Cloudflare, Inc. (NET.US), a cloud services provider focused on its "Connectivity Cloud" positioning and often referred to as the "Internet Gatekeeper," reported fourth-quarter results before market open on Wednesday, Beijing time, that surpassed Wall Street analysts' average expectations. The company also issued revenue guidance for the first quarter of fiscal year 2026 and for the full fiscal year 2026 that exceeded recently upwardly revised analyst estimates, a trend fueled by the viral popularity of OpenClaw (formerly known as Clawdbot and Moltbot). Driven by the strong results and outlook, Cloudflare's stock surged more than 14% in after-hours trading, echoing a more than 20% gain over three trading days in late January when OpenClaw first gained global traction.

For the fourth quarter of fiscal year 2025, ended December 31, Cloudflare's total revenue increased 34% to $614.5 million, exceeding the Wall Street consensus estimate of approximately $590 million, which had also been significantly raised by analysts recently. Non-GAAP earnings per share were approximately $0.28, representing a 47% year-over-year increase and beating the average analyst estimate of about $0.27. For the full fiscal year 2025, total revenue grew 30% to $2.1679 billion.

Other key Q4 performance metrics included a GAAP gross profit of approximately $452.6 million, with a gross margin of about 73.6%, compared to roughly $351.3 million in the same period of fiscal 2024. Non-GAAP gross profit was approximately $460.2 million, yielding a margin of about 74.9%, compared to approximately $356.8 million in Q4 fiscal 2024. Non-GAAP operating income for the quarter was approximately $89.6 million, representing about 15% of total revenue, a significant increase from approximately $67.2 million in the prior-year period.

Regarding the closely watched revenue guidance, Cloudflare's management provided a robust outlook for both Q1 fiscal 2026 and the full fiscal year 2026. For the first quarter, management expects total revenue between $620 million and $621 million, with the midpoint of this range notably higher than the Wall Street consensus estimate of approximately $615.5 million, which analysts had recently raised substantially due to the global wave of agentic AI sparked by OpenClaw. For the full fiscal year 2026, management anticipates total revenue in the range of $2.785 billion to $2.795 billion, with the midpoint also exceeding the upwardly revised analyst consensus of approximately $2.740 billion. For Non-GAAP EPS, management guided to $0.23 for Q1 fiscal 2026 and between $1.11 and $1.12 for the full year, largely in line with analyst expectations.

In the earnings statement, Cloudflare co-founder and CEO Matthew Prince highlighted that the company closed its largest annual contract value (ACV) deal ever in the fourth quarter, representing an average of $42.5 million per year. Total new ACV grew nearly 50% year-over-year, marking the fastest growth rate since 2021. Prince stated, "The global shift towards AI and agents represents a fundamental restructuring of the internet, revolutionarily driving strong demand for Cloudflare's Connectivity Cloud services and network security technology." He added, "If AI agents are the new users of the internet platform, then Cloudflare is the core platform they run on and the network they must traverse. This creates a virtuous flywheel: more agentic AI workflows on the Cloudflare Workers platform drive more code execution, thereby increasing demand for our performance, cybersecurity, and connectivity cloud services. We were built for this moment and the rise of the agentic internet."

Cloudflare is often called the "Internet Gatekeeper" because it operates at the "entry layer" for a vast number of websites and applications, acting as a reverse proxy and a "traffic adjudication point." User requests typically first reach Cloudflare's edge network before being forwarded to the origin server. Prior to forwarding, Cloudflare performs tasks like cache acceleration, TLS termination, WAF/bot identification, DDoS mitigation, and access control—decisions to "allow or block" traffic happen at this technical layer. Consequently, based on its position in the technology stack, it effectively stands at the "front door" of websites, inspecting each request: allowing legitimate traffic, blocking malicious traffic, and restricting or blocking certain automated access (like crawlers/bots) based on policy. Due to this network-layer position offering "visibility, control, and blocking capability," some research and media discussions directly describe it as the internet's "gatekeeper."

