Hyperliquid Labs has confirmed Singapore as its registered headquarters. Co-founder Jeff Yan and a development team of roughly eleven people relocated there in 2024. Company documents pointed to that registration, and recent hiring material still asked whether candidates could work from the Singapore office. In the same breath, the Monetary Authority of Singapore said it was not aware that Hyperliquid is regulated in any major jurisdiction. People familiar with the authority’s thinking pointed to decentralization as the reason the platform is treated as outside its remit. The company, for its part, acknowledged that it operates without regulation, said it has never claimed permission from the authority, and signaled that it is open to working with regulators.
That is a lot of precision packed into a short sequence. Headquarters, yes. License, no. Ban, also no. Investor alert, yes. A proposed American pathway through other registered firms, maybe, and only if approvals land. If you trade size on perpetual markets, the distinction is not academic. It is the difference between a city on a letterhead and a supervisor who can pick up the phone.
A City Address Is Not the Same Thing as a License
Start with the plain reading, because the plain reading is already strange enough. A firm can be physically present in a financial center and still not be a regulated firm in that center. Presence answers a logistics question. Authorization answers a legal one. Mixing the two is how rumors get born, and how they get expensive.
Hyperliquid’s Singapore presence, as confirmed, is a development base. The co-founder moved. The core team is small. Job posts as recent as the week before the confirmation still treated the office as a real workplace, not a mailbox fiction. I do not think that detail is trivial. Eleven people who can sit in one room and ship protocol changes are a different animal from a brand that rents a serviced address and calls it home. The concentration is the point. A lean team can move fast. A lean team also means key-person risk sits in a very small circle.
The regulator’s line was cooler. It said it was not aware of Hyperliquid being regulated in any major jurisdiction. That is not a finding that the protocol is illegal everywhere. It is a statement about the absence of a license the authority can point to. People close to the authority’s thinking tied the jurisdictional view to the platform’s decentralized design. In other words, the argument on the regulatory side is not “we have inspected this and blessed it.” The argument is closer to “this does not fall inside the box we supervise.”
We are not aware that Hyperliquid is regulated in any major jurisdiction.
Monetary Authority of Singapore, in a direct response
Hyperliquid’s own stance matches the gap rather than papering over it. The company said it remains unregulated, that it has never presented itself as authorized by the Singapore authority, and that it is willing to cooperate. I would rather hear that than a foggy claim of “compliance in principle.” Clarity about a missing license is more useful than a slogan.
What the Relocation Actually Shows
Relocation stories in crypto tend to get written as strategy memos. Tax. Talent. Time zone. A friendlier policy climate. Sometimes all of that is true, and sometimes the memo is written after the flight is booked. Here the public record is narrower, and narrower is better.
What is on the table is a registered headquarters, a 2024 move by the co-founder and a small team, and recruiting language that still treats Singapore as a place people are expected to work. What is not on the table is a payments license, a capital-markets license, or any statement that the trading application itself is supervised. If you only remember one split, remember that one.
- A registered headquarters answers where the company says it is based.
- A team on the ground answers who is building the software day to day.
- A license answers who can examine the business, set conditions, and sanction breaches.
- An investor alert answers a narrower fear: that someone might assume a license exists when it does not.
Perhaps the most interesting aspect is how ordinary the office detail feels next to the market the protocol supports. Perpetual futures are not a newsletter product. They are leveraged, they settle continuously, and they attract flow that used to live inside traditional derivatives venues. A small engineering group in a city-state can still sit at the center of that flow if the matching logic is onchain and the brand is strong. Geography did not disappear. It just stopped being the thing that clears the trades.
How the Authority Drew the Line
Jurisdictional line-drawing is where crypto arguments go to get philosophical, and also where real money gets stuck. The familiar version goes like this. If the order book is a smart contract, if anyone can interact with it, if no single company takes the other side of your trade, then whose customer are you? The trader wants a venue. The regulator wants an entity. The protocol offers a function. Those three nouns do not map cleanly.
People familiar with the authority’s position described decentralization as the basis for treating Hyperliquid as outside its jurisdiction. That is a specific claim, not a universal rule about every onchain market. It depends on how the activity is structured, who is soliciting users, where customers sit, and which token services are actually being provided. I have watched too many threads treat “decentralized” as a magic cloak. It is not a cloak. It is a factual description that a supervisor may accept, reject, or accept only in part.
The company’s willingness to engage matters in that light. A team that says it will work constructively with authorities is leaving the door open for a future license, a no-action style conversation, or a redesign that pulls certain services inside a regulated wrapper. None of that has happened just because the door is open. Engagement is a posture. Authorization is a document.
The Investor Alert Is a Warning Label, Not a Shutdown
On June 26, the authority added Hyperliquid to its Investor Alert List. The entry names both the foundation website and the trading application. If you have never read how that list is described, it is easy to overshoot. The list is a notice about entities that investors may wrongly believe are licensed, authorized, or regulated by the authority. It is not described as exhaustive. It records information available when an entry is published.
Hyperliquid’s response was unusually direct, and worth keeping in the original shape of the claim.
IAL listing does not constitute a ban, an enforcement action, or a finding of wrongdoing.
