The Founder Who Walked Away from YC

Every evening in Malad, a teenage Dhravya Shah would wander through Mindspace with a glass of Boost in hand, thinking about whatever he was building that week.

Just a kid who liked long walks and late-night code, who somehow ended up shipping over 60 products before the age of 19, hacking for Cloudflare, and qualifying for an O-1 “extraordinary ability” visa at 20.

Today, that same kid sits in San Francisco in a “solo together” house, running one of Silicon Valley’s most intriguing infra startups. Supermemory has already raised $3 million from Google DeepMind's chief scientist Jeff Dean, Cloudflare CTO Dane Knecht, Google DeepMind's Logan Kilpatrick, and a who’s who of AI founders betting on the memory layer of agents.

Supermemory is simple to describe and brutally hard to build: a universal memory API that gives AI agents long-term, cross-app memory.

It ingests everything—files, documents, emails, chats, app streams—and converts them into a personalised, evolving knowledge graph that any agent can query instantly. It already powers AI video editors, assistants, agentic workflows and multi-tool orchestration engines that need to remember users across months, not minutes.

https://www.youtube.com/watch?v=mmLzN6xNXXU

Why Memory, And Why He Walked Past YC

After applying for the Y Combinator Summer Fellows grant, Dhravya got in.

He revealed to AIM that he walked away from YC because investors outside the startup accelerator offered significantly better terms, deeper infrastructure expertise and immediate strategic value for the company.

“I got the summer fellows grant, but then I decided to build Supermemory as a company instead,” recalled Shah, in an interview on Front Page by AIM Network. He added that “the valuation that they offered was already six times more than YC.”

The investors he eventually chose, leaders from Google, Cloudflare and the broader infra ecosystem, now review architecture, guide research, open doors to design partners and validate the technical rigour of Supermemory’s roadmap.

“They are some of the best people in this industry,” he said. “They act as a testament to how rigorous our own processes are in terms of infrastructure and research.”

As a solo founder building in one of the hardest technical categories, particularly infra + memory, Shah said the right guidance mattered more than YC’s brand recognition.

“I do think that maybe getting YC is better for recognition,” he said. “But that’s the wrong kind of recognition for a B2B SaaS guy like me. For me, this is the best thing that could happen.”

He clarified that he never applied to YC’s traditional fundraising batch, only the Summer Fellows grant—and by the time the opportunity arose, external interest had already pushed Supermemory into a different orbit.

“We actually do industry partnerships as well with these companies,” he said, noting prior work with Cloudflare and ongoing collaboration with Google’s models.

The Noise, the Media, and Perplexity

Shah told AIM that most stories the Indian media publishes about him are simply wrong: the 'IIT dropout' tag, assumptions about his upbringing or that he comes from an elite network.

“There are multiple problems with the media narrative…None of it is true,” he said.

“People think I’m from an elite, rich family or that all my elite connections gave me money. I’m just a dude from Malad. My dad is a businessman. I know no one in tech.”

He said the oversimplification extends to his product as well. “People say, ‘This is just RAG, I built this over a weekend’, but they either don’t understand the product or only know the version from two years ago. This has taken years to build.”

Shah has given up, he doesn’t fight the noise anymore. “I just ignore it…play into it…use it as marketing, even to boost ARR.”

He also rejected the doom narrative around Perplexity with one of the sharpest lines of the interview.

“Pessimists sound smart, but optimists make all the money,” he said.

People underestimate how hard it is to build a product that even ten people genuinely love, he added, which is why he respects what Perplexity AI CEO Aravind Srinivas has achieved.

“I have huge respect for Perplexity and Aravind…They’ve built something millions of people love. That’s insanely hard.”

He believes Perplexity will continue to win because they have fundamentals and iteration velocity on their side.

“Once a company gets to that stage, they can keep people happy…They have it in their blood. Perplexity will win no matter what people think.”