Cloudflare focuses on its "Connectivity Cloud" positioning, integrating network acceleration (native CDN/Anycast network), security (WAF, DDoS, Zero Trust/SASE), and a developer platform (serverless Workers, KV/D1/Queues, R2 object storage, etc.) onto a single global edge network architecture, providing enterprises with "connectivity + protection + acceleration." This has been the company's core narrative and product packaging logic over the past two years. Unlike cloud giants like Microsoft or Google, which lean more towards central cloud/hyperscale training and inference clusters tied to their own or ecosystem model services and data platforms, Cloudflare emphasizes edge + network + security integration. It focuses on "locally deploying inference, RAG, data access, and security controls" close to users, using zero-egress R2, edge vector databases, AI Gateway, and Workers AI to string together "traffic – data – inference – security" in a pipeline. This approach can complement the "last mile" of central cloud architectures or coexist with multi-cloud environments. It positions Cloudflare more as a "network and edge operating system" for the AI inference era, leveraging its global edge network, SASE, Zero Trust, and serverless platform as an innovative architectural layer for connectivity and runtime operations in the AI era, helping numerous AI applications run efficiently, cost-effectively, and securely in edge environments.

The advancement of the Gemini 3 series and the viral popularity of Claude and OpenClaw in "programming/Agentic usage" have significantly accelerated the penetration depth and speed of generative AI and agentic AI workflows across both enterprise (business processes, development, customer service, analytics) and consumer (AI search/recommendation/consumer agents) segments. This objectively shifts compute demand further from "AI training systems" towards broader AI inference and online services, thereby substantially elevating the growth trajectory of global cloud IaaS infrastructure and platforms like Cloudflare's comprehensive cloud PaaS AI developer ecosystem.

The recent explosion in popularity of OpenClaw has helped Cloudflare concretize the narrative that "Agentic Internet workflows equal demand for a new generation of infrastructure." Cloudflare itself has begun a new narrative, directly referencing the phenomenon of "people frantically buying Mac minis to run Moltbot/OpenClaw" and introducing Moltworker to facilitate easy deployment on its developer platform. If an individual aims to deploy "cloud-based agentic AI workflows" using open-source OpenClaw—especially for public internet, multi-user, multi-channel access—and chooses the lower-barrier, highly efficient deployment path via Cloudflare's Moltworker, the key pipeline would deeply utilize various Cloudflare platform components: using a Worker for API routing and proxying at the entry point, running the OpenClaw gateway and integrations in a Sandbox container, utilizing R2 for persistent storage, and securing the management interface and APIs with Cloudflare Access/Zero Trust. Simultaneously, AI Gateway would provide unified access to model providers like Anthropic, offering key management, usage/cost observability, while Browser Rendering productizes "browser automation needed by agents" into a callable edge capability. These Cloudflare-focused components essentially address the fundamental底层 needs of an "agent system": an orchestrated compute entry point, an isolated execution environment, reliable storage, a stable proxy layer for external models and tools, and security with auditing.

However, as the OpenClaw open-source community leans towards a "local-first" philosophy for deploying agentic AI workflows—emphasizing local gateway control plane, unified entry for sessions/tools/events, multi-channel inboxes, and multi-agent routing—it means the workflows can run on personal hardware or self-built servers without requiring Cloudflare. Therefore, "locally deploying OpenClaw" has significantly less dependency on Cloudflare subscriptions, as users can handle domain management, reverse proxying, authentication, and persistence through other means. That said, "local deployment" does not equate to "completely不需要 Cloudflare." When needing to securely expose local agents for remote use by oneself or a team, or when desiring a unified layer for cost and reliability governance of model calls, Cloudflare can remain a high-ROI option—for instance, using AI Gateway as a unified proxy, logging, and fallback layer for multiple models, avoiding scattering provider API keys; using Zero Trust/Access to add enterprise-grade access control and observability to management interfaces and APIs; and combining object storage/edge entry points to externalize "agent memory and assets" more stably. However, these are considered "enhancements/engineering reinforcements," not prerequisites for running OpenClaw locally.