Coinbase Faces Extended Service Outage Due to AWS Dependency, Criticized by Gergely Orosz
Prominent engineer blogger Gergely Orosz pointed out that Coinbase experienced several hours of global trading disruption just days after the CEO publicly stated that non-technical teams could directly push production code.
The root cause of this failure was Coinbase's hard dependency on AWS; when AWS US-EAST-1 encountered issues, Coinbase's core services such as trading and wallets were simultaneously paralyzed.
Orosz believes this reflects a clear trade-off made by the engineering team between availability and development speed.
Source: Public Information
ABAB AI Insight
Gergely Orosz has long focused on the engineering culture of tech companies, previously commenting on Coinbase, Stripe, and others regarding their reliance on cloud infrastructure during periods of high growth. Coinbase CEO Brian Armstrong has been promoting a "company-wide coding" culture for the past two years, and this incident exposes the cost of that strategy in terms of infrastructure resilience.
On the capital front, Coinbase has prioritized engineering resources for the Base chain, institutional products, and rapid iteration of new features, rather than multi-cloud architecture or self-built disaster recovery. The shift in funding from high-redundancy infrastructure to user growth and transaction volume expansion aims to capture market share during the crypto bull market, but it also leads to trust risks when user assets cannot be traded during single point failures.
Similar to the multiple AWS regional outages from 2022 to 2025 that triggered chain disruptions across crypto platforms, and the regulatory requirements for traditional banks to enforce multi-cloud deployments, crypto exchanges are currently transitioning from extreme speed expansion to high availability control.
This essentially represents a technological substitution and capital concentration: a culture that pursues development speed has replaced traditional enterprise-level high availability architecture. The mechanism is that the single dependency on AWS significantly reduces operational costs but shifts the pricing power of critical availability to cloud vendors, forcing Coinbase to reallocate capital from feature iteration to infrastructure resilience building after the next major failure.
ABAB News · Cognitive Law
The faster the speed, the more dangerous the single point dependency.
When the whole company can push code, often no one is pushing the disaster recovery system.
A truly mature engineering culture evolves from "can go live quickly" to "never goes offline."