Building for “Google-Scale” Crypto
How Sei’s Jay Jog Aims to Redefine Throughput
In the rapidly shifting world of crypto, a chain that can handle thousands of transactions per second seems ambitious. One that wants to approach Google-level scale—100,000 transactions per second—sounds almost unimaginable. But for Jay Jog, co-founder of Sei, it’s a necessary goal if Web3 is to stand shoulder-to-shoulder with Web2 giants. In our chat, Jay laid out how Robinhood’s GameStop saga inspired Sei’s beginnings, why fundraising is an art of perseverance, and how building an L1 contrarian bet in a rollup-focused market can pay off—if you build for real performance.
From Facebook Bonuses to GameStop Realizations
Jay’s first taste of crypto came in 2017, watching friends dump $100k Facebook signing bonuses into ETH at $40. “I thought it was the worst idea ever,” he laughs—until they made money. Then lost it all. Still, it showed him the financial power of crypto speculation.
But the big turning point came at Robinhood in early 2021. Jay led an engineering team there when the platform halted buys of meme stocks like GameStop, angering users nationwide. “I felt powerless,” Jay recalls. “People wanted a decentralized Robinhood—so we started building it.” That impetus led to Sei, originally aimed at being a next-gen DEX. Over time, he and co-founder Jeff realized the real problem was infrastructure: existing blockchains lacked throughput and developer tooling to sustain true high-frequency trading.
A Contrarian L1 Amidst a Rollup Craze
“It wasn’t easy convincing investors to back another L1 in 2022,” Jay admits. Terra had collapsed, and everyone was hawkish on rollups, not layer-ones. But Multicoin saw potential and funded Sei’s contrarian approach. “Everyone else said no,” Jay says, “but you only need a few believers.”
Sei’s early days used CosmWasm exclusively, but they quickly pivoted to the parallelized EVM after discovering 85–90% of devs were EVM developers. “We realized if we want dev adoption, we can’t ignore EVM.” The result: a chain that offers both the EVM’s developer mindshare and the speed of a cosmos-based infrastructure.
Lessons in Fundraising & Team Building
Jay endured countless rejections while raising. “Every fund that passed was basically saying, ‘We don’t believe in you.’” The key, he notes, is pushing forward. “Investors can be wrong. If you believe in your mission and keep shipping, the right partners will come along.”
For team building, Sei took a part-time first approach. Their founding engineers—ex-Googlers, Facebookers, Databricks devs—contributed nights and weekends, so both parties saw real working relationships before signing on full-time. “Interview processes are shallow; actually coding together is how you truly vet people,” Jay insists. This method also helped reduce flight risk. After two months of shipping code, new hires rarely looked elsewhere.
The Performance Obsession
Sei’s push for a “decentralized Robinhood” requires near-instant finality and throughput that dwarfs existing EVM chains. “Ethereum or its rollups might do 50 TPS,” Jay says. “We’re already at around 5,000 TPS. But to compete with a Web2 giant like Google, we need 100k TPS.”
Why Google-level performance? Jay points out that in Web2, large-scale applications routinely hit tens of thousands of operations per second. True DeFi and real-time trading could demand similar throughput. “We want developers to build anything on-chain without clunky, capital-inefficient designs just to work around throughput constraints.”
Yet high throughput alone won’t lure devs. They come for opportunity—the chance to be early in a chain’s growth and reap the rewards. Sei now sees “killer app” adoption as the final step to proving product-market fit. “We have narrative market fit and community fit,” Jay explains. “What’s missing is that breakout DApp that onboards real users and traffic.”
What’s Next for Sei
Jay’s vision is to blur the line between Web2 and Web3 until “building on-chain” feels identical to “building in the cloud.” That includes:
• Low-latency finality (~100 milliseconds) so users see instantaneous confirmations.
• High throughput (tens of thousands of TPS) to handle real-world user loads.
• Developer-friendly: EVM compatibility + minimal friction. Tools for debugging, block explorers, and user login solutions akin to “Sign in with Google.”
• Open ecosystem: Let the broader community tackle improvements like caching, user onboarding, or bridging solutions.
Ultimately, Jay says, “If users can’t even tell they’re on a blockchain, that’s when we’ve succeeded.”
Takeaways for Founders
1. Ignore Most Investor Rejections: One partner who truly believes can outweigh 99 that say no.
2. Test Drive Hires: Working part-time first ensures deeper vetting than any interview can.
3. Pick EVM’s Mindshare: If dev adoption is the goal, meet them where they are.
4. Build For Real Throughput: Don’t settle for 50 TPS. Aim for scale that can handle tomorrow’s “Google moment.”
5. Wait for the Killer App: True product-market fit for an L1 (or L2) crystallizes once a breakout DApp emerges.
For anyone eyeing the next wave of crypto infrastructure, Sei stands as a bet that performance does matter—and that bridging EVM devs with cosmos-level speed might birth the chain that handles crypto’s “GameStop moment” at global scale.