Fastly puts AI traffic under tighter control

Fastly has rolled out new tools designed to help businesses monitor, secure and control artificial intelligence (AI) traffic as companies move AI systems into production.

The edge cloud company on Wednesday launched AI Runtime Control, AI Firewall and expanded API Security capabilities under its Fastly for AI portfolio.

The tools are designed to govern how applications access AI models, protect AI systems from attacks such as prompt injection and control how autonomous agents interact with enterprise APIs.

Fastly said the expansion comes as machine-generated traffic continues to surge across its network. AI traffic grew 6.5 times faster than human traffic from January to May, while machine-generated requests accounted for more than half of network traffic in July and August.

“Enterprises need control in production, at runtime, without delay, friction, or disruption,” Fastly chief product officer Kelly Shortridge said.

AI Runtime Control allows companies to route model requests through a single endpoint across different providers while monitoring token spending, enforcing budgets and setting failover policies.

AI Firewall, meanwhile, screens prompts in real time to detect and block malicious activity before requests reach AI models.

Fastly’s expanded API Security tools are designed to prevent AI agents from making unsupported or malformed requests or operating outside defined API boundaries.

The company said the tools run on the same edge platform it uses for content delivery and cybersecurity services.

Fastly said its global edge network had 622 terabits per second of capacity as of 30 June and processes more than five trillion requests daily.