Code Metal Raises $125M for AI-Powered Hardware Management
The startup develops software to manage hardware using artificial intelligence. The round was led by Salesforce Ventures.
Code Metal and $125M for Code Verification: Why "Writing Code" Is No Longer Enough
The Core: Not Hardware Management, but Mathematically Proven Safety
When Code Metal announced the closing of a $125 million Series B round at a $1.25 billion valuation on February 19, 2026, headlines buzzed with phrases like "AI-powered hardware management." That's a superficial reading. In reality, Code Metal is tackling a fundamentally different challenge—not code generation, but mathematically provable verification for environments where a mistake costs human lives.
The key term here is "neurosymbolic approach." The platform doesn't just generate code using a large language model. It breaks programs into testable symbolic components and mathematically proves that the generated code is functionally equivalent to the original and contains no security violations. This is not "vibe coding," where a developer trusts the model's intuition. It's formal verification—a discipline that has existed in academia for decades and is now, for the first time, embedded in a commercial product with such capitalization.
The market Code Metal is targeting was articulated by Salesforce Ventures partner Rob Keith: "Mission-critical industries cannot deploy what they cannot verify." That sentence encapsulates the entire business logic of the company.
Timeline and Context: From F-35 to $1.25 Billion in Two and a Half Years
Code Metal's trajectory deserves close scrutiny—it explains why the company raised $125 million just months after its Series A.
The founders—Peter Morales (CEO) and Alex Showalter-Bucher (CTO)—are alumni of MIT Lincoln Laboratory, where they developed AI reasoning systems for the F-35 fighter jet. These are not startup founders who spotted an opportunity in the AI hype. They are engineers who spent decades solving the problem of code porting and verification for weapons systems. They know the pain firsthand: when satellite communication protocols need to be updated or ported to new hardware, when legacy C++ systems need to be modernized into memory-safe Rust—this work is painfully manual, slow, and error-prone.
The company was founded in 2023. In July 2024, it raised a $16.5 million seed round. Then came the Series A—and almost immediately, a $125 million Series B led by Salesforce Ventures, with participation from Accel, B Capital, Smith Point Capital, J2 Ventures, Shield Capital, Overmatch, and RTX. The investor lineup is telling: RTX (formerly Raytheon) is not just a venture fund but a direct representative of the defense sector. Salesforce Ventures and B Capital are strategic investors with access to enterprise and government sectors.
A notable personnel move: Ryan Aytay, former CEO of Tableau, joined as President and COO. This signals that the company is preparing for aggressive scaling of commercial operations and enterprise sales. A person who managed one of the largest SaaS businesses doesn't join a startup just for a title.
Who Wins and Who Loses
The U.S. defense sector and its contractors win. Code Metal's client list already includes the U.S. Air Force, L3Harris, RTX, and Toshiba. These are not pilot projects—they are "programs of record," i.e., active defense programs. For the Pentagon, which recently stated that "developing capabilities using AI will redefine the nature of military operations in the next decade," Code Metal becomes load-bearing infrastructure.
Specialized chip manufacturers win. As computing shifts to the edge and specialized hardware, the problem of code portability across architectures intensifies. Code Metal ensures this portability while optimizing code for specific hardware. Chipmakers gain a tool that lowers barriers to migration onto their platforms.
Traditional RPA solutions and "ordinary" AI coding assistants lose. GitHub Copilot, Cursor, and the like are good for web development and enterprise software. In environments with zero tolerance for errors, they are useless—a generated hallucinated function or an unnoticed edge case could cost national security or human lives. Code Metal doesn't compete with Copilot; it creates an entirely different category.
Engineers whose job is manual code porting lose. If the company automates translation between programming languages with mathematical proof of correctness, the need for specialists who manually rewrite C++ into Rust decreases.
What the Media Misses
Insight One: Code Metal is not a software startup in the classic sense. It is a trust infrastructure for a future where machines write code.
The term "neurosymbolic approach" sounds like marketing. But behind it lies a specific architecture: generative AI combined with formal verification. Instead of predicting the next token, the system breaks a program into provable components. This is not an improved Copilot—it is a fundamentally different class of products. If AI generates more and more code for critical infrastructure, Code Metal becomes not a tool but a checkpoint—without its verification, code won't reach production.
Insight Two: Code Metal's market niche is not "hardware management" but the gap between the speed of AI development and security requirements.
Currently, this gap is closed by manual processes: code review, testing, formal methods by expensive specialists. Code Metal automates formal methods—an area that has always been considered too complex and expensive for mass adoption. If they have succeeded, then $1.25 billion is a fair valuation. If not, the company will collapse at the first audits against standards like DO-178C (aviation software certification).
Insight Three: Salesforce Ventures is leading the round not just for financial return—it's a strategic bet on controlling the infrastructure layer.
Salesforce is an enterprise company whose clients increasingly demand security and compliance guarantees for AI-generated code. By placing their person (Ryan Aytay, former CEO of Tableau, who worked at Salesforce for many years) as President and COO of Code Metal, they gain influence over a verification standard that could become industry-wide.
Forecast: Next 30 Days and 90 Days
30 days (through mid-March 2026). Expect an announcement of a new major defense contract. $125 million is money for scaling, and the company will use the coming weeks to strengthen its position in Washington. Also likely are 2-3 key hires in the enterprise sales team. Aytay will begin building partnerships with system integrators—without them, you can't enter the defense sector.
90 days (through mid-May 2026). The key indicator is expansion beyond the defense sector. Automotive, semiconductors, medical devices—all are stated verticals. If Code Metal announces a contract with a major automotive company or medical device manufacturer, it will confirm the platform's versatility. If not, the company risks remaining a niche defense contractor with an unjustifiably high valuation.
The defense angle is also important: with the Trump-Vance administration and the focus on DOGE (Department of Government Efficiency), budgets for modern defense technologies remain a priority. Code Metal fits perfectly into this trend but simultaneously becomes vulnerable to political shifts.
Bottom line. Code Metal is not about "AI for hardware." It's about who will certify code written by machines in a world where such code controls fighter jets, satellites, and drones. The $125 million bet is a bet that the future will require mathematical guarantees, not just fast generation. If the bet pays off, Code Metal will become what VeriSign became for the internet—invisible but indispensable trust infrastructure.
— Editorial Team
No comments yet.