Signal president Meredith Whittaker says they had no choice but to use AWS, and that’s a problem
Summary
Signal’s president Meredith Whittaker highlighted how the messaging app — despite its strong end-to-end encryption — is vulnerable to outages because it depends on Amazon Web Services (AWS). Signal experienced a brief outage during the substantial AWS failure on 19–20 October 2025, which industry estimates say cost businesses tens of billions of dollars in disruptions. Whittaker argues the issue isn’t Signal’s choice but the concentration of power among a small number of cloud hyperscalers (AWS, Google, Microsoft) that makes realistic alternatives impractical for global, low-latency platforms. She points to the immense capital and talent required to build and maintain a worldwide infrastructure and warns that dependence on a few providers poses availability, sovereignty and geopolitical risks.
Key Points
- Signal went briefly offline during the major AWS outage of 19–20 October 2025, underscoring availability risks tied to cloud reliance.
- AWS holds a very large share of the cloud market (commonly cited as ~33%, but insiders suggest usage is even broader), creating deep systemic dependency.
- Global, real-time messaging requires low-latency infrastructure — compute, storage and edge nodes across many regions — that is extremely costly to build and operate independently.
- Hyperscalers cross-subsidise massive infrastructure investments, creating lock-in and making it uneconomic for most organisations to replicate those services.
- Talent scarcity to run global cloud platforms further limits alternatives to dominant providers.
- Regulatory and sovereignty efforts (eg EU cloud debates) struggle to displace US hyperscalers and may unintentionally favour them.
- Concentration of infrastructure raises geopolitical risk: control by a few firms exposes services to policy, regulatory or directed-access pressures.
Content Summary
The Register reports on a thread by Signal president Meredith Whittaker reacting to a recent AWS outage that knocked multiple services, including Signal, offline. Whittaker frames the problem as structural: the economics and engineering of global, low-latency services effectively force many companies to run on hyperscaler infrastructure. She stresses that provisioning and maintaining a worldwide network costs billions and requires specialised staff — resources most organisations cannot muster. The piece also touches on market-share figures, third-party analyses, European cloud-sovereignty debates and the broader implications of a concentrated cloud market for resilience and national data strategy.
Whittaker wants the outage to refocus attention on the risks of centralised cloud power and to encourage work on realistic alternatives — whether through policy, new architectures, public investment or industry cooperation — though practical solutions remain elusive given the scale and cost involved.
Context and Relevance
This is important for anyone who runs or relies on internet services: it reframes outages as a systemic market and policy problem, not just an engineering hiccup. Businesses, regulators and technologists should note the piece because it connects operational availability to market concentration, talent shortages and geopolitical exposure. As more critical services and government workloads run on a few clouds, resilience planning, procurement policy and sovereignty strategies need to consider how to reduce single-provider risk or manage it more transparently.
Author style
Punchy: Whittaker’s point lands as a clear warning — outages are a symptom of concentrated power in cloud infrastructure. Read the detail if you care about service resilience, digital sovereignty or the limits of private-sector-only solutions.
Why should I read this?
Because it nails a big but often hidden problem: the internet now depends on a handful of mega-clouds. If you run services, buy cloud, or shape policy, this explains why an outage can ripple far beyond one vendor — and why ‘we’ll just switch cloud’ isn’t a realistic fallback. Short, sharp and worth the few minutes.