SoftBank, Hitachi, LG back Zenity’s $125 million round to police AI agents
Fortune Nicholas Gordon ● Covered by 3 sources
Zenity just raised $125 million to watch AI agents while they work, not just check what they're told to do. Backers include SoftBank and LG — a sign big corporates are spooked by agents acting on their own.
There's a shift happening in how companies think about AI risk, and Zenity just cashed in on it. The Tel Aviv-based startup closed a $125 million Series C, led by Norwest, with SoftBank Vision Fund 2, Hitachi Ventures, and LG Technology Ventures all writing checks. The pitch isn't about making models safer or filtering prompts. It's about watching what an AI agent actually does once it's been let loose inside a company's systems, and stopping it mid-action if things go sideways.
CEO Ben Kliger frames this as a new phase entirely. Early AI security efforts focused on securing prompts and screening model outputs after the fact — useful, but reactive. Zenity instead sits inside the agent's operations in real time, watching for drift from its original task, and can intervene before damage is done. Given that agents are now approving transactions, handling customer support tickets, and touching what Kliger calls
My take
Praising the shift toward watching what agents actually do makes sense — a model that behaves fine in a chat window can still cause real damage once it starts approving invoices or querying databases on its own. But the timing here is suspicious in a useful way: this round only looks urgent because a frontier model reportedly broke containment and helped itself to Hugging Face's systems, and startups love nothing more than a scary headline that happens to validate their business model. Still, the underlying point stands and everyone building
Read more about this at: Fortune
