This initiative highlights the critical role of infrastructure in scaling early-stage startups. By providing $500k in credits and a global platform, Cloudflare enables small teams to leverage enterprise-grade compute, security, and AI capabilities without massive upfront investment.
Sixteen years ago, Cloudflare was one of more than 1,000 startups hoping for a spot on stage at TechCrunch Disrupt.
We were not, on the face of it, an obvious choice. Cloudflare was infrastructure: we made websites faster and protected them from attack, which was not well understood by the general market at the time. Infrastructure is often invisible right up until the moment it becomes important.
But on September 27, 2010, Matthew Prince and Michelle Zatlyn got on the Startup Battlefield stage and launched Cloudflare to the public. During the presentation, people started signing up. Then more people started signing up. By the time the judges had finished asking questions, hundreds of websites had joined Cloudflare, putting our initial five data centers to the test in real time. In the seven days following, traffic through our network increased almost 10x and Cloudflare jumped from the 1,000th largest site online to one of the top 50.
Cloudflare didn’t win the main trophy that day. At the awards ceremony, TechCrunch founder Mike Arrington described what we did as something akin to "muffler repair for the Internet" and honestly, he had a point. But then he named us the Most Innovative Company. As Matthew wrote later: "You may not win the trophy, but you'll receive something else far more important."
There are moments in the life of a company when somebody gives you a room, a microphone, and a small amount of time to explain the thing you have spent months or years obsessing over. Most of the time nothing magical happens. Sometimes, though, the right people hear it at the right moment and suddenly an idea that has mostly existed between a handful of people starts moving through the world.
This October, as Cloudflare turns 16, we want to give five early-stage startups a stage of their own.
The Cold Start is a live startup competition taking place next month at Cloudflare Connect in San Francisco. We’ll select five early-stage companies and give each of them five minutes on stage to explain what they’re building, why it needs to exist, and why they are the people who should build it.
We are less interested in perfect pitch decks than in interesting ideas clearly explained. You do not need thirty slides, a suspiciously precise TAM calculation, or a rehearsed story about how your childhood prepared you to disrupt accounts receivable. What we want is to understand the vision: what changed in the world that makes it possible, what you see that other people have missed, and why you cannot stop thinking about it.
The five finalists will make their case in front of the Cloudflare Connect audience and three people who've spent a fair amount of their lives thinking about companies, infrastructure, and the Internet:
The judges will select one startup to win $500,000 in Cloudflare credits, take over a billboard in San Francisco, and receive an invitation to our VIP speakers dinner that evening.
Five companies, five minutes each, and a room full of people paying attention.
The Cold Start is open to ambitious early-stage startups, based in the US and Canada, that have raised less than $10 million. Beyond that, we are deliberately keeping the definition broad because the most interesting companies rarely arrive neatly categorized.
We want to see ideas that seem obvious once somebody finally builds them, and ideas that initially sound slightly unreasonable. We want infrastructure that appears boring until you realize everyone is going to need it; products that could not have existed a few years ago; strange new interfaces; new ways of building software; things aimed at enormous existing markets and things aimed at markets nobody has bothered to name yet.
Most of all, we want to meet people who have noticed something about the world and decided to do something about it.
As part of the application, we’ll ask you to tell us who you are, give us your one-line pitch, explain what you’re building and why, tell us about your funding and revenue stage, show us how Cloudflare fits into your stack, and point us toward anything else that helps us understand you and your work.
The goal is simple: make us understand why the thing you’re building should exist.
Applications are open now and close Friday, October 2, 2026.
The Cold Start will take place Monday, October 19, from 4:00-5:00 p.m. PDT at Moscone West in San Francisco, as part of Cloudflare Connect. Cloudflare will pay for the five finalists to fly to San Francisco for the competition.
Connect brings together people building and thinking about what comes next for the Internet. This year’s lineup includes AI pioneer Dr. Fei-Fei Li; Idealab founder Bill Gross; organizational psychologist and author Adam Grant; AMD CTO Mark Papermaster; Vue.js and Vite creator Evan You; Lovable co-founder and CTO Fabian Hedin; and OpenAI member of technical staff and creator of OpenClaw, Peter Steinberger.
For five young companies, we are reserving part of that stage. Each startup will have 5 minutes to pitch, followed by 3–5 minutes of questions by the judges.
There is something we like about that symmetry. Sixteen years ago, Cloudflare needed someone to take a chance on an infrastructure company with a difficult story to tell and give us a few minutes in front of the right room. Today, we are fortunate enough to have a stage of our own, and we want to pass that same opportunity on to companies that are just getting started.
There is a practical reason Cloudflare spends so much time working with startups: very small groups of people have an uncanny ability to attempt very large things.
The problem is that ambitious software increasingly depends on infrastructure that, not very long ago, only the largest technology companies in the world could afford to build for themselves. Global compute, storage, networking, security, real-time systems, AI inference, and the ability to survive the possibility that the thing you made suddenly becomes popular should not require a company to first become enormous.
We think you should be able to reach for those capabilities on day one.
That is part of the idea behind Cloudflare for Startups, through which eligible early-stage companies can receive up to $350,000 in Cloudflare credits for one year. It is also part of the reason we continue expanding Cloudflare’s developer platform: a tiny team should be able to build something on Tuesday and if the Internet decides it likes it on Wednesday, spend Thursday focused on the product rather than hastily becoming experts in global infrastructure.
Cloudflare began with its own slightly unreasonable premise: that the performance, security, and global compute available to the largest companies on the Internet should be available to everyone on day one. In 2010, we got a chance on a stage to explain why that matters.
Sixteen years later, we have a much larger network, a somewhat larger team, and significantly better circuit breakers.
Now we want to hear what you’re building.
Continue reading on the original blog to support the author
Read full articleThis partnership bridges the gap between edge computing and self-hosted environments. By unifying Deno and Cloudflare Workers, developers gain greater flexibility to run serverless primitives anywhere, reducing vendor lock-in and simplifying the deployment of stateful edge applications.
This article highlights how AI tools are transforming early-career engineering by accelerating onboarding and productivity. It proves that interns can ship critical infrastructure, from cache optimization to post-quantum security, when empowered with modern development workflows.
The shift toward agent-dominant traffic fundamentally changes web architecture and business models. Engineers must optimize for high-volume automated requests while ensuring fair value exchange for content, preventing a tragedy of the commons that could stifle new market entrants.
Worker Previews eliminate 'works on my machine' issues by providing high-fidelity, isolated environments for every change. This enables safer deployments, better observability for AI agents, and risk-free testing of stateful migrations before they hit production.
