Solana Founder: Crypto Is About to Change Finance Like the Internet Changed Everything Else | EP#204
Summary
Yakovenko’s investable Solana thesis is narrower than the “everything coin”: “Bitcoin is store value, Ethereum is settlement, Solana’s execution.” SOL’s technical purpose is to make network spam costly, while its economic value comes from traders paying for priority when “the opportunity cost of being late” is the entire profit on a trade. That makes valuable order flow—not raw transaction count—the core variable in his proof-of-stake valuation framework.
Stablecoin growth is the near-term catalyst Yakovenko expects to pull finance on-chain. With legislation in place and projections of $1 trillion to $10 trillion in digital dollars over five years, he expects trade, settlement and capital markets to migrate toward programmable rails; he claimed moving dollars from Ethereum to Solana can be “a million times cheaper” than moving them between banks. In Solana’s own phone presale, stablecoins generated half of purchases without incentives and saved the merchant 2% of gross sales.
Solana’s long-term target is a single global execution layer synchronizing markets near their sources of information. Concurrent block producers in Singapore, New York and other signal centers would ingest events locally, while stake and trading profits incentivize operators to co-locate with valuable data. Yakovenko sees no academic barrier: “It’s purely an engineering problem,” although localized exchanges would remain useful for sub-microsecond market microstructure.
The larger disruption is not cheaper payments but forcing financial intermediation to justify its price. Yakovenko contrasted roughly 50,000 meme coins launched in a day with an IPO market so costly that a panelist said sub-$3 billion listings have effectively disappeared; he estimated, with an “I think” hedge, that Figment’s process consumed about $3 billion, roughly 10% of its market capitalization. Deep on-chain dollar markets could support direct listings, global capital formation and deterministic liquidation without layers of brokers, transfer agents and two-year bankruptcy proceedings.
Cheaper AI intelligence could create an explosion of investable markets and market-governed organizations. Yakovenko points to Polymarket, Kalshi and especially futarchy, where participants can be required to put financial stakes behind proposed decisions. AI can digest more signals into buy-or-sell judgments, making markets viable for millions of corporate or collective decisions—though he openly says the aggregate result is like “the intelligence of the anthill,” beyond any individual’s comprehension.
On-chain systems remove mechanical trust but do not abolish law, fraud or jurisdiction. Cryptographic provenance can function like the web’s SSL lock, while cap tables, transfer restrictions and liquidations become executable code; false representations remain fraud adjudicated where the issuer is incorporated. The crucial legal unlock, Yakovenko argued, is limited liability for smart-contract participants: once liabilities are not commingled, “the only thing that you have is the code.”
Yakovenko rejects the near-term dystopian labor thesis, while the panel emphasized that adoption still depends on better interfaces. He expects AI disruption to be smaller than the steam engine’s, with people producing more rather than becoming broadly unemployable. The Seeker phone is his concrete consumer wedge: hardware-wallet security with an Apple Pay-like experience, coupled with an attempt to challenge Apple and Google’s roughly 20% digital-commerce rake.
Deep dive
1. Cheaper intelligence makes far more markets economically viable
Yakovenko’s opening mechanism is simple: markets require intelligence, and intelligence is getting cheaper. AI can analyze more of the world’s signals, while a permissionless chain can create a market for almost any decision; Polymarket, Kalshi and experimental futarchy are early specimens.
He does not pretend to foresee the aggregate result. The outcome resembles “the intelligence of the anthill,” too complex for one person to hold, but he trusts market losses as a corrective: “The forcing function of losing money” exposes bad intelligence and forces participants to adjust.
The Truth Terminal discussion supplied the comic edge. Meme coins may be a “weird Keynesian beauty contest for attention” through which autonomous agents pay inference costs; Yakovenko joked that AGI’s real test may be whether it can tell jokes or keep a toddler entertained. Wissner-Gross suggested corporations already serve as economic bodies for AIs.
2. Solana was designed for execution, not monetary purity
Yakovenko’s clean taxonomy: “Bitcoin is store value, Ethereum is settlement, Solana’s execution.” He was uninterested in the first two because execution—moving bits around the world as quickly as possible—is the hands-on engineering problem that attracted him.
At Qualcomm, beginning in 2003, he became a self-described “performance nerd,” optimizing how bits moved through memory. In 2017 he was also building deep-learning machines for transformers and mining crypto in the background so the GPUs could offset their capital cost.
