Cloudflare Just Made Shadow AI Spending a Thing of the Past

Every company has that one question nobody can answer: who’s actually using AI, and what is it costing us? Cloudflare just built the tool that answers it. On August 6, 2026, the connectivity cloud giant unveiled Identity-Aware AI Gateway alongside a companion feature called User Insights — giving IT and security teams their clearest window yet into enterprise AI usage. Alongside it, Cloudflare also announced Cloudflare OS, an open-source AI workspace built to run entire companies on.

Together, these launches tackle two headaches growing in lockstep: runaway AI costs and invisible security risks.

What Cloudflare Announced

Why This Matters Right Now

AI adoption inside companies has outpaced the tools meant to manage it. Individual AI requests are getting cheaper, but employees and automated bots are sending so many more of them that bills spike with zero explanation. Worse, sensitive data — employee names, passwords, confidential files — can quietly leak to outside AI providers without anyone noticing. Existing tools can cap total spending, but they can’t tell you who caused it or what got sent.

As Cloudflare CTO Dane Knecht put it, clunky security becomes “a speed bump that slows down innovation,” pushing companies to just lock everything down instead of enabling safe AI use.

Identity-Aware AI Gateway, Explained

This is the core fix: every AI request now carries a verified identity, whether from a person or an automated system. By connecting AI Gateway with Cloudflare Access, IT teams can finally see exactly who sent what, set spending limits per person or team, and automatically flag risky content — all without disrupting daily workflows or building new systems from scratch.

User Insights, Explained

Where the Gateway tracks identity, User Insights tracks behavior. It builds a usage baseline for every employee and system, then alerts IT the moment spending or activity spikes beyond what’s normal. It also flags a subtler waste: employees using expensive, high-powered AI models for tasks that a cheaper, lighter model could easily handle — surfacing savings that are nearly impossible to spot manually at scale.

Cloudflare OS, Explained

This is the bigger swing. Most enterprise AI tools know a lot about the world but nothing about how a specific company actually operates — its internal systems, approval chains, or daily workflows. Cloudflare OS flips that: it’s the same platform Cloudflare uses internally, now open-sourced, giving every employee a secure AI-powered workspace connected to real company systems from day one.

Key capabilities include an agent workspace usable without coding skills, the ability for anyone to turn AI output into a shareable working app, Zero Trust security by default (AI agents start with zero permissions), and full model flexibility through AI Gateway so companies aren’t locked into one AI vendor.

CEO Matthew Prince summed it up simply: “Cloudflare OS is how we run Cloudflare… We built this because nothing else did what we needed. Now any company can start from where it took us years to get.”

The Bigger Picture

Both launches point to the same underlying shift: enterprises don’t need less AI access, they need smarter visibility into how it’s used. Rather than restricting adoption out of fear, Cloudflare is betting that identity-based controls and open, self-owned infrastructure let companies move faster and safer. Given Cloudflare’s role as one of the internet’s largest connectivity networks, its approach to AI governance is likely to influence how enterprise IT handles AI adoption going forward.

For more breaking tech news and enterprise AI updates, check out the latest coverage on technosports.co.in.

FAQs