More than 10 million developers are now building applications on Cloudflare's network, a jump from 7.4 million in the previous quarter. The milestone signals something larger than company growth: it reflects how quickly software development itself is changing as artificial intelligence takes on a greater share of writing code, deploying applications, and acting on behalf of users.
From a Niche Compute Tool to a Full Development Stack
Cloudflare's developer platform began nearly a decade ago with Workers, a serverless compute product built on an architecture known as isolates rather than the virtual machines or containers that have long underpinned cloud computing. That distinction matters. Isolates are lighter-weight execution environments that start almost instantly and use fewer resources, allowing code to run closer to users across a distributed network instead of being confined to a handful of centralized data centers. What started as a single product has since expanded into a broader collection of tools spanning storage, databases, AI inference, media handling, real-time communication, and monitoring - the kind of full stack that developers once had to assemble from multiple vendors.
Rita Kozlov, Cloudflare's vice president of product management, described this expansion as a response to a shift in who, or what, software is now being built for. As she put it, the software lifecycle is moving from developers building for people toward developers and autonomous agents building for both people and other agents. That framing captures a change already visible in how applications are designed, tested, and deployed.
A Measurable Shift in Who - or What - Uses the Internet
The clearest evidence of this shift is traffic itself. Cloudflare reports that daily requests from AI agents on its network grew by more than 1,700% over the past year, and that for the first time, more than half of global Internet traffic did not originate from humans. That statistic alone reframes how engineers need to think about infrastructure: systems once built to handle browsers and mobile apps now have to accommodate automated agents querying data, triggering workflows, and communicating with other software continuously, around the clock.
To support that reality, developers are turning to tools such as Workers AI, Workers KV, AI Gateway, Durable Objects, D1, and R2 - components designed to let AI-native applications run globally without being stitched together from incompatible, legacy infrastructure. The goal, according to Cloudflare, is to let a single application serve people, other applications, programmatic interfaces, and autonomous agents without separate engineering for each.
Why Global Network Scale Has Become a Competitive Factor
Modern applications increasingly need to respond in milliseconds regardless of where a request originates, whether from a person in one country or an automated agent operating on a server elsewhere. Cloudflare's network now spans more than 340 cities in over 130 countries, interconnects with nearly 14,000 other networks, and offers close to 600 Tbps of external capacity. Dane Knecht, the company's chief technology officer, argued that this scale removes a historical trade-off in cloud computing, where developers often had to sacrifice either speed, security, or geographic reach to optimize for the others.
That trade-off is becoming less tolerable as applications grow more distributed and as agentic systems demand constant, low-latency communication across borders. Whether this concentration of infrastructure in a handful of large providers raises its own questions about resilience and dependency is a debate the industry will likely revisit as agent-driven traffic continues its rapid climb.