We Have a Centralization Issue



October 20, 2025: AWS US-EAST-1 crashes.

Coinbase freezes. Robinhood stutters. Infura disrupted MetaMask connections. Layer 2 networks - Polygon, Optimism, Arbitrum, Base, Linea, Scroll - all offline within minutes.

ManFromHell put it bluntly: "So AWS went down and half of crypto just stopped working. Decentralization vibes are off the charts today"

Ethereum kept producing blocks. Bitcoin hummed along undisturbed.

No blockchain failed - yet exchanges went offline, wallet connections dropped, and users found themselves locked out of "decentralized" finance by a single cloud provider's DNS hiccup.

An industry built to eliminate trusted third parties just proved it can't function without trusting Jeff Bezos to keep the lights on.

When your censorship-resistant revolution stops working because Amazon's servers sneezed, who's really in control?

Credit: AWS, CryptoSlate, ManFromHell, Skipper, The Register, Decrypt, Ethernodes, CoinTelegraph, statista, FinanceFeeds, U Today, Vet, Lefteris Karapetsas, MattFlint, Carla, Gracy Chen, LexisNexis, concept board, Akash, ICP

Infura, the Consensys-backed infrastructure service that connects MetaMask to blockchains, reported outages across Polygon, Optimism, Arbitrum, Linea, Base, and Scroll when AWS crashed.

When Amazon's US-EAST-1 region suffered DNS resolution failures in its DynamoDB endpoint at 3:11 AM ET, Infura's status page lit up like a Christmas tree.

Ethereum Mainnet JSON-RPC API: down. Polygon: down. Optimism, Arbitrum, Linea, Base, Scroll: down, down, down, down, down.

Six supposedly "decentralized" Layer 2 networks went dark simultaneously because they all depended on the same centralized pipe to reach users.

MetaMask users lost access to Ethereum and Layer 2 networks, leaving transactions stuck and dapp interfaces unresponsive during the AWS outage.

Not because anything broke on-chain - validators kept validating, blocks kept getting produced - but because the road to access them ran through a single company's servers, and those servers ran on Amazon's infrastructure.

Coinbase and its Base app highlighted crypto’s centralization irony.

During the AWS outage, both the Coinbase app and Base were temporarily unusable, leaving users unable to log in, buy, sell, or withdraw crypto.

The company acknowledged the issue, stating it was actively working on “reorganizing services” to prevent similar outages in the future - essentially admitting that their infrastructure isn’t as decentralized as marketing suggests.

Meanwhile, roughly 2,368 Ethereum execution nodes sit on AWS infrastructure, representing almost 37% of the network's total.

Not enough to stop the chain, but enough to cripple access for most users who don't run their own nodes.

If decentralized systems need centralized infrastructure to function, what exactly are we decentralizing?

Feels Like Deja Vu

April 15, 2025: AWS experiences "connectivity issues." Binance suspends withdrawals. KuCoin goes dark. MEXC, Gate.io, Coinstore, DeBank, Rabby Wallet - at least eight platforms report outages within minutes of each other.

Six months later: Almost the same script.

AWS controls roughly 30% of the global cloud market. Microsoft Azure holds 20%. Google Cloud Platform takes 13%.

Three companies - Amazon, Microsoft, Google - control 63% of the infrastructure that powers the internet, including the vast majority of crypto's "decentralized" ecosystem.

Binance runs on AWS. So does Coinbase. BitMEX, Huobi, Crypto.com, Kraken - all tethered to Amazon's infrastructure for the low-latency, high-volume transaction processing their users demand. When Amazon stumbles, crypto's biggest players fall with it.

ManFromHell’s tweet wasn't hyperbole. It was an observation.

But here's what separates the posers from the real ones: XRP Ledger kept closing blocks throughout both outages.

Validators distributed across AWS, Google Cloud, Hetzner, DigitalOcean, and independent servers meant no single point of failure could take down the network.

"That's the hard work of decentralization, especially geographical and hosting wise," noted contributor Vet.

Proof that distributed infrastructure works - if you're willing to pay for it and build it intentionally. Most projects aren't.

How many more outages before "decentralized" stops being a marketing term and starts being an engineering requirement?

The Outage Heard Around Crypto

"AWS is down and then the internet stops working. But the blockchain, it never goe… wait a minute. Scratch that. This sector is a joke. Everyone preaching decentralization and censorship resistance but in reality … it's all 100% reliant on the cloud." - Lefteris Karapetsas, founder of Rotkiapp

"The platform preaching 'decentralization' just got taken out by one centralized cloud. Irony level: off the charts." - MattFlint, capturing what thousands were thinking about Coinbase.

"If an AWS outage affects your coin, then it's not decentralized nor is it money. Bitcoin, not crypto." - Carla, Bitcoin maximalist, stating the uncomfortable truth that protocol purists have been screaming for years.

Lefteris Karapetsas, dropped another philosophical hammer: "The whole vision behind blockchain was decentralized infrastructure, which we have completely failed on."

Not "struggling with." Not "working toward." Failed.

Gracy Chen, CEO of Bitget exchange, said back in April during the previous AWS outage: "AWS data center issues impacted several CEXs - no need to panic. It's a solid reminder: maybe it's time to explore decentralized cloud services."

Six months later, it still rings true.

Dr. Max Li, CEO of OORT, offered the solution everyone knows but few want to pay for: "Decentralized cloud computing offers a powerful alternative by distributing data and processing across a network, reducing the risk of single points of failure."

Coinbase responded to the backlash by announcing it's "reorganizing services" to prevent future outages. Translation: they're finally admitting their architecture contradicts their marketing.

