Runpod’s AI cloud platform reaches a $120 million annual revenue milestone.
Runpod, an AI cloud platform for hosting and running AI applications, has reached an annual revenue run rate of $120 million. The company was founded four years ago by Zhen Lu and Pardeep Singh, two former corporate developers who never planned to build a venture-backed startup. What began as a side project in their basements has turned into one of the fastest-growing platforms for AI developers.
The founders shared their story with TechCrunch, describing a journey shaped by timing, persistence, and a single Reddit post that helped change everything.
From Crypto Mining to a New Idea
Lu and Singh met while working as developers at Comcast in New Jersey. Outside their day jobs, they shared a hobby that was popular at the time: cryptocurrency mining. Each of them built specialized computer systems with powerful GPUs in their basements to mine Ethereum.
While they managed to mine some cryptocurrency, the returns were disappointing. The income was not enough to recover the money they had spent on hardware. On top of that, Ethereum was preparing for a major network change known as The Merge, which would end GPU-based mining altogether.
After a few months, the hobby stopped being interesting. Lu described it as boring. But there was another problem. Between them, they had spent around $50,000 on the equipment. Convincing their families to support that expense had not been easy, and letting the hardware sit unused was not an option.
Turning GPUs Into AI Servers
At work, both founders were already involved in machine learning projects. They decided to repurpose their mining rigs and turn them into servers for AI workloads. This was before tools like ChatGPT or DALL-E 2 became widely known. As they started experimenting, they ran into constant frustrations. The software required to manage GPUs was messy, complicated, and difficult to use.
As developers, they saw a clear problem that needed fixing. Lu explained that Runpod was created because developing software on GPUs felt unnecessarily painful. They believed developers deserved better tools and a smoother experience.
Building Runpod for Developers
By early 2022, they were ready to share what they had built. Runpod became a platform designed to host AI applications with a focus on speed, flexible hardware options, and developer-friendly tools. These included APIs, command-line interfaces, and easy integration with popular tools like Jupyter notebooks. At first, the platform supported only a few integrations. The next challenge was finding people willing to test it.
As first-time founders, they had no marketing plan and no experience promoting a product. Lu decided to try something simple. He posted about Runpod on Reddit in a few AI-focused communities, offering free access to their servers in exchange for feedback.
The response was immediate. Developers signed up, tested the platform, and shared their thoughts. Some of those early users became paying customers. Within nine months, Lu and Singh left their jobs and reached $1 million in revenue.
Growing Without Outside Funding
Early success brought new challenges. Business customers wanted to run serious workloads on Runpod, but they were not comfortable relying on servers located in private basements. The founders realized they needed more capacity and more reliable infrastructure. At first, raising venture capital was not even part of their thinking. Instead, they formed revenue-sharing partnerships with data centers to expand their hardware footprint.
This approach worked, but it was stressful. Demand for GPUs was growing quickly, and any shortage could send customers elsewhere. Singh explained that if users could not find capacity, they would simply move on to another provider. At the same time, Runpod’s community kept growing on Reddit and Discord, especially after the launch of ChatGPT increased interest in AI development.
A Reddit Post Leads to Venture Capital
As Runpod gained visibility, investors started paying attention. Radhika Malik, a partner at Dell Technologies Capital, came across Runpod through Reddit and reached out. It was their first serious conversation with a venture capitalist. Lu admitted he had no idea how to pitch a startup. Malik helped guide him, explaining how investors think and what they look for. Even then, Runpod continued operating without outside funding for nearly two years.
Because the business had to support itself, Runpod never offered a free tier. The company charged from the start, covering costs even if profits were limited. The founders also avoided taking on debt, unlike other AI cloud services that emerged from crypto mining backgrounds.
A Big Bet Pays Off
By May 2024, interest in AI applications had exploded. Runpod’s early decision to focus on AI hosting was paying off. The platform had grown to 100,000 developers. That same month, the company raised a $20 million seed round co-led by the venture arms of Dell and Intel. Other investors included Nat Friedman and Hugging Face co-founder Julien Chaumond. Chaumond became involved after using Runpod himself and reaching out through the support chat.
Where Runpod Stands Today
Today, Runpod serves more than 500,000 developers. Its customers range from individual builders to Fortune 500 companies, spending millions each year. The platform operates across 31 regions worldwide and counts companies like Replit, Cursor, OpenAI, Perplexity, Wix, and Zillow among its users. Competition in the AI cloud space is intense. Developers can choose from major providers like Amazon, Microsoft, and Google, as well as specialized firms such as CoreWeave and Core Scientific.
Still, Runpod sees itself differently. The founders believe coding is not disappearing but evolving. Developers are becoming builders and operators of AI-driven systems. Runpod aims to be the platform that this next generation of developers grows up using. As Lu put it, the goal is simple: build the tools that future software creators rely on from day one.
About The Author
