Jeff Yan's Plan to Bring the Entire Financial Industry Onto Hyperliquid | Odd Lots

Summary

• Jeff Yan’s stated ambition for Hyperliquid is aspirational but serious: a neutral blockchain protocol whose “tagline, or aspirational goal, is housing all the finance,” which he frames as “a step-function improvement over electronic but fragmented markets,” the way electronic trading improved on pit trading. He hedges only on degree — “even 90% of all finance would be a huge success” — and insists the chain is not meant to displace incumbents: “legacy implies that they will be disrupted,” whereas he sees banks and exchanges adopting it as companies adopted AWS.
• The market structure is permissionless but rationed: staking 500,000 HYPE currently unlocks deployment functionality, with scarce market slots allocated through a Dutch auction held every 31 hours — an interval chosen because it is “relatively prime to the number 24,” so reset times rotate across time zones. The price falls linearly, the first bidder wins, and it resets to twice the last clearing price; liquidity itself is “a free market” order book.
• The regulatory unlock may run through deployers, not the protocol. Trump’s discussion of bringing Hyperliquid to the U.S. was “a huge surprise” to Yan. U.S. users are barred by the terms of service of Hyperliquid.xyz and other interfaces from using the derivatives trading, and Yan tells VPN users: “you’re breaking the rules.” In response to Joe’s framing, Yan says he can see a world where deployers are regulated while the protocol remains infrastructure analogous to TCP/IP.
• Joe’s democratization case rests on IPO gatekeeping: when trillion-dollar companies go public, “most of the wealth creation has happened at that point,” so instruments tracking pre-IPO shares could give an order of magnitude more people access to an opportunity. Joe acknowledges that this is not necessarily “a strict improvement over the system,” since synthetic exposure carries no claim on corporate cash flows; his argument is that access and investor education are both necessary.
• The infrastructure claim is striking for investors pricing compute: Hyperliquid Labs is a very small team, the project raised no VC money, roughly 30-something validators participate, and the top 27 by stake operate the network at any moment, with nodes requiring about 128GB of RAM. Yan, formerly at Hudson River Trading, built down to the consensus layer in 2023 because nothing existing could house finance: “the fact that it was hard was almost more inspiring to me… if we don’t at least give it a shot, who will?”
• Oracle design is a central risk, and Yan’s answers are specific: perp volume dwarfing the underlying is not necessarily a problem, citing S&P futures, so long as someone can say, “That 1% move is not correct. Whatever size you want, I’m here to trade against you.” The protocol clamps oracle updates to 1% per push but makes no subjective listing judgments; the SK Hynix bad-quote episode was described as the deployer’s “implementation issue,” addressed through a postmortem and refunds — “Apple ships operating systems with bugs, and they patch.”
• The hosts’ debrief flags what is structurally distinct: unlike Ethereum, Hyperliquid bakes exchange rules such as HIP-3 auctions and staking requirements into the protocol, helping prevent the kind of listing spam represented by 10,000 dog coins. Joe also identifies a deployer “metagame” around bidding for the next hot market. His bottom line: whatever happens to Hyperliquid specifically, perps and 24/7 trading “feel very inevitable.”

Deep dive

1. The pitch: a neutral layer “housing all the finance”

• Tracy’s setup is her “inverse pyramid”: in media, original reporting is now the smallest layer with content stacked on top; finance looks similar, with the underlying dwarfed by derivatives, tokenized stocks, prediction markets and other instruments. Joe calls it “financialization on steroids,” with Hyperliquid combining many of these developments.
• Yan defines Hyperliquid as a protocol whose “tagline, or aspirational goal, is housing all the finance.” He describes it as a neutral layer and “a step-function improvement over electronic but fragmented markets,” analogous to electronic trading improving on pit trading.
• The U.S. wrinkle: Tracy says Donald Trump discussed U.S. regulators bringing Hyperliquid to America. Yan says he was unaware Trump would mention it: “That was a huge surprise to me.” U.S. users are currently barred by the terms of service of Hyperliquid.xyz and other interfaces from using the derivatives trading. On the VPN workaround Tracy says she read about on Reddit, Yan is categorical: “It’s not a laughing matter… you’re breaking the rules.”