Two coffees and a beer at San Francisco’s Café Soleil kept him awake until 4 a.m., when he realized recursive SHA-256 hashing could establish a hard-to-fake minimum passage of time. A faster chip might beat the process modestly—he cited roughly a factor of two—but not arbitrarily.
His analogy was cellular TDMA: simultaneous transmissions collide, so a clock schedules channel access. Bitcoin’s long block interval likewise reduces collision risk by sacrificing channel efficiency; his back-of-envelope conclusion was that a clock could make the system at least 10,000 times faster. “I built TDMA for blockchain.”
3. Proof of History can retire after identifying the real bottleneck
Yakovenko now thinks he initially sold Proof of History as a novel consensus algorithm when its deeper contribution was exposing a channel-efficiency problem. That was difficult to communicate while he was focused on implementation and shipping.
Solana’s success let it recruit a consensus team from ETH Zurich to develop next-generation approaches such as Alpine Glow. Those designs can address channel efficiency without retaining Proof of History’s complexity or much of the original code.
His change of mind is unusually explicit: the founding invention can be discarded once its function is understood. “Now I get to see my baby retired,” he said, treating architectural replacement as evidence of progress rather than a repudiation of Solana’s origin.
4. The end state is one global machine with block production near the signal
Today’s NASDAQ or NYSE matching engine is effectively centralized memory in one physical location, enabling sub-microsecond execution. But if a ship carrying iPhones sinks near Singapore, that information must still cross fiber at light speed before a New York market can price it.
Yakovenko’s proposed layer runs concurrent block producers in Singapore, New York and other signal centers. Proof-of-stake participants would move stake toward those locations because being first to ingest market-moving data captures the trade’s profit; being second can mean losing all of it.
Trades are small messages compared with streaming 4K video to a remote driver. Solana’s task is to feed those messages into “this single giant state machine” as quickly as the network can synchronize, moving conceptually from TDMA’s alternating channel to CDMA-like concurrency.
Localized markets would survive because co-location still creates a valuable “light cone” around queues and microstructure. The global layer synchronizes them rather than abolishing them. Yakovenko offered no firm date, only the categorical claim that no computer-science theorem prevents it: “It’s purely an engineering problem.”
5. Digital dollars are the bridge from legacy settlement to on-chain finance
Yakovenko expects stablecoin legislation and projections of $1 trillion to $10 trillion in digital dollars over the next five years to “massively accelerate things.” Those dollars would be minted for domestic and global trade, leaving businesses with a simple programmable interface to money.
His provocative cost comparison was that moving dollars from Ethereum to Solana is “a million times cheaper” than moving them between two banks. The exact application matters less than having liquid dollars already available on-chain for settlement and collateral.
DeFi, in his framing, is programmable escrow: money is placed at risk and conditionally released according to an oracle, market price or other future signal. That primitive extends to borrowing, business financing and any activity requiring continuously updated risk calculations.
6. Financial infrastructure works—but human coordination makes it slow
Yakovenko praised the US financial system as robust precisely because regulation incorporated decades of trial and error. Its weakness is that it was built before the internet and depends on people, making two-day settlement difficult to compress into one day, four hours or milliseconds.
Blockchains were designed after ubiquitous networking and can replace some procedural assurance with cryptographic correctness. He even wondered whether NATO could reliably partition today’s internet, underscoring how much these systems assume persistent global connectivity.
A panelist’s boardroom example made the friction tangible: a director claimed there had not been an IPO below a $3 billion valuation in roughly a decade because legal and administrative costs are unaffordable. The panel contrasted that scarcity with approximately 50,000 meme coins launched in a day.
A panelist pressed for a wealth-creating “killer app,” not merely payment efficiency. Yakovenko’s answer was deliberately macro rather than consumer-facing: better global pricing, fewer financial errors, faster GDP growth and declining poverty may matter enormously without producing one application his parents would recognize.
7. On-chain capital markets could force finance to earn its fees
Yakovenko cited Figment’s IPO—hedging with “I think”—as having lost roughly $3 billion in the process, about 10% of the company’s market capitalization. With deep on-chain dollar markets, a company could instead direct-list through a smart contract.
Analysts or other third-party providers could still be paid, but only for information, liquidity or another identifiable service. Listing access itself would be nearly free, pushing each intermediary’s compensation toward “the actual value it’s providing.”
One panel statistic put 40% of corporate American profits over the past decade in financial services. Yakovenko called much of finance “a tax” consuming more value than it creates; the intended endpoint is talent anywhere raising capital from investors anywhere, including an engineer in Ukraine funded from China, America or Australia.
