As synthetic intelligence reshapes software program improvement, a small startup is betting that the trade’s subsequent huge bottleneck will not be writing code — it is going to be trusting it.
Theorem, a San Francisco-based firm that emerged from Y Combinator’s Spring 2025 batch, introduced Tuesday it has raised $6 million in seed funding to construct automated instruments that confirm the correctness of AI-generated software program. Khosla Ventures led the spherical, with participation from Y Combinator, e14, SAIF, Halcyon, and angel traders together with Blake Borgesson, co-founder of Recursion Prescribed drugs, and Arthur Breitman, co-founder of blockchain platform Tezos.
The funding arrives at a pivotal second. AI coding assistants from firms like GitHub, Amazon, and Google now generate billions of strains of code yearly. Enterprise adoption is accelerating. However the skill to confirm that AI-written software program truly works as supposed has not saved tempo — creating what Theorem’s founders describe as a widening “oversight hole” that threatens essential infrastructure from monetary methods to energy grids.
“We’re already there,” stated Jason Gross, Theorem’s co-founder, once we requested whether or not AI-generated code is outpacing human evaluate capability. “When you requested me to evaluate 60,000 strains of code, I would not know how you can do it.”
Why AI is writing code sooner than people can confirm it
Theorem’s core know-how combines formal verification — a mathematical method that proves software program behaves precisely as specified — with AI fashions skilled to generate and examine proofs routinely. The method transforms a course of that traditionally required years of PhD-level engineering into one thing the corporate claims will be accomplished in weeks and even days.
Formal verification has existed for many years however remained confined to essentially the most mission-critical purposes: avionics methods, nuclear reactor controls, and cryptographic protocols. The method’s prohibitive price — typically requiring eight strains of mathematical proof for each single line of code — made it impractical for mainstream software program improvement.
Gross is aware of this firsthand. Earlier than founding Theorem, he earned his PhD at MIT engaged on verified cryptography code that now powers the HTTPS safety protocol defending trillions of web connections each day. That venture, by his estimate, consumed fifteen person-years of labor.
“No person prefers to have incorrect code,” Gross stated. “Software program verification has simply not been economical earlier than. Proofs was written by PhD-level engineers. Now, AI writes all of it.”
How formal verification catches the bugs that conventional testing misses
Theorem’s system operates on a precept Gross calls “fractional proof decomposition.” Moderately than exhaustively testing each attainable conduct — computationally infeasible for advanced software program — the know-how allocates verification assets proportionally to the significance of every code element.
The method just lately recognized a bug that slipped previous testing at Anthropic, the AI security firm behind the Claude chatbot. Gross stated the method helps builders “catch their bugs now with out expending a number of compute.”
In a current technical demonstration referred to as SFBench, Theorem used AI to translate 1,276 issues from Rocq (a proper proof assistant) to Lean (one other verification language), then routinely proved every translation equal to the unique. The corporate estimates a human workforce would have required roughly 2.7 person-years to finish the identical work.
“Everybody can run brokers in parallel, however we’re additionally in a position to run them sequentially,” Gross defined, noting that Theorem’s structure handles interdependent code — the place options construct on one another throughout dozens of information — that journeys up typical AI coding brokers restricted by context home windows.
How one firm turned a 1,500-page specification into 16,000 strains of trusted code
The startup is already working with prospects in AI analysis labs, digital design automation, and GPU-accelerated computing. One case research illustrates the know-how’s sensible worth.
A buyer got here to Theorem with a 1,500-page PDF specification and a legacy software program implementation stricken by reminiscence leaks, crashes, and different elusive bugs. Their most pressing downside: bettering efficiency from 10 megabits per second to 1 gigabit per second — a 100-fold enhance — with out introducing further errors.
Theorem’s system generated 16,000 strains of manufacturing code, which the shopper deployed with out ever manually reviewing it. The boldness got here from a compact executable specification — a couple of hundred strains that generalized the huge PDF doc — paired with an equivalence-checking harness that verified the brand new implementation matched the supposed conduct.
“Now they’ve a production-grade parser working at 1 Gbps that they’ll deploy with the boldness that no info is misplaced throughout parsing,” Gross stated.
The safety dangers lurking in AI-generated software program for essential infrastructure
The funding announcement arrives as policymakers and technologists more and more scrutinize the reliability of AI methods embedded in essential infrastructure. Software program already controls monetary markets, medical units, transportation networks, and electrical grids. AI is accelerating how rapidly that software program evolves — and the way simply delicate bugs can propagate.
Gross frames the problem in safety phrases. As AI makes it cheaper to seek out and exploit vulnerabilities, defenders want what he calls “uneven protection” — safety that scales with out proportional will increase in assets.
“Software program safety is a fragile offense-defense steadiness,” he stated. “With AI hacking, the price of hacking a system is falling sharply. The one viable answer is uneven protection. If we would like a software program safety answer that may final for quite a lot of generations of mannequin enhancements, it is going to be by way of verification.”
Requested whether or not regulators ought to mandate formal verification for AI-generated code in essential methods, Gross provided a pointed response: “Now that formal verification is affordable sufficient, it is likely to be thought of gross negligence to not use it for ensures about essential methods.”
What separates Theorem from different AI code verification startups
Theorem enters a market the place quite a few startups and analysis labs are exploring the intersection of AI and formal verification. The corporate’s differentiation, Gross argues, lies in its singular deal with scaling software program oversight quite than making use of verification to arithmetic or different domains.
“Our instruments are helpful for methods engineering groups, working near the steel, who want correctness ensures earlier than merging modifications,” he stated.
The founding workforce displays that technical orientation. Gross brings deep experience in programming language principle and a observe document of deploying verified code into manufacturing at scale. Co-founder Rajashree Agrawal, a machine studying analysis engineer, focuses on coaching the AI fashions that energy the verification pipeline.
“We’re engaged on formal program reasoning so that everybody can oversee not simply the work of a median software-engineer-level AI, however actually harness the capabilities of a Linus Torvalds-level AI,” Agrawal stated, referencing the legendary creator of Linux.
The race to confirm AI code earlier than it controls all the pieces
Theorem plans to make use of the funding to increase its workforce, enhance compute assets for coaching verification fashions, and push into new industries together with robotics, renewable vitality, cryptocurrency, and drug synthesis. The corporate at present employs 4 folks.
The startup’s emergence alerts a shift in how enterprise know-how leaders might have to guage AI coding instruments. The primary wave of AI-assisted improvement promised productiveness positive factors — extra code, sooner. Theorem is wagering that the subsequent wave will demand one thing totally different: mathematical proof that velocity does not come at the price of security.
Gross frames the stakes in stark phrases. AI methods are bettering exponentially. If that trajectory holds, he believes superhuman software program engineering is inevitable — able to designing methods extra advanced than something people have ever constructed.
“And with out a radically totally different economics of oversight,” he stated, “we’ll find yourself deploying methods we do not management.”
The machines are writing the code. Now somebody has to examine their work.