The platform also said its permissionless infrastructure was unchanged, repeated that it had never presented itself as authorized, and said it would work with regulators and institutions on clearer rules for onchain finance. I think that reply is doing two jobs at once. It stops a panic reading. It also refuses to convert the listing into a badge of innocence. No ban is not the same sentence as no concern.
Why do lists like this exist? Because retail investors use city names as shortcuts. Singapore, London, New York, Zurich. Each name carries a halo from firms that really are supervised there. An alert list is a clumsy but practical way to say the halo does not automatically transfer. You can dislike the method and still see the problem it is trying to solve.
For a trader, the operational consequence is narrower than the headline. Access to the protocol did not flip off because a list was updated. The legal character of that access did not suddenly become a local license either. If your risk process treats an investor alert as “they got banned,” you are misreading the document. If your risk process treats it as “nothing to see,” you are also misreading it.
Singapore’s Overseas Token Rule Still Sits in the Background
The headquarters story landed in a city that had already tightened a different door. In 2025, following a notice at the end of May and a clarification in early June, covered providers of certain overseas digital token services faced a hard date. Obtain a license or stop the covered activity by June 30, 2025. No extra transition period was offered.
The framework targets digital token service providers that serve only customers outside Singapore, in digital payment tokens or in tokens that represent capital market products. The authority set a high licensing bar and said it would generally not grant licenses for that business model. The stated reasons were blunt: higher money-laundering risk, and the practical difficulty of supervising a provider whose substantive regulated activity happens overseas.
Read that twice if you allocate capital for a living. A model built on serving foreign customers from Singapore, without local customers as the core, was not treated as a welcome export industry. It was treated as something the authority mostly does not want to license. Providers without the required license were told to cease covered services from the effective date.
Does that rule automatically capture a decentralized perpetual venue with a local development team? The authority’s later posture on Hyperliquid suggests it does not see the platform as inside its jurisdictional box, at least on the reasoning shared with reporters. I would not treat that as a permanent carving. Rules get reinterpreted when products start to look like solicitation, custody, or a service aimed at local users. A development office is one fact. Marketing into a market is another. Onboarding local residents would be a third.
Three different questions, often collapsed into one: Where do the builders sit? Who is being solicited? Who has the power to examine the books?
Decentralization Is an Argument, Not a Mood
I get why builders reach for the word. A matching engine that lives in contracts, an order book anyone can read, liquidations that do not wait for a support ticket. That architecture really is different from a brokerage app with a hidden internal book. Different is not the same as unsupervised-by-definition.
Supervisors tend to look through the diagram and ask who can change the parameters, who is paid, who publishes the interface people actually use, and who holds the relationship with the customer. A foundation website and a trading application both appeared on the investor alert entry. That pairing matters. Even if the core matching is permissionless, the front door people click is a product. Products have publishers.
Hyperliquid’s line has been that the infrastructure stayed permissionless and that the listing changed none of that. Fair enough as a technical statement. Permissionless infrastructure can still be wrapped, indexed, marketed, or cleared by firms that are very much permissioned. That is exactly the shape of the American proposal that showed up later. The protocol does not have to become a broker for a broker to build a fenced yard on top of it.
There is a metaphor I keep reaching for, and it is not elegant, but it works. Think of a public square with a chess clock bolted to a table. Anyone can sit down. The city did not license the clock. A cafe across the street might still sell you a membership, keep your coat, and call the game a tournament under house rules. The clock is still the clock. The membership is a different business. Regulators are usually more interested in the membership than in the bolts.
What Traders Should Separate from the Headquarters Headline
Price does not wait for jurisdiction memos. Perpetual markets mark to funding, liquidate on a schedule, and punish anyone who confuses narrative with margin. Still, the legal wrapper changes the kind of risk you are actually running. I would split it like this, on a notepad, before I let a headline move size.
- Protocol risk: smart contract behavior, oracle inputs, liquidation design, upgrade keys if any exist.
- Operator risk: the small team, the foundation, the interface, and what happens if those people stop shipping.
- Access risk: whether your path in is a wallet click, a third-party front end, or a registered intermediary.
- Legal risk: investor alerts, overseas token rules, and whatever your own home regulator thinks of offshore perpetuals.
- Counterparty risk: if you later use a cleared wrapper, the clearinghouse and the carrying broker become the names that matter.
A Singapore confirmation touches the second item and brushes the fourth. It does not retire the first. It does not, by itself, create the fifth. People who trade as if a city registration collapses all five into one green check are doing storytelling, not risk management.
There is also the social proof problem. Screenshots travel faster than footnotes. A line like “based in Singapore” will get cropped, reposted, and paired with a chart. The missing half of the sentence, the part about no known license in any major jurisdiction, travels slower because it is less fun. In my experience, the slower half is the one that saves accounts.
The American Idea Is a Different Building Entirely
Offshore-style perpetual futures for retail customers do not have a neat on-ramp under existing United States law. That is the sober view from Ashley Ebersole, a former senior counsel at the securities regulator and now co-founder and chief legal officer at tx, laid out in late August. Commodity-linked contracts would likely sit with the derivatives regulator. Products tied to securities could pull in the securities regulator as well. A compliant structure could need registrations across the venue, the clearing layer, and the intermediaries, plus rules or exemptions for the products themselves.
