~/AI AGENTS/cloudflare-reports-ai-agent-traffic-surges-1-700-yoy-reaching-60-of

Cloudflare Reports AI Agent Traffic Surges 1,700% YoY, Reaching 60% of Web Traffic

Cloudflare's Goran Risticsevich announced that internet traffic generated by AI agents has surged by over 1,700% year-over-year, now accounting for approximately 60% of all web traffic. This rapid growth has introduced new security risks, including "Shadow AI" and "Shadow MCP" (Model Context Protocol) vulnerabilities. The shift from human-driven to machine-driven web traffic significantly expands the digital attack surface, as AI agents can query thousands of servers simultaneously. Organizations must rapidly adapt their security frameworks to defend against unauthorized AI tools and unmonitored protocol integrations. Cloudflare highlighted "Shadow MCP" as a rising threat where unauthorized Model Context Protocol connections bypass enterprise controls to link AI applications to internal databases. Additionally, the use of unapproved AI tools by employees ("Shadow AI") risks exposing sensitive internal data to external models.

## BACKGROUND

The Model Context Protocol (MCP) is an open-source standard introduced by Anthropic in late 2024 to standardize how large language models integrate with external data sources and tools. "Shadow MCP" refers to the unauthorized deployment of these protocol servers within an organization, similar to how "Shadow IT" describes unauthorized software usage.

## REFERENCES

## KEYWORDS

#AI Agents#Internet Traffic#Network Security#Cloudflare

$ subscribe --daily

Cloudflare Reports AI Agent Traffic Surges 1,700% YoY, Reaching 60% of Web Traffic | Daily News