2. How the machine actually clears a trade

• Yan’s first-principles case for consensus in finance is agreement: “If we don’t agree on what the ledger of balances is… then all bets are off.” Under the assumption that two-thirds of the network’s stake is non-Byzantine, independent validators converge on a consistent global state.
• He distinguishes decentralization from regulation as “an orthogonal question.” Regulated entities can deploy markets on-chain and benefit from resilience. A 24/7/365 exchange operated by one company can suffer a single-point failure in matching, clearing or vendors; a diverse validator set can continue operating even if one or many validators go down.
• Joe’s walkthrough of a $1 million Bitcoin perpetual: the counterparty is whoever is in the order book, orders must have the collateral necessary to back them, and liquidation is “rules based,” not subjectively set by a person. It is an on-chain analogue of a traditional futures exchange’s default waterfall, with clearing and liquidation executed autonomously by HyperCore, a submodule of the blockchain state.

3. Deployers, 500,000 HYPE and the 31-hour Dutch auction

• Discussing the Nvidia perpetual associated with TradeXYZ, Yan prefers “deploy” to “issue.” A deployer currently stakes 500,000 HYPE to unlock deployment functionality, sends a transaction specifying the contract and defines or arranges the oracle system that tracks the perp’s exogenous index price. The deployer inherits the protocol’s order book, clearing system and USDC infrastructure: “This idea of modularity… is unique to on-chain financial systems.”
• Market slots are scarce because each order book and clearing system consumes network resources, so they are allocated through a Dutch auction every 31 hours. Yan says 31 is chosen because it is “relatively prime to the number 24,” causing reset times to rotate across time zones. The price falls linearly, the first bidder wins, and it resets to twice the last clearing price. Yan contrasts this with normal auctions, where bidders may be incentivized to “penny the other bidders, withhold information, and spam it at the very end.”
• Liquidity is unmanaged by the protocol: “It’s a free market. It’s an order book.” The protocol is the mechanism by which buyers and sellers meet, while orders must be collateral-backed. Yan does not personally track auction winners and says he is too busy to examine the data in depth, though the chain makes deployment activity transparent, verifiable and auditable.

4. The democratization thesis — and Joe’s cash-flow pushback

• Yan’s strongest concrete example is IPOs: when trillion-dollar companies go public, “you can pretty confidently say that most of the wealth creation has happened at that point.” He says the resulting gatekeeping is not intentional but “a byproduct of how the markets work right now,” making it a limitation of the existing system.
• Joe’s pushback is that an unaccredited investor buying an instrument tracking Anthropic shares receives no claim on future corporate cash flows. Joe then clarifies that he is not saying such access is “a strict improvement over the system.” Yan responds that futures likewise do not provide the underlying cash flow. Joe’s broader claim is that an order of magnitude more people could benefit from access, provided education helps them assess whether the opportunity fits their risk profile.

5. Origin story: rebuilding the stack from consensus up, with no VC

• Yan worked at Hudson River Trading after college and speaks positively about the experience. In 2023, he pursued a narrower goal: centralized crypto venues were offering perpetual futures with clear product-market fit, but the opportunity was to combine that user experience with on-chain properties such as self-custody and reduced exchange counterparty risk.
• The project expanded downward into infrastructure, like building consumer technology in the early 2000s and discovering that “maybe the internet doesn’t actually work that well.” Yan draws an AWS analogy. “The fact that it was hard was almost more inspiring to me… it felt more like, well, if we don’t at least give it a shot, who will?”
• The infrastructure footprint is deliberately modest: the current validator set is 30-something, with the top 27 by stake operating the network at any moment. Running a node requires roughly 128GB of RAM. Yan recalls a validator’s surprise that a humble machine could run what he described as a global financial system in miniature. It cannot run on a phone, but it is not intended to require football-stadium-sized data centers.
• Joe adds that Hyperliquid has no Wikipedia page; he says a moderator described the page as “typical crypto spam” and said the topic was “very doubtful” in notability. Joe jokes that the moderator may feel foolish if Hyperliquid eventually hosts most of finance.

