Anatoly Yakovenko, founder of Solana (now worth ~$47B), recounts his journey from a Soviet childhood to building a high-performance blockchain, sharing lessons on founder psychology, fundraising, co-founder dynamics, the FTX collapse, and how he now uses AI agents to 12x his coding output while still chasing the “hard engineering problems” that drive him.
Childhood in the USSR and immigration to the US
Grew up in the Soviet Union with vivid but surreal memories: uniform, red tie, three state TV channels (one Disney cartoon a week — DuckTales), and unsupervised outdoor play.
Parents had no fond memories; they wanted out since the 1970s but were blocked. Grandparents met in WWII; the USSR was “an evil empire that persecuted everyone.”
Immigration happened after the USSR fell in 1991. Landed in Chicago with $50 per person (legal limit), no English, in their 40s.
Parents took blue-collar work: mom (a pianist) taught music in schools; dad (a civil engineer who “rearranged mountains in Siberia with dynamite”) did roofing and studied for electrician certs — Anatoly helped with his circuit-analysis homework as a kid.
Culture shock: expected coast-to-coast Manhattan; found farms, suburbs, and an explosion of media (Michael Jordan, Bulls, internet). Got a computer in 1993, dove into Linux and open source.
Early career: Qualcomm, side projects, and the Bay Area move
Studied CS at University of Illinois (chose coding over bioengineering after “weed-out” lab classes). Co-founded a VoIP startup (GrandCentral/Google Voice precursor) that failed because they didn’t grasp global internet distribution — lesson learned a decade later.
Joined Qualcomm in San Diego (2001–2014, ~13 years). Stayed for the interesting work (mobile hardware leaping generations yearly) and lifestyle: surf, bike, run, underwater hockey with a cohort of Ironman engineers who constantly spun up side projects (14+ tiny startups, none panned out).
Moved to Bay Area ~2011 for wife’s Airbnb career. Bored working remote, tried Mesos (~5 months) then Dropbox (~9 months, 2015–16). Dropbox team was “hardcore and smart” (compression + AI) — would have stayed, but the Solana idea hit.
The Solana insight: Proof of History and the “Eureka moment”
2017: mining crypto on GPU rigs (built for deep learning) to pay hardware costs. Got “nerd-sniped” by proof-of-work.
At Cafe Sole with two coffees and a beer, stayed up till 4 AM with the insight: use a proof-of-work-like scheme to measure time instead of work — later called Proof of History (technically a Verifiable Delay Function, but he didn’t know the term, so he thought it was unique).
Qualcomm background (TDMA/cellular radio protocols) shaped the approach: time-division multiple access applied to blockchain consensus.
Wife’s advice (from her Facebook-competitor startup): there’s a ~6-month window when “everyone knows a fast blockchain will be built” — raise then or miss the shot.
Bay Area culture gave instant permission: Dropbox managers said “come back in 6 months if it fails.” That founder-friendly network is unique to the Bay.
Fundraising: 1,000 meetings and the power of small funds
Raised with co-founder Raj Gokal (met through mutual friend; Raj brought BD/hustle/fundraising experience; Anatoly brought engineering/architecture).
Literally ~1,000 meetings. Early investors: Foundation Capital, Slow Ventures, Multicoin (tiny fund at the time, became most pivotal).
Multicoin did all non-engineering work: marketing, PR, intros, hiring. Lesson: small funds that make a big bet out of necessity often add more operational value than big brands.
Only became selective after they didn’t need money. Later criteria: VCs with board seats/connections in traditional finance (Stripe, banks) to help navigate partnerships.
Mission and product-market fit: “You can’t build a slow expensive blockchain anymore”
Mission stayed constant: unblock founders to build permissionless markets; optimize for price discovery/trading as the killer use case.
Hardest part: open-source permissionless systems compete with forks that verticalize/centralize for short-term speed, but long-term the open model wins.
