Harmonic Expands Aristotle to C and Verilog After Raising $295 Million
Updated
Updated · WIRED · Oct 29
Harmonic Expands Aristotle to C and Verilog After Raising $295 Million
2 articles · Updated · WIRED · Oct 29
Summary
Harmonic said this week its AI agent Aristotle can now verify software written in C and hardware designs in Verilog, widening its push to automate formal verification.
The move comes as startups argue generative AI can slash verification costs and timelines—Stanford researcher Chuyue Sun reported one project finished in under half a day for less than $500, versus five months by three experts.
Harmonic is testing Aristotle with potential users including Tensordyne and Nvidia, whose venture arm backs the company, while rival &Truth launched publicly after raising $70 million.
Formal verification already protects security-critical code in systems from Boeing planes to more than 2.5 billion Apple devices, but adoption remains limited by scarce specialists, high costs, and fears of slowing software releases.
Skeptics say mathematically verified code can still miss failures tied to user behavior or hardware, leaving open whether AI can push the technique beyond niche, high-stakes systems.
As &Truth and Harmonic raise huge funding, can AI-checked proofs beat traditional testing for real-world software and hardware security?
If AI can verify encryption code in hours for under $500, what still prevents formal verification from becoming standard software practice?
Did Chuyue Sun’s Signal-related breakthrough solve formal verification’s biggest bottleneck—or just move the trust problem to specifications and proof tools?