His geopolitical framing was equally strong: “The only place for finance that can compete with the United States is the internet.” Crypto replicates US-style access to capital online and, in his view, may export American values and influence rather than simply displace the country.
8. Cryptography removes administrative trust, not fraud or courts
Securities rules arose partly because a railroad certificate might be counterfeit, refer to a nonexistent railroad or rely on false disclosures. Yakovenko praised Paul Atkins, Hester Peirce and David Sacks, then argued that blockchains change which safeguards require human administration.
A token can be traced cryptographically to an issuer’s keys and disclosures, much like the SSL lock made online credit-card entry trustworthy. The software verifies provenance and transfer without asking a broker or transfer agent to certify every step.
Wissner-Gross’s pushback was that contracts still rely on courts and rule of law. Yakovenko agreed on substance: lying about a fundraising purpose remains fraud, and the issuer’s local jurisdiction ultimately enforces liability. The chain removes “all the boring stuff” around ownership and transfer, not accountability.
A ledger could become the cap table itself, with smart contracts enforcing clawbacks, company approvals and transfer restrictions. Yakovenko called the implementation “a bunch of database kind of gnarly code,” while Diamandis noted that startup disputes already often go to JAMS because courts are too slow.
9. Programmatic liquidation compresses bankruptcy from years to blocks
FTX illustrated what remains broken when a crypto-trading company itself operates as a centralized institution. Its bankruptcy took about two years to unwind; a shared ledger could have shown immediately that more money was leaving than entering.
Aave’s liquidations effectively run a bankruptcy process on every Ethereum block, which Yakovenko described as every 12 seconds. Kamino performs comparable checks every 400 milliseconds on Solana, preventing undercollateralized positions from accumulating into a multiyear ownership dispute.
The investor benefit is deterministic downside handling across jurisdictions: participants can know in advance how collateral will be processed and receive the available recovery within roughly a minute. Diamandis estimated FTX investors received only 50–60 cents on the dollar, attributing substantial loss to procedural friction.
10. Solana defines decentralization as permissionless participation
Salim Ismail’s foundational distinction was that crypto’s value is not merely being digital but being programmable. His Byzantine-generals analogy emphasized authenticated coordination and the double-spend problem: a network must remain correct even when a participant sends malicious or conflicting messages.
Yakovenko dismissed Crypto Twitter’s tribal definition—whatever coin someone owns is decentralized—as useless. His test is operational: can anyone run a validator, produce blocks, transact, deploy code, copy the state and recover the network without third-party approval?
A global financial layer must let banks in France, England and New York participate throughout the stack because even allies will not trust one another completely. It must also survive misconfigured or deliberately adversarial nodes.
Formal proofs are insufficient when implementations still fail. Solana tried to solve performance and decentralization simultaneously while behaving like a fast-moving startup; Yakovenko described the resulting implementation problems and repairs as “growing pains” involving substantial “blood, sweat, and tears.”
11. SOL captures value by pricing scarce priority, not ordinary commerce
Yakovenko rejected the “everything coin” label literally. SOL’s underlying purpose is to prevent an adversary from sending infinite messages: “The only problem that it can solve is decentralized spam.” For purchases, he explicitly encouraged merchants to use USDC, PayPal’s USD or another digital dollar.
What surprised him was that anti-spam machinery could capture substantial value. NFT congestion exposed a database-hotspot problem: when money is at risk, traders voluntarily pay more to enter the block first, turning priority into network revenue.
Consequently, transaction volume is less important than “the opportunity cost of being late.” Block producers stake SOL for the right to make a percentage of blocks, then receive tips from users competing for first access to profitable trades.
Tips could arrive in SOL, USDC or “a sack of potatoes”; the valuation link is the stake required to access that revenue. Yakovenko suggested comparing the risk-adjusted return with Treasury bills and using Kelly-style allocation. Blockworks has compiled relevant network data; Bitcoin remains a “special snowflake” because its Sybil resistance is energy, not a staked token.
12. Crypto can remain invisible while stablecoins win business adoption
Yakovenko thinks only about one person in 200 needs to understand crypto—the Linux-capable friend in an extended wedding-sized trust circle. Most users can rely on that person’s guidance while the infrastructure stays under the hood.
Solana Mobile’s presale offered credit cards and stablecoins; half the purchases arrived via stablecoin without incentives. For Southeast Asian buyers, that rail was often easier than an international card, while the merchant saved 2% of gross sales—“three or four engineering salaries” on one product.
