When you build AI for memes, you can afford vibes. When you build AI for missile systems, you better bring receipts. Code Metal just secured $125M in Series B funding at a $1.25B valuation, led by Salesforce Ventures with Accel, B Capital, RTX, Shield Capital, J2 Ventures, Overmatch, and Smith Point Capital in the mix. That is not tourist capital. That is “we’ve read the contracts” capital.
Congratulations to Peter Morales, CEO and Founder, and Alex Showalter-Bucher, SVP of Technology and Founder, for turning formal methods into something venture investors actually line up for. And credit to Ryan Aytay, President and COO, plus Laura Shen, Senior Vice President of Growth, for building the kind of commercial engine that can translate deep tech into 8-figure contracts without losing its nerve.
Let’s talk about the name. Code Metal. Not Code Marshmallow. Not Code Maybe. Metal is forged under heat. It holds shape under pressure. That is the thesis. AI-generated code is cute until it flies a drone, drives a vehicle, or sits inside a semiconductor pipeline. In defense, automotive, and industrial systems, “close enough” is how you end up on a congressional hearing livestream.
Code Metal’s platform focuses on verifiable AI-powered code translation and optimization. Move between programming languages. Optimize across chips. Validate every line so it survives compliance reviews and real-world stress. Formal methods plus AI. That combination is not flashy, but it is lethal in the best possible way. It closes the trust gap that has kept mission-critical industries skeptical of generative code tools.
Customers like Toshiba, RTX, L3Harris, and even the U.S. Air Force are not experimenting for sport. They are modernizing legacy systems, reducing hardware lock-in, and accelerating deployment in environments where failure is not an option. Code that can migrate across SDKs and operating systems without breaking under pressure is not a feature. It is leverage.
Here is the business lesson hiding in plain sight. Code Metal did not chase the loudest use case. It chased the hardest one. Instead of competing for prompt engineers building landing pages, they went after regulated industries where verification is oxygen. That focus creates pricing power, long-term contracts, and investors who understand that durable revenue beats viral demos.
Boston roots at MassRobotics. Deep MIT Lincoln Laboratory DNA. 8-figure revenue contracts already in motion. This is what happens when you pair research-grade rigor with operators who know how to scale.
