Ryan Aytay is president and COO of Code Metal, which translates, verifies and optimizes code from any source to any hardware using formal methods, so mission-critical systems are provably correct. Before Code Metal, he ran Tableau as CEO for three to four years, after about fifteen years at Salesforce going back to 2006 or 2007. He joined us in Episode 48 (June 26, 2026), about five and a half months into the job.
The name is the thesis: lower the barrier from code to metal, software to hardware (03:41). Ryan's examples of what runs on code were the military, cars, aerospace, the power grid and medical devices like MRI machines. Much of that code is legacy, written years ago and hard to change safely.
Code Metal takes that legacy code, translates it to a new language or target hardware, verifies with formal methods that it keeps the same intent and behavior, and optimizes it for the device. "Provably correct," he said, are the two words he likes to use. The company was founded about two years before the episode by CEO Peter Morales, who started at MIT Lincoln Lab and worked on F-35 simulation, with co-founder Alex.
Ryan walked us through slides rather than a live product run. At 08:29 he showed an early look at a platform Code Metal had not fully launched yet: a code-in, code-out view and a test generation suite. The idea is to look back at the original code's intent, behavior and origin, then use formal methods, testing and fuzzing to prove the new version is right and will run on the target hardware. We didn't see a translation run end to end, so we can't speak to speed or output quality ourselves.
We are all about, you know, we translate, we verify, and we also optimize to make sure that we're doing it in the most efficient way because we know that in this world, there are certain systems that just can't afford to fail.
— Ryan Aytay, Episode 48
But the reality is, like, in a production world, AI generated code actually isn't quite safe enough and really never will be for what we'll call a mission critical system at the edge.
— Ryan Aytay, Episode 48
So we have this belief, right, where, you know, generating code actually, the price is going down. It doesn't really cost anything, but if you get it wrong, it's pretty substantial.
— Ryan Aytay, Episode 48
Jordan admitted he has spent his career avoiding mission-critical software, sticking to web SaaS and mobile apps. In the news segment, Ryan talked about wearing early Meta Ray-Bans in Yosemite and was cautiously open on prediction markets, which Sam tied back to our Kalshi prediction market trading bot.
For more on the industries Code Metal works in, see AI in aviation, AI in automotive and AI in government.
Ryan Aytay is president and COO of Code Metal. Before joining about five and a half months before his June 2026 appearance, he ran Tableau as CEO for three to four years after roughly fifteen years at Salesforce.
Code Metal translates code from any source to any hardware, verifies it with formal methods so it keeps the same intent and behavior, and optimizes it for the target device. It focuses on mission-critical systems in defense, aerospace, automotive, medical devices and the power grid.
Ryan's view is no, not on its own. He said AI-generated code is mostly correct, but mission-critical systems at the edge need code that is provably correct, which takes formal verification, testing and physical testing on the hardware.
Yes. Ryan said Code Metal can modernize a million lines of C++ to Rust quickly and guarantee it works on the target device. He said harder pairs, like MATLAB to VHDL, are also possible.
War Matrix is a modeling and simulation platform Code Metal built with the military. It brings siloed data together and runs war-gaming scenarios on a simulation engine.