Google DeepMind analysis proves Signal and WhatsApp encryption is secure

Google DeepMind has formally verified the security of the Signal protocol, the end-to-end encrypted messaging design that Edward Snowden has notably advocated and that has long been treated as the gold standard. The Signal app, WhatsApp, and Google Messages all use that same protocol for their end-to-end encrypted messaging. 9to5Mac's report by Ben Lovejoy, published at 4:39 am PT on November 10, 2026, calls the analysis a landmark development.

This piece is for people who rely on Signal, WhatsApp, or Google Messages and want a full account of what was proved. It follows the November 10 report from the definition of formal verification through the agentic AI method and the practicality comment Google gave Wired. It then covers DeepMind's statement that the result is one milestone, the line between the protocol and the Signal app, the flags on older vulnerable versions, and a four-week run built on earlier proofs completed without AI.

01Gold-standard messaging protocol

The Signal protocol has long been considered the gold standard for end-to-end encrypted messaging, with Edward Snowden a notable advocate of that standing. A Google DeepMind analysis has now formally verified the security of the protocol. The gold-standard reputation Snowden has supported is the context, and the formal verification of the protocol's security is the new result.

In addition to its use in the Signal app, the same protocol is also used by WhatsApp and Google Messages. The verification applies to that shared protocol, the encryption design each of those three products uses. After naming Signal, WhatsApp, and Google Messages, the November 10 account describes the work as a landmark development.

02Security claims in all conditions

Formal verification is the term given to mathematically proving or disproving that a cryptographic algorithm or protocol satisfies its security claims under all conditions. It guarantees that there is no situation in which the protocol will fail. The term includes both directions, a mathematical proof that the claims hold and a disproof when they do not, and DeepMind's Signal result takes the proving side.

This is a far higher standard than traditional testing methods, which seek to find vulnerabilities but cannot promise that other vulnerabilities do not exist. A test that finds no flaw has still not promised that other flaws are absent. Formal verification, as the November 10 account defines it, requires the security claim to hold under all conditions, with no situation left in which the protocol will fail.

03Agentic AI compute cost

The November 10 account calls the use of agentic AI to carry out formal verification a huge breakthrough for cryptography and AI alike. The breakthrough belongs to both fields at once: cryptography, through a mathematical proof of the protocol, and agentic AI, the method that carried out the proof. Beside that claim, the account states the constraint in plain terms: the work is extremely computationally demanding.

Google told Wired it is unclear as yet how practical it will be to use the approach more broadly. The November 10 account ties that uncertainty to how computationally demanding the method is. The open question is broader use of agentic formal verification, and the Signal protocol analysis is the verification already completed.

04Milestone before worldwide scale

DeepMind researchers emphasize that their findings are just one milestone in a larger process. Those findings are the formal verification of the Signal protocol, and the researchers locate that result inside work that is still continuing. The November 10 account records the caution in their words: the verification is a milestone, and a larger process still extends beyond it.

Swarat Chaudhuri, a DeepMind research scientist, tells WIRED, "We are very excited by this result, but there is a long distance to go before we are able to scale this to all the software of the world, and we want to do that." He names the goal as all the software of the world, states that a long distance remains, and says DeepMind wants to scale the method that far.

05Protocol scope after four weeks

The validation is for the protocol itself rather than the Signal app as a whole. Ben Lovejoy marks that split in the November 10 account so the formal result stays a statement about the protocol. The Signal app as a whole lies outside the validation, while the protocol that the app uses is the piece DeepMind formally verified.

Encouragement can be taken from a check against older versions of Signal that contained known vulnerabilities. When the method was tested on those versions, it identified the vulnerabilities and flagged them as failing formal verification. The known flaws were caught and marked as failures, which is the outcome the November 10 account treats as encouragement.

Google was able to build on some past formal verification work already completed without the use of AI, and this analysis took four weeks. The scale of resources required has not yet been disclosed. A whole array of other protocols should not be expected to be validated overnight, and all the same the report calls this result a very big deal.