The outage didn’t move markets, but it moved minds.

Suddenly, the conversation turned to the alternatives - what decentralization could actually look like if anyone built it.

Smart money doesn't just complain about centralization. It bets on the alternative.

When the rhetoric finally matches the infrastructure, will anyone still remember what decentralization was supposed to mean?

Why Some Chose the Easy Path

Running your own nodes means buying expensive hardware, securing stable electricity, maintaining bandwidth, and hiring people who actually know what they're doing.

AWS offers the same thing for a fraction of the cost, with 99.99% uptime promises and the kind of infrastructure reliability that took Amazon two decades to build.

For a startup trying to get to market before funding runs out, the choice isn't even close.

Multi-cloud strategies cost more. Self-hosting requires expertise most teams don't have.

Geographic redundancy creates latency issues that traders notice immediately.

Every philosophical stance about decentralization runs headlong into the brutal economics of cloud computing: centralized infrastructure is cheaper, faster, and "good enough" - until it isn't.

Most projects chose speed over sovereignty. Can't blame them. Try explaining to VCs why you're spending twice as much on infrastructure for the sake of principles.

But there's another angle nobody wants to discuss: the CLOUD Act.

US law gives authorities the power to demand data from American cloud providers regardless of where that data physically sits.

Whether it’s stored on European servers, Asian data centers, doesn't matter - if AWS, Azure, or Google hosts it, US law enforcement can access it with the appropriate legal authorization - no foreign court approval needed.

This reality complicates crypto's "censorship-resistant" narrative, especially when a significant portion of its infrastructure relies on servers subject to U.S. government data requests.

European regulators have increasingly viewed American cloud dominance as a sovereignty issue, with some authorities warning against the use of U.S.-based cloud services for sensitive data.

Decentralization promised freedom from institutional control.

Instead, most crypto infrastructure ended up in the hands of three corporations answerable to one government.

If your "trustless" system requires trusting Amazon not to comply with law enforcement, how trustless is it really?

The Hidden Cost of Convenience

Multi-cloud isn't rocket science.

Spread your infrastructure across AWS, Azure, and Google Cloud. Add geographic redundancy beyond US-EAST-1.

Accept the higher costs and complexity as the price of actual resilience.

XRP Ledger proves it works. Validators distributed across multiple providers kept the network running while everything else burned.

Not because XRPL has better technology - because they made different architectural choices and paid for redundancy.

Decentralized alternatives exist but remain marginal.

Filecoin, IPFS and Arweave offer decentralized storage.

Akash Network provides decentralized cloud computing.

Internet Computer Protocol promises full-stack decentralization.

All still in relative infancy. Adoption remains slow. Developers default to what they know, and what they know is AWS.

Two weeks after the April outage, Vanar launched Neutron specifically to address this dependency.

"This unlocks entirely new possibilities: from simply storing a file fully on-chain without relying on third parties, to querying and verifying the actual information inside the file," said CEO Jawad Ashraf.

Possibilities. Potential. Maybe someday.

Meanwhile, every project faces the same choice: rent from Big Tech and accept the systemic risk, or build truly decentralized infrastructure and accept the cost.

There's no middle ground. You're either dependent on centralized providers or you're not.

Most choose dependence because it's cheaper today, ignoring what it costs when Amazon's DNS hiccups tomorrow.

How many billions need to be temporarily frozen before "temporarily" becomes unacceptable?

Ethereum kept producing blocks. Bitcoin was unaffected. No blockchain actually failed on October 20th - the protocols worked exactly as designed.

What failed was the access layer, and that failure exposed crypto's dirtiest secret: you can verify transactions all you want, but if AWS controls whether you can submit them in the first place, your sovereignty is performative.

Two major outages in six months. Same root cause, different region.

The pattern isn't going away because the incentives haven't changed - centralized infrastructure remains cheaper and easier, so projects keep choosing it despite knowing the risk.

Crypto built itself on "don't trust, verify" while depending greatly on trusting three corporations not to go down.

Built to escape institutional gatekeepers, then handed the keys to the biggest gatekeepers on earth.

Protocol-level decentralization means nothing when the infrastructure layer is centralized.

The next AWS outage is coming. Then another. Then another. Each one will spark the same Twitter outrage, the same promises to do better, the same pivot back to convenience once the crisis passes.

Either crypto adapts - multi-cloud, geographic redundancy, actual decentralized alternatives - or it admits "decentralization" was always just marketing for a slightly different flavor of the same old centralized system.

When Amazon's next DNS failure locks you out of your "self-custodial" wallet, will you still believe in trustless systems?


compartir artículo

REKT sirve como plataforma pública para autores anónimos, nos deslindamos de la responsabilidad por las opiniones y contenidos alojados en REKT.

dona (ETH / ERC20): 0x3C5c2F4bCeC51a36494682f91Dbc6cA7c63B514C

aviso legal:

REKT no es responsable ni culpable de ninguna manera por cualquier Contenido publicado en nuestro Sitio Web o en conexión con nuestros Servicios, sin importar si fueron publicados o causados por Autores ANÓN de nuestro Sitio Web, o por REKT. Aunque determinamos reglas para la conducta y publicaciones de los Autores ANÓN, no controlamos y no somos responsables por cualquier contenido ofensivo, inapropiado, obsceno, ilegal o de cualquier forma objetable, que se pudiera encontrar en nuestro Sitio Web o Servicios. REKT no es responsable por la conducta, en línea o fuera de línea, de cualquier usuario de nuestro Sitio Web o Servicios.