Akamai CEO Tom Leighton announced the company is developing sandbox protections and API security measures to manage risks associated with AI agents. The initiative aims to apply existing acceptable use policies to AI customers to prevent fraud and data exfiltration while maintaining human oversight.
The shift toward autonomous AI agents requires a fundamental hardening of the delivery stack to prevent automated fraud and unauthorized data access. By implementing sandbox protections at the infrastructure level, Akamai aims to contain agent behavior before it reaches critical internal systems. This move signals a broader industry transition where CDNs must evolve from simple content delivery into active gatekeepers for machine-to-machine traffic. As streaming providers increasingly use AI for personalization and encoding, these guardrails will become essential for maintaining platform integrity. Watch for the release of specific API security benchmarks to see how these protections impact latency for real-time AI applications.
Akamai has been expanding its security portfolio beyond traditional DDoS mitigation to address the specific threats posed by autonomous AI agents operating across distributed infrastructure. The company's 2025 State of the Internet report documented a 400 percent year-over-year increase in API attacks targeting AI endpoints, with credential stuffing and bot-driven scraping representing the fastest-growing vectors. That data informed the sandbox architecture Leighton described, which isolates agent workloads at the edge before they reach origin systems. Akamai's approach positions it against Cloudflare, which launched its own AI Gateway product in early 2026 to provide rate limiting, caching, and observability for LLM API calls, and against Fastly, which has focused on programmable edge security through its Compute platform.
The business implications for streaming and media companies are significant because CDN providers are increasingly the enforcement layer for content access policies. Akamai reported $1.04 billion in security revenue for Q2 2025, representing 37 percent of total revenue and growing 12 percent year over year, signaling that security has become a primary growth engine alongside delivery. For streaming platforms that rely on Akamai for both CDN and edge security, the new AI agent protections could consolidate vendor relationships rather than requiring separate API security tools. The company's existing Media and Entertainment segment already serves major studios and OTT platforms, meaning these sandbox protections will likely reach streaming workloads first.
Competing CDN and edge providers are pursuing similar AI security strategies, creating a market where streaming buyers must evaluate which vendor's agent governance model fits their architecture. Cloudflare announced in August 2026 that its Workers AI platform now processes over 1 billion inference requests daily, giving it visibility into agent traffic patterns that inform its security rules. Meanwhile, Fastly acquired security startup Signal Sciences in a deal that expanded its edge security capabilities for API protection, and the combined platform now offers runtime application self-protection that could extend to AI agent traffic. For video infrastructure teams, the key evaluation criteria are whether these protections add measurable latency to real-time streaming pipelines and whether they support the specific API patterns used by encoding, packaging, and personalization agents.
Akamai CEO Tom Leighton has introduced new security measures to address risks posed by autonomous AI agents. By implementing sandbox protections and API firewalls at the infrastructure level, Akamai aims to prevent data exfiltration and fraud. This shift is critical as CDNs evolve into active gatekeepers for machine-to-machine traffic.
Akamai is developing sandbox protections to restrict AI agents to specific systems, alongside enhanced API firewalls and system segmentation to prevent data exfiltration.
Yes, human oversight remains a mandatory requirement for all AI agents operating on the Akamai platform.
Akamai will apply its existing acceptable use policies to AI customers to block applications that cause harm, violate local laws, or engage in fraudulent activities.
As streaming providers increasingly use AI for personalization and encoding, these guardrails are essential for maintaining platform integrity and preventing unauthorized data access.
Add StreamingMeme as a preferred source on Google to see more of our streaming news at the top of your Search results.
Add as preferred source
