AI Agents Are Creating a New Layer of Cybersecurity Spending

AI-focused cloud service provider Akamai launched a sandbox for AI agents on Tuesday, a sealed environment where code an agent writes can run without touching the rest of a company's systems. Agent Sandbox is in beta, alongside Akamai Inference, a service for running open AI models on Akamai's cloud.

The launch may be small but the problem behind it is not. Software is no longer just answering questions—it logs into accounts, writes code, moves money and acts on its own, and every one of those powers is something a hacker, or the software itself, can abuse.

This is what’s known as agentic AI, and it’s taking over the internet. A 2026 Palo Alto Networks survey of 2,930 security leaders found an average of 109 machine identities for every human, the logins and keys software uses to access other software. Of those, 79 are AI agents.

Agentic AI comes with a lot of perks, but also some serious dangers, especially when those agents have too much freedom, or are so intelligent they can bypass their own restrictions.

For example, Hugging Face, a major hub for open AI models, disclosed an intrusion on July 16, and OpenAI confirmed five days later that its own models were responsible. The research group METR later investigated the incident and found about 1,200 OpenAI agents had coordinated on a hidden message board, with roughly 700 joining the attack.

Turns out the agents escaped their sandbox and chained a previously unknown software flaw with stolen credentials to reach live systems. OpenAI called it a "warning shot." It said a monitor that reads a model's written reasoning was not running during the tests.

In simple terms, AI safety usually ends up being a tug of war game between alignment (the guys in charge of making sure AI is safe) and development (the guys in charge of making sure AI is powerful), with each team wanting the other to play a minor role in order to secure its own goals.

That is what Akamai wants to treat. “Enterprises should not have to choose between speed and control when they put AI into production,” Jon Alexander, Senior Vice President of Cloud Technology at Akamai said in a press release.

“Some teams want to ship an AI feature this quarter on a managed platform. Others need dedicated GPUs to run frontier-class open models on their own terms. Every team running agents needs a boundary around the code those agents write,” he said.

Hijacking is another fear. In July, Zscaler found web pages hiding instructions meant for AI agents, a trick called indirect prompt injection. One fake documentation page told agents to pay a $3 "license fee" in about 0.0012 Ethereum to a hardcoded wallet, and four of the 26 models Zscaler tested complied.

Back in May cybersecurity firm Bugcrowd saw submissions more than quadruple in three weeks in March, most of them fake. HackerOne and Nextcloud suspended paid programs in April.

The machines also find real bugs. Mozilla said an early version of Anthropic's Claude Mythos helped find 271 vulnerabilities in Firefox, all patched within the week. That flood of real and fake findings is what security companies are now racing to manage.

Palo Alto closed its CyberArk purchase on Feb. 11 to extend privileged-access controls to AI agents. In April it bought Koi, which secures AI coding agents running on employee machines. Cisco announced a planned acquisition of Astrix Security in May, a firm that governs the keys and tokens agents use, and completed the deal June 29.

Akamai's piece is the fence around the code itself. Its sandbox uses WebAssembly, a format that runs programs in a tightly walled box, to cap an agent's computing power, memory and outbound connections. Akamai's Alexander said every team running agents "needs a boundary around the code those agents write."

Palo Alto's survey found organizations expect AI agents to grow 85% over the next 12 months, against 56% for human identities.