As the software development landscape evolves with the rise of artificial intelligence, one small startup is placing a bold bet on the future of code verification. Theorem, based in San Francisco and a product of Y Combinator’s Spring 2025 cohort, has announced a significant milestone with a $6 million seed funding round led by Khosla Ventures. The company’s mission is to address the growing challenge of ensuring the reliability of AI-generated software, a critical need as AI coding assistants from major tech companies like GitHub, Amazon, and Google are churning out billions of lines of code annually.
Theorem’s co-founder, Jason Gross, highlights the urgency of the situation, noting that the pace at which AI is generating code is surpassing human capacity for thorough review. The company’s innovative approach combines formal verification, a rigorous mathematical technique for verifying software behavior, with AI models that can automatically generate and validate proofs. This blend of technologies streamlines a process that traditionally demanded extensive expertise and time, making it accessible for mainstream software development.
Formal verification has long been reserved for high-stakes applications like avionics and cryptography due to its complexity and cost. Gross, drawing from his experience in cryptography research at MIT, emphasizes the transformative potential of automating this process with AI. Theorem’s system employs a method called “fractional proof decomposition” to allocate verification resources efficiently, catching bugs that traditional testing methods might miss.
The startup’s success stories include a case where they translated a 1,500-page specification into 16,000 lines of trustworthy code, enabling a significant performance boost for a client without manual review. This demonstration underscores the practical value of Theorem’s technology in enhancing software reliability and efficiency across various industries, from AI research labs to electronic design automation.
With the increasing reliance on AI systems in critical infrastructure, the need for robust software verification is more pressing than ever. Gross warns of the security risks posed by unchecked AI-generated code and advocates for a proactive approach to ensure system integrity through formal verification. Theorem stands out in a crowded field of AI code verification startups by prioritizing scalability and practical application in real-world software development scenarios.
Looking ahead, Theorem plans to leverage its recent funding to expand its team and venture into new sectors such as robotics, renewable energy, and cryptocurrency. The company’s vision aligns with a paradigm shift in how organizations approach AI-assisted development, emphasizing the importance of safety and reliability over sheer speed and productivity. As AI continues to advance at an unprecedented rate, Theorem’s focus on rigorous oversight and verification represents a crucial step towards ensuring that humans retain control over the systems they create.
