Cybersecurity’s biggest asymmetry right now is the one between machines that attack and machines that defend. A six-week-old Tel Aviv startup just raised $60M from Sequoia Capital, Khosla Ventures and Coatue on the bet that defense needs its own frontier models.
Corma runs simulations that lay the gap bare: leading AI models asked to breach modeled Fortune 500 networks succeeded 88% of the time, while the same models asked to spot the intrusions caught only 12% of threats. General-purpose systems trained on code, the company argues, transfer well to finding flaws and writing exploits but poorly to the messy job of defending a large corporate network.
Its answer is a foundation model built specifically for defensive security, powering AI agents that work across functions rather than sitting on top of another dashboard. Corma says deployments inside Fortune 100 and Fortune 500 companies in healthcare, finance, energy and retail have cut threat response times by more than 94%, though those figures have not been independently verified.
The founding team mixes researchers from Google and DeepMind with veterans of Israel’s Unit 8200. “The race to general intelligence in cybersecurity has already begun, and the attackers have a significant head start,” said co-founder and CEO Alon Pluda.
The size of a seed round for a company this young shows how worried investors are that autonomous attackers will outpace human defense teams.