6. Who exactly are you trusting? Pseudonymity and governance

• Joe asks whether a pseudonymous co-founder and relatively secretive team can remain behind the scenes if the chain becomes important financial infrastructure. Yan reframes the team as “much bigger than Labs”: a constellation of deployers, service providers and other groups participating in the ecosystem, including a Nasdaq-traded ETF or digital-asset treasury company whose representative, Dave, is well known.
• Joe presses that Hyperliquid Labs is still a governing entity because it sets or implements rules and software associated with HIP-3. Yan responds with an Ethereum analogy: developers convert community ideas into code, but “the core ideas behind the network are” not owned by developers. The protocol should be fair, transparent and supportive of financial tools.
• Joe’s sharper formulation asks what users are trusting Hyperliquid itself to do in a HIP-3 market. Yan says the trust assumption concerning Labs is that its node software adheres to the specification, as with any blockchain client. Market-level trust runs from the user to the deployer: anyone trusting a contract to track gold is trusting the deployer to publish something representing what it claims to represent.

7. Oracles, manipulation risk and oil perps at 1 a.m. Sunday

• Joe raises the Mango Markets incident on Solana, where an accused manipulation of the underlying token affected oracle-driven trading, and the SK Hynix bad-quote episode that caused losses. Yan describes the Hynix event as an “implementation issue”: TradeXYZ published a postmortem, the system had not behaved intuitively, and refunds were made. He compares this to software generally: “Apple ships operating systems with bugs, and they patch.”
• On perp volume overwhelming a thin underlying, Yan says the ratio does not necessarily need to be capped and can be healthy, citing S&P 500 futures and ETFs. The key is having informed, capitalized participants willing to reject an erroneous move: “That 1% move is not correct. Whatever size you want, I’m here to trade against you.” Price discovery can occur in the derivative while the underlying remains a reflection of real value.
• The protocol does not make subjective liquidity judgments: “We can’t do subjective liquidity judgments.” It instead enforces mathematical constraints. Between oracle updates, the relative price difference cannot exceed 1%; a larger update is clamped to 1%. Yan says this may help the leveraged state machine mutate orderly, though he presents that as a possible rationale rather than a certainty.
• The 24/7 WTI and Brent perpetuals that drew attention after the Iran war began are defined by the deployer’s contract specification, not by the protocol. Yan says a continuously updated oracle is not strictly necessary: funding can allow derivative prices to float for periods and then converge toward the underlying. Convergence may be slower when funding is less frequent, but the mechanism can still work.

8. The regulatory endgame — and the hosts’ debrief on what is genuinely new

• Joe asks whether the CFTC or CME could regulate deployers rather than the protocol. Yan calls this “an astute way to frame it” and says he can “very much see that world.” As with TCP/IP, the protocol can serve as technology and liquidity infrastructure while regulation sits at the layer where explicit controls exist — the deployers.
• Yan says CFTC core principles are “agnostic of the underlying technology.” He acknowledges that regulators are rightly conservative because the existing system broadly works and contains protections, but argues that on-chain transparency creates a useful property: anyone can audit that user balances equal the USDC held at an address. Solvency becomes “a property rather than a process.”
• On incumbents becoming deployers, Yan rejects the term “legacy”: “Legacy implies that they will be disrupted or displaced.” He compares adoption to AWS, where companies with their own servers eventually adopted cloud infrastructure. Banks, State Street or similar institutions becoming deployers would be important, but not remarkable.
• The hosts’ debrief identifies a distinction from Ethereum: Hyperliquid’s protocol itself establishes exchange rules, including the HIP-3 auction mechanism and staking requirements. Joe says this helps explain why Hyperliquid does not have the equivalent of 10,000 dog coins: the auction process prevents anyone from spamming listings whenever a theme becomes popular.
• Joe and Tracy also identify a deployer “metagame”: choosing which scarce market to bid on, perhaps a Kazakhstan index or another future popular instrument. Joe’s takeaway is that, regardless of Hyperliquid’s specific outcome, perpetuals and 24/7 trading “feel very inevitable.”