“Cinderella crypto phase” (hopes/dreams funding without traction) is over. Now must build real businesses with revenue, users, token value accrual — “boring” but permanent.
PMF is unpredictable: trading turned out to be the wedge, but it was just as likely as any other narrative. Parallel to AI: big visions precede actual business models by years.
FTX collapse: the scariest moment
Happened right after record-breaking Breakpoint conference (ecosystem felt booming despite bear market). On flight back, Twitter exploded.
Fear: portfolio companies had treasury on FTX; VCs also wrecked; market tanking. Panic-called teams to check runway, tried to bridge funding.
80–85% of teams had been conservative (treasuries, not FTX). Lesson: put runway in short-term T-bills (treasurydirect.gov) — take startup risk, not custody/bank risk.
Test of decentralization: “you don’t know how distributed your ecosystem is until a pillar breaks.”
Stress management, co-founder dynamics, and leadership
Self-describes as “stress seeker” (USSR trauma?); cautious but pushes to the edge. Manages via sleep, no alcohol, weightlifting (knees blown from running).
Executive coaching helps — sometimes you need someone to repeat back what you already know.
Co-founder Raj: “work married.” Argued until out of arguments → Pareto-efficient options remain → coin flip, but CEO must decide. Trust = total delegation without micromanagement.
Picking co-founders: “just pick one… it’s like dating… you never know.” No spreadsheet filters; commit and trust.
Hardest decision: firing great engineers pre-launch to extend runway to the bone. Still friends with them.
Ongoing skill: balancing build vs. partner vs. wait. Argue to death with good executives; eliminate bad options.
Current focus: capturing more on-chain trading TAM
Growth is the perpetual struggle (some segments grow, some don’t; macro matters). Bear markets force signal-over-noise; bull markets drown signal in noise.
Optimistic: businesses with revenue exist → sustainable → crypto won’t die. Finance and law are permanent primitives; blockchain is a finance system.
Personal success definition: solving a fundamentally hard engineering problem that required sweat and deep design. Now peeling new layers: consensus/cryptoeconomics experts ripping out his old code (Ship of Theseus with engines).
Last 6 months (especially post-December) flipped from “slop” to usable for systems engineering.
Runs 4 persistent agents on different projects (perpetual derivatives risk engine, video games, etc.). Each agent ≈ 1 FTE; 24/7 → ~12x human output.
Philosophy: single-agent, max context, highest token quality. Multi-agent “skeomorphic” orchestration = midcurve (you don’t understand the product enough).
No company-wide AI mandates: trust ICs to pick best tools. Some use Claude, some ChatGPT, some none — delegate completely.
Early 10x engineer had 6 monitors + foot pedals; AI hasn’t matched that raw output yet, but ~2 years away on current trajectory.
Public persona, politics, money, and personal life
Introverted engineer pushed into “cult leader” Twitter role by Raj: “it’s your baby, it has to come from you.” Sales = do it 1,000 times.
Political shift: always anti-socialist (parents’ stories); loves arguing online. Not a big transition.
Money didn’t change life much: Bay Area tech success already exceeded parents’ dreams. Only scarce resource is time (kids vs. work). No luxury hobbies (not a foodie, car guy, traveler).
Hobbies: skiing/snowboarding (~100k vertical ft/season). Curiosity: quantum computing (if it works, solves protein folding/material science — the “atomic scale” problems classical compute can’t touch). 5 or 50 years? Unknown.
10-year dream: ski full-time, autonomous AI CEO running Solana while he’s on the lift. Won’t miss building — agents build for him.
Philosophy and closing thoughts
Belief: most people are good; assume good intent, trust by default — works almost always. Likely from parents.
Advice to 10-year-old immigrant self: “take it really easy and slow coming back to running after a break” (blew knees sprinting after 2-year burnout break; hopes quantum computing fixes cartilage regrowth someday).
Sports = flow state for creativity: ruminate, discard, find new ideas.
Life philosophy in one sentence: “Try to be mostly good. Assume other people are good. Things will work out.”
