Nikil Viswanathan on pivoting 16 times to build a $10B crypto platform

Executive overview

Most founders attach identity to a product, hold on too long, and never find what people actually want. Nikil Viswanathan and co-founder Joe spent nearly a decade iterating through 16 pivots before Alchemy took off. The insight that unlocked everything: stop brainstorming in a room and start solving a problem you personally have. Alchemy emerged from the pain of building in crypto and finding the developer tooling impossibly hard.

The only metric that matters at the early stage is product-market fit — everything else is noise.

Building confidence before building companies

• Grew up in Lubbock, Texas — small enough to be best at everything (sports, math, music, student council)
• 18 years of repeated wins built foundational confidence that persisted through years of startup failure
• Immigrant mother read psychology books and deliberately reinforced capability across every activity
• At Stanford, prioritised relationships over grades; believed childhood confidence was sufficient capital
• Turned down an offer from Zuckerberg to run Messenger — valued the option value of trying over any salary

Why startup risk is often overstated

• In the US, worst case for a failed startup is getting a good job — that is most people's dream outcome
• Personal worst case was returning to Facebook; he and Joe joked they'd "crush it at McDonald's"
• The real risk is not trying: opportunity cost of not starting cannot be recovered
• Earlier is almost always better — responsibilities accumulate with age

The wrong way to find product-market fit

• Brainstorming in a room produces products too complex and not solving real pain
• Ego gets wrapped up in the concept, making bad ideas harder to kill
• Nikhil spent close to a year on one early product nobody wanted — far too long
• Correct upper limit: two weeks without clear user demand means scrap or pivot
• Iteration speed is the number one metric in the early days — get cycle time down to minutes

How Dandelion was actually built

• After years of failed apps, the team simply wanted to make something anyone would use
• The idea came from personal annoyance: coordinating lunch with nearby friends required too many texts
• Built a one-button app that mass-texted all friends — expected no one outside their circle to use it
• It blew up organically, distributed through a Facebook enterprise build link, never submitted to the App Store
• Best consumer apps start as side projects for a singular purpose and don't look like real businesses at first

How Alchemy was born

• In 2017, neighbours in their San Francisco loft were building a Bitcoin payments company
• Used a summer lull to experiment: if this is the next internet, what infrastructure is missing?
• Started building a hedge fund ML platform; realised if two Stanford AI graduates found it hard, tooling was the real problem
• Nikhil's sister's tweet crystallised it: whatever you're building, pivot to the infrastructure for that thing instead
• Alchemy became the developer platform for Web3 — the Windows/Mac equivalent that hides blockchain complexity

Picking a small market that grows fast

• The canonical great-company pattern: dominate a tiny market, then ride its growth — requires conviction when others don't yet believe
• In 2017–18, total addressable market was estimated at $5–10M if they captured every customer
• The bet: if crypto works, this is massive; if not, back to McDonald's
• Crypto prices dropped 84% from the 2021 peak; Alchemy's developer base still grew 4.5x over the same period
• Usage-based compute revenue rather than transaction-based kept growth stable through the crypto winter

What Alchemy actually does

• Powers the backend of crypto applications the same way AWS powers Netflix and Uber
• When a user buys crypto on Robinhood or views an NFT on OpenSea, Alchemy is serving blockchain data behind the scenes
• California's blockchain-based digital driver's licence, usable at airports, runs on this infrastructure invisibly
• Enterprise use cases include Shell tracking global equipment supply chains on-chain
• Revenue spans NFTs, DeFi, real-world assets, consumer apps, and enterprise — benefits whenever anything in crypto succeeds

Advice for founders stuck in the grind

• Product-market fit first — talk to customers obsessively, stay deep in their world, nothing else matters until you have it
• Become a domain expert by living the problem, not theorising; DoorDash founders delivered orders themselves
• If you're doing it for money you won't last — seven to ten years minimum if you succeed
• Speed is the only structural advantage a startup has: the fast eat the slow, not the big eat the small
• Be completely comfortable scrapping your product and starting over

On work-life balance and regrets

• Work is the hobby — no meaningful separation between the two
• Looking back: would have slept more and invested more in relationships
• Early-stage balance is largely impossible; treat it as a fast marathon, not a sprint
• In-person office is a deliberate, costly choice: limits remote hiring but produces team loyalty he values most
• Several early Alchemy hires came through birthday parties and Instagram, not job postings