That evidence supports his view of blockchains as primarily B2B infrastructure with consumer touchpoints. Ismail’s counterweight was usability: decentralized exchanges and liquidity pools can still require a “hardcore crypto sherpa,” so adoption depends on further interface improvements.
13. Shared digital state turns even valueless objects into markets
Meme coins surprised a team whose original tagline was “blockchain at Nasdaq speed.” Yakovenko called them “assets with no value and all price,” yet found the behavior familiar from Ultima Online, where players traded gold and wood and he wrote scripts to mine digital resources automatically.
The broader rule is that shared state attracts economic activity. Digital wood has no external backing, but participants still trade it for fun; video games now teach younger users that digital assets can carry scarcity, although Wissner-Gross noted that board games long used fictional money too.
Ismail’s examples showed rapid institutional evolution: Bored Ape owners could license character IP; CyberKongz issued tradable bananas; failed NFT collections created a turnaround and M&A market. After rug pulls, some projects required tokens to be held for six months before sale.
Yakovenko’s key point was bounded catastrophe risk. A smart contract tells a buyer what can happen to the bananas without relying on an eBay stranger and cashier’s check. That assurance permits experimentation that would be uneconomic with bespoke lawyers—or would never happen at all.
14. On-chain organizations need limited liability more than permission
A token financing and a Y Combinator SAFE both hinge on whether “this founder is a jackass.” The SAFE provides legal guarantees but little liquidity; a token provides immediate secondary-market access and reliable transfer. Either can still lose everything, while the success upside may be similar.
Wissner-Gross argued that virtual labor and AI output could migrate outside dollar valuation, payroll taxes and national rails. Yakovenko pushed back: most markets may still clear through US dollars, and US residents owe mark-to-market taxes regardless of the asset—an obligation backed by a state capable of collecting it.
Wissner-Gross cited MetaDAO as a current model for an on-chain corporation. Futarchy asks participants to price its token conditional on a decision—effectively, “I will buy your Apple stock higher if Apple builds this product”—so governance requires capital at risk, not a costless shareholder vote. Salim said a couple of related projects had attracted more than $150 million in commitments.
Wyoming, Panama and other jurisdictions already offer DAO-friendly structures, but Yakovenko argued the decisive rule is that one participant’s liability must not be commingled with everyone else’s. Once laws or courts recognize that limitation, “the only thing that you have is the code,” and losses caused by accepted contract risk have no external defendant.
15. AI, mobile hardware and abundance still leave human choice unresolved
AI can turn worldwide signal into trading actions and scale futarchy across large organizations. Wissner-Gross argued that counting agents will sound dated within a decade: “agents is a very 2025-era term”; the reality may be a continuous information funnel to GPUs.
On employment, Yakovenko opposed Balaji Srinivasan’s unrest thesis. His parents came to the United States from the USSR in 1992 with $50 per person, and he expects AI’s labor disruption to be smaller than the steam engine’s: jobs become safer and easier, productivity rises and poverty falls. Wissner-Gross’s experience watching Claude code with “half an open eye” illustrated the value of experienced oversight.
Ismail offered local resilience rather than centralized rescue: communities could quickly issue tokens for mostly local transactions. His analogy was Ireland’s 1970s banking strike, when people kept passing checks for roughly six months on the assumption they would eventually clear.
Yakovenko’s preferred current product is Seeker, which repurposes secure-enclave technology originally built for DRM into a phone-based hardware wallet with an Apple Pay-like experience. The second objective is economic: challenge Apple and Google’s roughly 20% digital-spending rake with lower, volume-based transaction fees.
AI-generated content might eventually overcome the enormous catalog advantage of YouTube or TikTok. Yakovenko thought today’s systems remain too expensive and insufficiently good, but could produce infinite entertainment streams within five years; he also called that prospect “kind of a scary thought” because shared human curation helps create culture.
The closing wealth debate never fully converged. Ismail chose time and healthspan; Blundin chose compute; Wissner-Gross proposed “future freedom of action”; Ismail also emphasized purpose and physical manipulation through nanotechnology. Yakovenko moved from Shannon-style information-processing capacity to “degrees of freedom,” while warning that humans will still want more than whoever has less.
In a Dyson-swarm future, sufficiently powerful computation might allocate resources without commerce. Yakovenko nevertheless resisted that endpoint: “Personal freedoms are far more important than efficiency,” and competition for tokens—even bananas—can supply purpose. Abundance may change the unit of wealth without ending the human drive to choose.