AWS Goes Down, Takes Crypto's Decentralization Story With It

On October 20, 2025, AWS's US-EAST-1 region went down — and a large swath of crypto infrastructure went down with it. Coinbase froze up, Robinhood stalled, and Infura's outage broke MetaMask's connections to multiple chains. Layer 2 networks including Polygon, Optimism, Arbitrum, Base, Linea, and Scroll all dropped offline within minutes of each other.

As trader ManFromHell summarized it on X: "So AWS went down and half of crypto just stopped working. Decentralization vibes are off the charts today."

Notably, the base layers themselves kept running — Ethereum continued producing blocks without interruption, and Bitcoin was entirely unaffected. No chain went down. What broke was the layer users actually touch: exchange logins failed, wallets couldn't connect, and access to supposedly "decentralized" finance evaporated because of a DNS failure at a single cloud vendor. An industry premised on removing trusted intermediaries turned out to depend, in practice, on one company's servers staying online.

01How the Outage Propagated

Infura — the Consensys-run infrastructure layer that MetaMask relies on to reach various blockchains — reported outages across Polygon, Optimism, Arbitrum, Linea, Base, and Scroll as soon as AWS faltered. The root cause, per reporting from The Register, was a DNS resolution failure affecting AWS's DynamoDB endpoint in US-EAST-1 starting around 3:11 AM ET. Infura's status page reportedly lit up almost instantly: Ethereum Mainnet's JSON-RPC API went down, along with Polygon, Optimism, Arbitrum, Linea, Base, and Scroll.

Six chains marketed as decentralized became unreachable at the same moment, for the same reason: they all funneled user access through the identical centralized pipe. MetaMask users found transactions stuck and dapp interfaces frozen — not because validators stopped validating or blocks stopped forming, but because the on-ramp to that activity ran entirely through one vendor's infrastructure.

Coinbase supplied its own illustration of the irony: both the Coinbase app and its Base network were knocked offline for hours, blocking users from logging in or trading. Coinbase confirmed the disruption and said it was "reorganizing services" to guard against a repeat — an admission that its systems weren't as distributed as its branding implies.

The scale of AWS's footprint in Ethereum itself is notable too: roughly 2,368 Ethereum execution nodes — close to 37% of the network's total — run on AWS. That concentration wasn't enough to halt the chain, but it was more than enough to degrade access for anyone not running their own infrastructure.

02Not the First Time

This wasn't an isolated event. Six months earlier, on April 15, 2025, AWS suffered its own "connectivity issues": Binance suspended withdrawals, KuCoin went dark, and at least eight platforms — including MEXC, Gate.io, Coinstore, DeBank, and Rabby Wallet — reported disruptions within minutes of each other. October's outage largely repeated that script.

The underlying concentration risk is structural. AWS holds roughly 30% of the global cloud infrastructure market, with Microsoft Azure at 20% and Google Cloud at 13% — meaning three companies control 63% of the infrastructure underpinning the internet, crypto included. Binance and Coinbase both run on AWS; so do BitMEX, Huobi, Crypto.com, and Kraken, all drawn by the low-latency performance high-volume trading demands. When AWS falters, so do crypto's largest platforms.

03An Exception That Proves the Point

Amid the outages, the XRP Ledger kept closing blocks throughout both incidents, because its validator set is spread across AWS, Google Cloud, Hetzner, DigitalOcean, and independent operators — no single provider's failure could take the network down. As contributor Vet put it, "that's the hard work of decentralization, especially geographical and hosting wise." It demonstrates that distributed infrastructure is achievable — but only for teams willing to build and pay for it deliberately, which most aren't.

04Voices From the Fallout

The outage produced a wave of blunt commentary. Lefteris Karapetsas, founder of Rotkiapp, wrote: "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."

MattFlint captured a similar sentiment aimed at Coinbase: "The platform preaching 'decentralization' just got taken out by one centralized cloud. Irony level: off the charts."

Bitcoin maximalist Carla was more pointed still: "If an AWS outage affects your coin, then it's not decentralized nor is it money. Bitcoin, not crypto."

