What percentage of Signal's infrastructure is hosted on AWS?

1. The simple truth: the exact percentage is not disclosed in the reporting

None of the core outlets provided — Signal’s official blog, The Verge, TechRadar, Computerworld, Wikipedia or AboutSignal — publish a granular breakdown by percentage of how much of Signal’s compute, storage, edge or other services live on AWS versus Google Cloud, Microsoft Azure, or other providers; Signal’s blog states it “rents server infrastructure from a variety of providers like Amazon AWS, Google Compute Engine, Microsoft Azure, and others” but does not quantify shares ^{[1] [5]}. The Verge and Computerworld relay comments from Signal leadership about reliance on hyperscalers but likewise stop short of numerical disclosure ^{[2] [6]}.

2. Why no headline percentage surfaced after the outage

Signal’s leadership reframed scrutiny after the outage to emphasize structural concentration in cloud infrastructure — that global, real-time platforms realistically must depend on a “pre-built, planet-spanning network” that only a few companies can provide — an argument repeated across reporting and Signal’s public statements ^{[3] [2]}. That rhetorical framing shifts attention from “what percent is on AWS” to “why are there only 3–4 realistic providers,” which explains why journalists focused on systemic dependency rather than publishing a vetted percentage figure ^{[2] [3]}.

3. Beware of stray numerical claims and their provenance

A handful of secondary pieces and aggregators float precise-seeming numbers (for example a blog-style item listing percentages or quoting “95%”/“88%” style figures), but those items in the provided set do not cite primary data from Signal or independent audits and therefore cannot be treated as authoritative ^[4]. When outlets report percentages, verification typically requires either a public disclosure from the company, a provider billing breakdown, or an independent measurement study — none of which appear in the supplied material.

4. What the available sources do confirm about Signal’s cloud usage

Signal openly affirms it uses multiple major cloud providers and that end-to-end encryption is designed so rented infrastructure cannot read message contents, a technical point the company emphasizes when explaining why using commercial clouds is compatible with privacy goals ^[1]. Reporting also documents episodic moves in Signal’s history between clouds (for example past domain‑fronting work and earlier switches between AWS and Google) that illustrate an operational pattern of multi-provider usage rather than exclusive single‑provider hosting ^[5].

5. Competing narratives and possible agendas in the discussion

Signal’s executives present infrastructure choices as pragmatic and constrained by global scale needs, which positions Signal as a victim of industry concentration rather than an imprudent single-provider dependency ^{[2] [3]}. Critics and commentators sometimes weaponize outages to question privacy claims or to push for decentralised alternatives; for instance, privacy rivals and decentralization advocates urge moving off hyperscalers for sovereignty reasons, a stance highlighted in TechRadar quoting Proton and Matrix founders ^[7]. Some aggregators or pundits amplify numeric claims without sourcing them, which can feed misperception ^[4].

6. Bottom line and how the percentage could be known

The correct answer to “what percentage of Signal’s infrastructure is hosted on AWS?” based on the provided reporting is: it cannot be determined from these sources; Signal affirms multi‑provider hosting but offers no percentage breakdown in the supplied material ^{[1] [2] [3]}. A verifiable percentage would require Signal to publish a detailed infrastructure report, access to supplier billing/usage records, or an independent measurement/audit — none of which are present in the current reporting.