Cloudflare’s AI-agent bet is changing the stock case

Cloudflare is turning AI agents and Anthropic ties into a bigger part of its growth case.

Cloudflare is no longer being judged only as a fast global network with a pricey valuation. By late May 2026, the company had paired a preliminary proxy filing for its annual meeting on June 30 with fresh AI moves, including an Anthropic collaboration and the purchase of VoidZero.

Cloudflare’s investment case now depends on AI workloads

The old Cloudflare pitch was easy to explain: sell security, performance, and developer tools on top of a global edge network. The new pitch is more ambitious. Cloudflare wants to become infrastructure for AI agents, which means handling inference, orchestration, and secure deployment close to users and data.

That matters because the company still has to prove it can turn heavy product investment into durable profits. In the Simply Wall St framing, the upside case depends on Cloudflare converting its network reach and AI stack into a broader platform, while the main risk is that spending runs ahead of monetization.

The Cloudflare Agents work and the new Cloudflare Environments for Claude Managed Agents product link that strategy to a real deployment path. That is a lot more concrete than generic AI branding. It gives customers a place to run agentic workloads on Cloudflare’s edge, which is exactly where the company wants to win.

• Cloudflare is still carrying losses, with Simply Wall St citing current earnings of -$86.7 million.
• The 2029 model expects revenue of $4.8 billion, which implies 27.6% annual growth.
• The same model expects earnings to rise to $328.5 million by 2029.
• Simply Wall St’s fair value estimate of $234.18 sits about 6% below the current price.

The Anthropic tie-up is more than a press release

The Anthropic relationship matters because it gives Cloudflare a credible partner in one of the hottest parts of the AI market. Anthropic’s Claude family has become a common reference point for enterprise AI adoption, and Cloudflare’s job is to make those workloads easier to run safely at the edge.

That is the sort of partnership investors should care about. It ties product strategy to actual usage rather than to vague “AI exposure.” It also helps explain why Cloudflare has been buying smaller tools like VoidZero, which can fill in developer workflow gaps around modern application building.

“We are in the early stages of the AI platform shift.” — Matthew Prince, Cloudflare co-founder and CEO, in Cloudflare’s 2024 annual letter

Prince has been making the same basic argument for years: the internet keeps moving closer to the edge, and Cloudflare wants to sit in the middle of that movement. The Anthropic deal gives that argument a sharper shape, because AI agents need security, low latency, and predictable deployment more than they need marketing language.

There is also a governance angle here. In late May, Cloudflare filed a preliminary proxy statement recommending support for board nominees Michelle Zatlyn, Scott Sandell, and Karim Lakhani. On its own, that filing does not change the operating story. But it does show management wants continuity while it pushes a more aggressive AI roadmap.

What the numbers say about the valuation debate

Cloudflare’s valuation debate is still the hard part. The company is asking investors to believe that AI-agent infrastructure can become a meaningful revenue line before competition and spending pressure erase the upside. That is a tall order, especially when the stock already embeds a lot of optimism.

Simply Wall St says the most optimistic analysts before this news were looking for about US$4.6 billion in revenue and roughly US$170 million in earnings by 2028. That is already a strong view, but it is still below the 2029 numbers in the narrative model. The spread tells you how much uncertainty sits inside the AI thesis.

Here is the practical comparison:

• Optimistic 2028 revenue view: about US$4.6 billion.
• Simply Wall St 2029 revenue view: US$4.8 billion.
• Optimistic 2028 earnings view: about US$170 million.
• Simply Wall St 2029 earnings view: US$328.5 million.

That gap is what investors are really pricing. If AI-agent adoption through Workers AI, Pay Per Crawl, and the Claude-managed-agent stack grows fast, Cloudflare can justify a premium multiple longer. If adoption is slower, the valuation compresses before the earnings inflection arrives.

For readers who want the broader market context, OraCore’s recent coverage of AI infrastructure names like other AI platform plays shows the same pattern: investors reward companies that can attach AI demand to existing distribution, not just companies that mention AI in earnings calls.

What investors should watch before June 30

The June 30 annual meeting is useful mostly as a checkpoint. The real story is whether Cloudflare can keep turning product launches into usage that shows up in revenue, margins, and eventually free cash flow. The board vote matters less than whether the AI-agent push becomes sticky with developers and enterprises.

Three things will tell the story faster than the proxy filing:

• Whether Cloudflare Environments for Claude Managed Agents gets real customer traction.
• Whether the company can keep AI spending from outrunning revenue growth.
• Whether management keeps adding developer tools that make the platform harder to replace.

My read is simple: Cloudflare has moved from “interesting edge company with AI optionality” to “AI infrastructure bet with an edge network attached.” That is a better story, but it is also a harder one to execute. If the next few quarters show agent workloads landing in production, the investment case gets stronger fast. If they do not, the market will keep asking whether the premium multiple is doing too much work.

For now, the most important question is whether Cloudflare can turn AI agents from a strategic talking point into a repeatable revenue engine before investor patience runs out.