Cloudflare and OpenAI Launch AI Search Real-Time Indexing Pilot
Claire Weston
Cloudflare and OpenAI announced a research pilot on July 8 to use Cloudflare's real-time network signals — drawn from over 20% of global internet traffic — to improve AI search indexing. This marks the first infrastructure-layer partnership pushing AI search from crawl-and-index toward signal-driven, real-time discovery.
What exactly are they building together?
Cloudflare will share real-time network signals — content freshness, traffic quality, and actual page changes — while OpenAI contributes frontier models and real user queries for testing.
This means → AI search no longer relies on crawlers visiting pages one by one; instead, the infrastructure itself tells the search engine what just changed.
Cloudflare's global network carries over 20% of all internet traffic. That "whole-web view" is why OpenAI chose this partner — traditional crawlers can no longer keep up with AI-era demands for speed and accuracy.
Blocking crawlers and opening data at the same time — what is Cloudflare's game?
On July 1 Cloudflare announced that starting September 15, 2026, all sites on its platform will block mixed-use AI crawlers by default, forcing AI companies to separate search crawlers from training crawlers entirely.
In plain terms = one hand shuts the door on training data; the other opens a dedicated lane for search indexing — essentially compelling AI firms to split "search" and "training" into two clean operations.
This reflects Cloudflare repositioning itself: no longer just a website shield, but an aspiring rule-setter for AI-era traffic.
Where does the AI search race stand now?
OpenAI has moved fast: it launched the Atlas search browser on June 19 to challenge Chrome directly, then upgraded ChatGPT search on June 24.
BrightEdge data shows OpenAI accounts for 96% of real-time AI user-agent activity and 51% of crawler activity used for AI model building.
This means → OpenAI has opened a clear lead in AI search. Google and Perplexity are accelerating, but the gap remains significant.
What is Wall Street's take on Cloudflare?
Scotiabank analyst Patrick Colville upgraded Cloudflare to "outperform" and raised his target from $225 to $300.
His three key reasons: ① Cloudflare's Workers platform is becoming the default infrastructure for "vibe coding" — building apps through natural-language prompts — a trend investors have not yet priced in; ② traffic trends typically lead revenue by about three quarters, and agentic AI could push second-half results roughly 5 percentage points above consensus; ③ Cloudflare is winning top AI-native customers.
In plain terms = this analyst believes the market has not yet recognized Cloudflare's real position in the AI wave, and the stock is undervalued.
How wide is the valuation divide?
Cloudflare shares are up over 34% year-to-date, trading near $260 — close to the 52-week high of $272.66.
The consensus rating from 34 analysts is "buy," yet the average target of $246.07 sits roughly 8.5% below the current price. The range is extreme: $136 at the low end, $305 at the high end.
This means → Wall Street is split on the same core question: can AI-era cloud infrastructure opportunities justify this premium valuation? The pilot's results will be a key proof point.
Content is for reference only, not financial advice.