The most expensive software bug in history cost a European Space Agency rocket $370 million and ten years of work. The most common ones just waste developer time and customer patience. The math to prove they don't exist, however, has historically been so labor-intensive it was reserved for life-or-death systems like aircraft controls and nuclear reactor software. A new startup, &Truth, is betting that generative AI can finally make that math cheap enough for the rest of the codebase, and it just raised $70 million in seed funding to prove it [Wired, October 2026].
The universal verification machine
&Truth calls its product a "universal verification machine." In practice, it's software that reads code,with an initial goal of handling any programming language,and uses a combination of AI and formal methods to generate a mathematical proof that the code is safe, secure, and correct [Perplexity Sonar Pro Brief, October 2026]. The proof is independently checkable, moving verification from a matter of expert opinion to a deterministic yes-or-no answer. The company's technical foundation is Lean, a proof assistant and programming language also used for verifying complex mathematics [Perplexity Sonar Pro Brief, October 2026]. The bet is that AI can handle the tedious, logical scaffolding of proof construction, dramatically reducing the human effort required.
A team built for the proof
The founding team reads like a roster assembled to tackle this specific, deep technical problem. It blends academic authority in formal verification with executive experience in scaling enterprise software.
| Role | Name | Key Background |
|---|---|---|
| Co-Founder, Co-CEO | Clark Barrett | Stanford professor, founding co-director of the Stanford Center for AI Safety, director of the Stanford Center for Automated Reasoning [Perplexity Sonar Pro Brief, October 2026]. |
| Co-Founder, Chief Science Officer | Ilya Sergey | Founded the VERSE Lab at the National University of Singapore; co-created the open-source verifiers Veil and Velvet [Perplexity Sonar Pro Brief, October 2026]. |
| Co-Founder, Co-CEO, Exec. Chairman | Chris Hsu | Former CEO of Kilometre Capital and managing director at Citadel Investment Group; former CEO of Micro Focus [Perplexity Sonar Pro Brief, October 2026][Bloomberg, March 2018]. |
| Co-Founder | Sebastian Ullrich | Co-architect of the Lean 4 proof assistant and co-founder of the Lean Focused Research Organization [Perplexity Sonar Pro Brief, October 2026]. |
The commercial push is led by Chief Business Officer Drew Harman, formerly of Insight Partners and Shasta Ventures, and the team includes other veterans from finance and enterprise software [Perplexity Sonar Pro Brief, October 2026].
The wedge and the first proof
The company's initial wedge is making formal verification faster and cheaper, specifically for developers of security-sensitive, safety-critical, or AI-generated code. The first public demonstration of its approach was on code implementing encryption for the Signal messaging application [Perplexity Sonar Pro Brief, October 2026]. This is a clever proof point: Signal is widely respected for its security, and verifying its encryption is a non-trivial, credibility-building task. It's important to note that this demonstration does not establish Signal as a paying customer, but it serves as a powerful technical reference.
The $70 million seed round, announced as the company emerged from stealth in October 2026, is notably large for a seed, especially in a deep-tech field with a long research runway [Business Wire, October 2026]. The backers were described only as a group of "AI safety and venture capital leaders," with no specific names disclosed [Perplexity Sonar Pro Brief, October 2026]. The size of the round suggests investors are buying into a long-term, capital-intensive vision of building a new foundational layer for software development.
The scale of the ambition
The ambition is vast. If &Truth succeeds in making verifiable code as routine as running a linter, it could reshape software development in high-stakes industries. The potential customers are any organization that cannot afford bugs: aerospace, automotive, financial infrastructure, critical hardware, and the growing universe of AI-generated code where traditional testing is insufficient.
- Technical moat. The combination of world-class formal methods expertise, control over the core Lean foundation, and proprietary AI for proof generation creates a high barrier to entry.
- Market timing. The AI safety crisis and the proliferation of AI-generated code have created a new urgency for provable correctness, a tailwind that pure-play formal verification companies a decade ago did not have.
- Commercial bridge. The team's blend of deep technical and seasoned commercial talent is explicitly designed to cross the chasm from academic research to enterprise product.
The incumbent to beat
The real competition isn't another startup. It's the entrenched incumbent: the entire culture of software development that accepts bugs as a cost of doing business. Manual code review, unit testing, and dynamic analysis are the default. They are good, but not provably perfect. For &Truth to win, it must make its mathematical proof not just better, but economically sensible. The back-of-the-envelope calculation is simple: if a team of five engineers spends six months manually verifying a critical module, that's roughly $1 million in fully loaded cost. If &Truth's machine can do it in a week for a fraction of the price, the unit economics of safety change.
The company it must beat, then, is the status quo. It must become cheaper to prove code correct than it is to fix the bugs later. For the industries where a single line of faulty code can mean a recalled car, a halted trading desk, or a breached satellite, that math is already compelling. &Truth's job is to make it compelling for the rest of us.
Sources
- [Wired, October 2026] The 'Pain in the Ass' Math That Could Save Vibe Coding From Itself | https://www.wired.com/story/software-verification-is-the-hot-new-answer-to-the-ai-safety-crisis/
- [Perplexity Sonar Pro Brief, October 2026] &Truth company brief
- [Business Wire, October 2026] &Truth Emerges From Stealth, Pioneering the Universal Verification Machine for Code | https://www.businesswire.com/news/home/20261029287395/en/Truth-Emerges-From-Stealth-Pioneering-the-Universal-Verification-Machine-for-Code
- [Bloomberg, March 2018] Micro Focus Shares Collapse After Sales Warning and CEO Exit | https://www.bloomberg.com/news/articles/2018-03-19/micro-focus-issues-sales-warning-alongside-ceo-resignation