Karapetsas followed up with a second post: "The whole vision behind blockchain was decentralized infrastructure, which we have completely failed on." Not a work in progress — a failure, in his framing.

The theme wasn't new. Bitget CEO Gracy Chen had said much the same back during April's 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 on, the observation held up. Dr. Max Li, CEO of OORT, offered a similar diagnosis: "Decentralized cloud computing offers a powerful alternative by distributing data and processing across a network, reducing the risk of single points of failure." Coinbase's response — announcing it would "reorganize services" — amounted, in effect, to conceding the point. Markets didn't move much, but the conversation about alternatives did pick up.

05Why Everyone Still Uses AWS Anyway

The economics explain the pattern even if they don't excuse it. Running your own node infrastructure means buying hardware, securing reliable power and bandwidth, and hiring specialized staff. AWS offers a comparable service far more cheaply, backed by a 99.99% uptime guarantee built on two decades of infrastructure investment. For a startup racing to ship before funding dries up, that trade-off isn't close. Multi-cloud setups cost more; self-hosting demands expertise most teams lack; geographic redundancy introduces latency that traders notice right away. Centralized infrastructure remains cheaper and "good enough" — until an outage proves otherwise. Most projects have simply chosen speed over resilience, since justifying doubled infrastructure spend to investors on principle alone is a hard sell.

There's a second, less-discussed dimension: the US CLOUD Act gives American authorities the power to compel data disclosure from US cloud providers regardless of where the underlying servers physically sit — European or Asian data centers included, without requiring approval from a foreign court, as long as AWS, Azure, or Google is the host. That complicates crypto's censorship-resistance claims given how much of its infrastructure sits on servers subject to US legal reach. European regulators have already flagged American cloud dominance as a sovereignty concern, with some advising against using US cloud services for sensitive data altogether. A movement premised on escaping institutional control has, in practice, concentrated much of its infrastructure under three companies answerable to a single government.

06What Alternatives Exist — and Why They Haven't Caught On

Building real resilience isn't complicated in principle: spread workloads across AWS, Azure, and Google Cloud, add geographic redundancy beyond a single US region, and absorb the added cost and complexity. The XRP Ledger's performance through both outages shows the approach works — not because of superior technology, but because of deliberate architectural choices and a willingness to pay for redundancy.

Decentralized alternatives to Big Tech cloud services do exist, but adoption remains limited. Filecoin, IPFS, and Arweave offer decentralized storage; Akash Network provides decentralized compute; the Internet Computer Protocol aims at full-stack decentralization. All remain relatively early-stage, and most developers still default to the tools they already know — which is AWS.

Two weeks after April's outage, Vanar launched a product called Neutron aimed directly at this dependency. CEO Jawad Ashraf described the goal as enabling "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." For now, that remains potential rather than proven at scale. Every project ultimately faces the same binary choice: rent from Big Tech and accept the systemic risk, or build genuinely decentralized infrastructure and pay for it. Most, so far, keep choosing the former because it's cheaper today.

07The Bigger Picture

Ethereum kept producing blocks. Bitcoin was untouched. No blockchain actually failed on October 20 — the protocols performed exactly as designed. What failed was the access layer sitting on top of them, and that failure exposed an uncomfortable truth: verifying transactions on-chain means little if a single company's cloud region determines whether you can submit them at all. That makes claims of sovereignty largely performative in practice.

Two major cloud-driven outages inside six months, both traced to the same underlying cause in different regions, suggest the pattern won't correct itself — the economic incentives that make centralized infrastructure attractive haven't changed, so the industry keeps choosing it despite the known risk. An ecosystem built around "don't trust, verify" currently depends, in large part, on trusting three corporations to stay online.

More outages are inevitable. Each will likely trigger the same cycle: public frustration, promises of architectural reform, and a quiet return to convenience once attention moves on. Whether crypto actually shifts toward multi-cloud setups, geographic redundancy, and genuine decentralized alternatives — or simply keeps running centralized infrastructure under decentralized branding — remains an open question.