent of global cloud infrastructure spend, with Microsoft Azure around 20 to 23 percent and Google Cloud about 12 to 13 percent. Oracle Cloud, by comparison, holds only a 3 percent slice of the market. These three, AWS, Azure and Google Cloud, have effectively become the backbone of the modern internet. When one stumbles, all are reminded of their interdependence. Even Azure and Google Cloud have suffered major outages in recent years. In short, no system is foolproof, regardless of reputation or size.
All three hyperscalers are chasing the same AI and enterprise workloads, but AWS’s growth has notably slowed relative to its smaller peers. For example, in Q3 2025 AWS revenue grew about 20 percent year-over-year, whereas Microsoft’s Azure grew roughly 40 percent and Google Cloud about 34 percent. AWS is bigger in absolute terms, over 33 billion dollars in Q3 versus Google’s 15 billion, so even smaller percentage gains translate to huge dollars. In earlier quarters, similar gaps appeared. Q2 2025 saw AWS revenue rise approximately 17 percent versus Google Cloud’s 32 percent and Azure’s approximately 26 percent. The pattern is clear. AWS is the market leader by size, but growing more modestly as the latter two catch up. This dynamic partly reflects AWS’s earlier start. It launched in 2006, years before Azure or GCP, giving it scale but making further percentage growth harder.
The Multi-Cloud Move: A Hedge Against Outages
The AWS outage has added fresh urgency to a practice already underway: multi-cloud adoption. Simply put, many companies are now spreading their services across multiple clouds by design, so that no single vendor outage can cripple them. In practice, this means running parallel deployments on AWS, Azure, Google or even on-prem clouds, with traffic able to fail over automatically if one side goes down.
Gartner has noted that over 80 percent of large enterprises now use multiple public clouds in some way. For example, a 2024 report estimated that more than 92 percent of large enterprises operate in a multi-cloud environment, often mixing AWS, Azure, Google, and others. Even smaller firms increasingly rely on two or more providers to diversify. The reasons are obvious: major outages at any one cloud have become too painful to ignore.
Diversification across providers adds costs and complexity, of course, but for mission-critical businesses it’s becoming non-negotiable. For example, during a 2025 Google Cloud outage that left many e-commerce sites offline, some firms with multi-cloud setups, including Mercado Libre, automatically shifted traffic to AWS or Azure and remained online. Others weren’t so lucky. Likewise, Lloyds Bank, HMRC and telecoms in the UK all got hit hard when AWS went down. The lesson: single-cloud is a single point of failure. As Cornell professor Ken Birman observed, engineers need to build better fault tolerance and use tools like cross-provider backup precisely for this reason.
Of course, not every workload can easily move clouds. Vendor lock-in and data gravity mean many firms have some legacy dependencies. But the fact that even big US banks, popular games, and government systems were all simultaneously affected shows why more IT leaders are now re-examining their cloud strategies. This outage could be a wake-up call. Maybe it is good that it happened, so companies take resilience more seriously.
Hyperscalers Learn Tough Lessons and Consumers Benefit
No cloud provider will intentionally publish downtime, but these incidents bring lessons. All three major cloud vendors are undoubtedly expanding redundancy and encouraging customers to use multi-zone and multi-region designs. Microsoft and Google themselves have had high-profile outages, so they (quietly) empathize with AWS’s blunder. In fact, Azure and Google Cloud have been pushing their own multi-region and geo-redundant offerings for years, partly citing AWS blackouts as cautionary tales.
The long-term effect may be a more decentralized cloud ecosystem. AWS was the original leader and made cloud mainstream, but as Azure and Google catch up, both in technology and in reassuring enterprise-grade reliability, companies have more confidence to mix providers. Diversification also weakens vendor lock-in. If firms see that switching is possible, and even beneficial in a crisis, they can negotiate better deals or flip to cheaper clouds when prices rise.
AWS itself will, of course, survive, and for readers who recently bought AMZN, will know, thrive, and continue growing, but its days of commanding indefinite loyalty simply by market share may be over. That is likely a good thing. A healthier internet is one where critical services are not all pinned on a single vendor’s fate. The recent AWS outage, while painful, may accelerate the diversification that our cloud ecosystem needs as compute requirements with AI continue to scale.
In short: The outage was a clear reminder that cloud centralization has risks. AWS, Azure and Google Cloud each can fail, so multicloud is quite clearly the answer. Businesses are already moving to dual or triple cloud architectures to avoid bringing down half the internet next time. This trend toward spreading workloads across providers will deepen resilience and spur healthy competition, which is ultimately good news for the cloud-dependent AI boom all of the hyperscalers are vying for.