Runpod, an AI app internet hosting platform that launched 4 years in the past, has hit a $120 million annual income run price, founders Zhen Lu and Pardeep Singh inform TechCrunch.
Their startup journey is a wild instance of how if you happen to construct it effectively and the timing is fortunate, they may undoubtedly come.
The story contains bootstrapping their technique to over $1 million in income; touchdown a $20 million seed spherical after VC Radhika Malik, a companion at Dell Applied sciences Capital, noticed some Reddit posts; and gaining one other key angel investor, Hugging Face co-founder Julien Chaumond, as a result of he was utilizing the product and reached out over the help chat, the founders inform TechCrunch.
All of it started in late 2021 when the 2 associates, who labored collectively as company builders for Comcast, determined the interest they have been doing wasn’t enjoyable anymore.
That they had constructed setups of specialised computer systems used to generate Ethereum of their respective New Jersey basements. Whereas they did efficiently mine a little bit of the cryptocurrency, it wasn’t sufficient to pay again their funding, they stated. Plus, mining was going to finish after the much-ballyhooed community improve referred to as “The Merge.”
On high of that, it was “boring” after a few months, Lu stated.
However they’d talked their wives into letting them spend a great $50,000 on the interest between them, they estimated. Lu and Singh knew that residence concord trusted discovering a method to make use of these GPUs.
Techcrunch occasion
San Francisco
|
October 13-15, 2026
The devs had been engaged in machine studying initiatives at work, in order that they opted to transform their mining rigs into AI servers. This was earlier than ChatGPT, even earlier than DALL-E 2.
As they repurposed the rigs, “We have been seeing how actually god-awful the software program stack was for coping with these GPUs,” Lu stated. As builders, they discovered an issue they needed to resolve.
Runpod was born “as a result of we felt that the precise expertise of creating software program on high of GPUs was simply scorching rubbish,” Lu described.
Just a few months later in early 2022, they have been able to share what they’d constructed. Runpod is a platform for internet hosting AI apps, emphasizing pace, simply configured {hardware} (together with a serverless choice that automates configuration), and dev instruments like APIs, command-line interfaces, and different integrations.
Again in 2021, they solely had a number of such integrations (like help for common net app device Jupyter notebooks). The following drawback: discovering beta testers.
“As first-time founders, we didn’t actually know methods to market or methods to do something,” Lu recalled. “So I’m like, all proper, let’s simply put up on Reddit.”
So, they posted in a few AI-oriented subreddits. The supply was easy: free entry to their AI servers in trade for suggestions. It labored. They landed beta clients, which led to paying clients. Inside 9 months, they’d stop their jobs and hit $1 million in income, they stated.
Bootstrapping development
However that led to a different drawback. “Six months in, enterprise customers have been like, ‘Hey, I need to truly run actual enterprise stuff in your platform. However I can not run it on servers which might be in folks’s basements,” Lu stated.
It had not occurred to the New Jersey founders to lift capital from VCs. As a substitute they fashioned revenue-share partnerships with information facilities to develop capability. However it was worrying. The founders wanted to remain three steps forward.
“If we don’t have the GPUs, the market sentiment, the person sentiment adjustments. As a result of after they don’t see capability from you, they go some other place,” Singh described.
In the meantime, their person base was rising on Reddit and Discord, particularly after ChatGPT launched.
VCs have been additionally on the prowl for investments. Malik noticed them on Reddit and reached out, their first VC name. However Lu didn’t know methods to pitch to an investor. “Radhika was tremendous useful, even on the first dialog,” he stated. She principally defined to him how a VC thinks and informed him she’d keep in contact.
In the meantime, Lu had a enterprise to run that needed to pay for itself. “It was virtually two years the place we actually didn’t have any funding,” he stated. So Runpod by no means provided a free tier. It needed to at the very least pay for itself, even when it wasn’t throwing off a lot revenue. Not like different AI cloud companies that started as crypto miners, these founders refused to tackle debt, they stated.
By Could 2024, with AI app fever spreading, their fortunate resolution to launch AI internet hosting for devs two years earlier was paying off. Their enterprise had grown to 100,000 builders, they usually landed a $20 million seed deal co-led by the VC arms of each Dell and Intel, with participation from large names like Nat Friedman and Chaumond.
They haven’t raised more cash since however at the moment are planning to, armed with a enterprise that, they consider, ought to command a wholesome Sequence A.
At the moment, Runpod counts 500,000 builders as clients, starting from people to Fortune 500 enterprise groups with multimillion-dollar annual spend, the founders stated.
Their cloud spans 31 areas globally and counts clients like Replit, Cursor, OpenAI, Perplexity, Wix, and Zillow as customers.
Competitors can also be fierce. Devs have all the main clouds to select from (AWS, Microsoft, Google), plus loads of industry-specific selections like CoreWeave and Core Scientific.
However in addition they see their place on the planet a bit otherwise — as a dev-centric platform. They don’t see coding ever going away however altering. Programmers will turn out to be AI agent creators and operators.
“Our aim is to be what this subsequent era of software program builders grows up on,” Lu stated.
