RadarMakers editorial
A startup raised $70 million to prove code safe, not just test it
&Truth emerged with $70 million in seed funding and a plan to build independently checkable mathematical proofs about code. The ambition is large; the product is still being built.
Most software teams ask whether tests passed. &Truth wants a harder answer: mathematical proof that code meets stated safety and security requirements. The company says it emerged from stealth with $70 million in seed funding on October 27. Its mission page describes a verification machine that aims to reason about code and produce proofs that others can check.
That promise is much broader than a test suite. Tests sample chosen inputs and behaviors. Formal verification can establish that a precisely stated property holds under explicit assumptions. That still does not prove every possible meaning of the word safe. The specification, system model and proof scope matter, and the company itself says those assumptions must be visible.
The money makes this a major bet on a difficult problem. &Truth says it is building on Lean, an open source proof system, and plans to release its libraries as open source. The public page presents a mission and a signup, not a product available for teams to validate today.
The useful question is not whether a machine can declare software correct in general. It is which properties customers can state, prove and reproduce on real code, and how much work remains to connect that proof to changing software. Funding shows conviction. Only inspectable results can show whether the universal claim holds up.