Flexxon wins Singapore funding for a hardware gateway to police AI agents

Software that supervises AI agents can be switched off by anyone who takes over the machine it runs on. Flexxon wants to move that supervision into hardware.

The Singapore company has been named an awardee of the Cybersecurity Industry Call for Innovation (CyberCall) 2025, run by the CyberSG TIG Centre, a joint initiative of the National University of Singapore and the Cyber Security Agency of Singapore. The award, made under the Cybersecurity for AI challenge, was announced at CyberSG Innovation Day on 10 November. No funding figure was disclosed.

The project, X-PHY AI Armor, will be a hardware-rooted Model Context Protocol (MCP) gateway. It will sit between an AI agent and the systems it reaches, inspect each request for injected instructions and hijacked tools, and hand anything high-risk to the hardware to lock down. Authorisation for those operations will be verified by the device, and the audit record written to it, so an attacker who controls the host cannot alter the decision or erase the trace. Flexxon contrasts that with generic API gateways, which work at the network layer and can be reconfigured by whoever holds the host.

The risk it targets is prompt injection. An agent holds credentials and acts without a person checking each step, so one poisoned line in a document, web page or another agent's reply can redirect it while every action stays technically authorised.

Flexxon cites Gartner's forecast that 40% of enterprise applications will include task-specific AI agents by the end of 2026, up from under 5% a year earlier. It also points to the CSA's Singapore Cybersecurity Health Report 2025, which surveyed 2,005 businesses and non-profits and found almost nine in 10 had met at least one cybersecurity incident during the year.

This is Flexxon's third CyberCall award. It was one of eight selected from 72 submissions in 2019 for its X-PHY solid state drive, and an awardee in 2021 for a server security module.

Camellia Chan, Flexxon's co-founder and CEO, said agents are useful because they act without waiting for a person, which is also what makes them dangerous. Willis Lim, executive director of the TIG Centre, said securing AI agents is the kind of problem CyberCall was set up to address.

To stay across the latest in cloud, AI and enterprise tech analysis from Compare the Cloud, subscribe to our weekly newsletter at https://www.comparethecloud.net/newsletter

More News