Cloudflare Sees Agentic AI Opportunity Dwarfing Past Growth Drivers

Cloudflare views agentic AI as a growth opportunity larger than its previous drivers and has raised its long-term “North Star” growth target to 50% from 40%.

Demand is emerging through the Workers platform and consumption-based revenue, with AI customers driving greater traffic and usage; Cloudflare is focusing on low-latency AI inference rather than training the largest models.

Cloudflare is developing new monetization infrastructure for the agentic economy, including “pay per crawl,” a Monetization Gateway and digital wallets, though significant technical and pricing challenges remain.

Cloudflare NYSE: NET sees its opportunity in agentic artificial intelligence extending beyond its current growth drivers, with Chief Financial Officer Thomas Seifert saying the company’s network architecture is positioned to support low-latency, secure and distributed AI inference workloads.

Speaking at an investor event, Seifert said Cloudflare has moved from its earlier focus on improving sales execution and productivity into what he described as a broader shift driven by agentic AI.

He said the company views its business in a series of “acts,” including its core network and security offerings, Workers and AI products, and a newer effort to help content creators protect and monetize their content.

“We think that that opportunity is going to dwarf what is already behind us,” Seifert said, referring to the company’s AI-related opportunities.

He added that Cloudflare raised its long-term “North Star” growth target to 50% from 40% at its Investor Day and expects to exit the year with momentum toward that objective.

Seifert pointed to growth in Cloudflare’s developer community and variable consumption revenue as early indicators of demand.

He said the company added 2 million developers quarter over quarter and that developers are increasingly building AI applications on its Workers platform.

According to Seifert, companies including Lovable, Figma, Anthropic and OpenAI are building on Workers.

He also said Cloudflare estimates that about 80% of significant AI enterprise infrastructure customers use its services.

The CFO said AI is changing software spending from seat-based subscriptions toward more variable, consumption-driven models.

In Cloudflare’s recent second quarter, he said upside relative to the company’s guidance came from greater-than-expected customer consumption, including more agentic AI traffic, requests and use of products through pool-of-funds contracts.

However, that shift also makes forecasting more difficult.

Seifert said Cloudflare is rethinking its forecasting processes and needs to tie more deeply into infrastructure data to identify the indicators driving consumption.

He said the company may need to shift guidance midpoints to preserve ranges as consumption becomes less predictable.

Cloudflare does not currently view training the largest AI models as its principal market, Seifert said.

Instead, it is focused on inference, including the CPU and memory-intensive work of orchestrating agentic workflows, such as determining where data comes from, what needs to be computed and which APIs need to be accessed.

Seifert said Cloudflare’s experience scheduling workloads across its network has helped it optimize AI inference.

The company deploys GPU capacity across its points of presence and data centers, he said.