Even if both agencies leaned in, Ebersole put the timeline around ten to twelve months. Agencies would have to identify their authority, build a framework, and run whatever proposal, comment, and implementation steps the law requires. Reliance on existing powers or exemptions could shorten that. Litigation, a turf fight, or a need for new legislation could stretch it. I read that range as a planning band, not a promise. Regulatory calendars slip for reasons that have nothing to do with the quality of the product.
Then, in mid-September, Payward, the parent of Kraken, said it planned regulated Hyperliquid perpetual markets using the protocol’s HIP-3 infrastructure. The proposal needs approval before eligible United States customers can trade. Bitnomial Exchange would create and administer the contracts. Bitnomial Clearinghouse would handle clearing and settlement. Matching and recording would use Hyperliquid’s onchain order book. Customer accounts would be carried through NinjaTrader Clearing, and traders would need to clear onboarding and appear on allowlists at both NinjaTrader and Bitnomial.
Notice what that design does not do. It does not hand a retail customer the entire open menu of Hyperliquid markets. It limits them to the regulated products Bitnomial deploys under its exchange rules. Hyperliquid was named as the first protocol for the initiative. No launch date was given. Approval is the gate, and the gate is not decorative.
If that structure ever goes live, the user’s legal relationship looks nothing like a wallet connected to a public front end. The user would be a customer of registered firms. The protocol would be plumbing. Plumbing can be excellent and still not be your counterparty. I suspect a lot of commentary will blur those roles the week any approval headline hits. Do not blur them in your own notes.
Two Doors, Two Sets of Rules
It helps to put the paths side by side, because they answer different customers. One path is the permissionless market as it exists. The other is a proposed fenced market that borrows the order book and wraps it in exchange rules. They can share a name and still not share a regulator, a customer agreement, or a dispute process.
| Question | Open protocol path | Proposed US wrapper |
| Who confirmed a Singapore base | Hyperliquid Labs, registered headquarters | Same underlying protocol, different access firms |
| Known major-jurisdiction license for the platform | Authority says it is not aware of one | Depends on Bitnomial, clearing, and broker approvals |
| Investor alert in Singapore | Listed June 26, described as not a ban | Does not by itself authorize or forbid the US plan |
| Who matches the trade | Onchain order book | Same order book, under exchange-administered contracts |
| Who carries the customer account | Typically the user and the interface they chose | NinjaTrader Clearing, after allowlist onboarding |
| Product set | Whatever the protocol lists | Only contracts Bitnomial deploys under its rules |
| Status | Live and unregulated, by the company’s own account | Proposed, approval required, no date set |
Tables flatten nuance, so here is the nuance. A shared order book does not merge the customer protections. If a cleared contract fails to list, the open market does not become cleared by sympathy. If the open market has a rough day, the proposed wrapper does not inherit that day unless it is actually live and routing. Sequence matters. Approval first, then flow, then any claim about “regulated Hyperliquid.”
Why a Small Team Changes the Feel of the Risk
Eleven people is not a rounding error. It is a design choice. Protocols with giant payrolls still ship bugs. Protocols with tiny payrolls ship a different kind of fragility: vacations, visas, burnout, a single engineer who understands a liquidation edge case. The Singapore move makes that circle easier to picture. They are not an anonymous swarm distributed across twelve time zones in the way the marketing sometimes implies. A core group chose a city and stayed.
I do not say that as a knock. Concentrated talent is how a lot of serious trading systems got built, in crypto and outside it. The honest version includes the bus factor. If the registered headquarters is also the room where upgrades are argued, then operational continuity and legal identity are sitting in the same postcode. That can help a regulator find someone to talk to. It can also make the “no one is in charge” story harder to tell with a straight face.
Hiring language is a quiet tell. When ads still ask if you can work in the office, the office is part of the operating model, not a relic of the relocation announcement. For anyone underwriting the project, that is mildly comforting and mildly concentrating. Comforting, because software of this kind benefits from people who can argue in person. Concentrating, because the map of responsibility gets smaller, not larger.
Money Laundering Concerns Were Already on the Record
The overseas token framework did not appear out of nowhere. The authority explained the high bar by pointing to money-laundering risk and to the trouble of supervising activity whose substance is abroad. You do not have to agree with the breadth of that policy to see why a perpetual venue makes supervisors nervous. Leverage, global access, fast settlement, and pseudonymous wallets are a combination that compliance teams already know how to worry about.
None of that is a finding about Hyperliquid’s customers. The investor alert was not framed as an enforcement case, and the company said as much. The point is structural. A jurisdiction that has already said it generally will not license an overseas-only token service model is an awkward place to imply, even accidentally, that a trading app enjoys local blessing. Hyperliquid has been careful not to imply it. Other people on the internet will be less careful. That is a communications risk as much as a legal one.
Institutions notice lists. A treasury desk, a fund administrator, a prime broker’s compliance officer. They do not need a ban to hesitate. They need a question they cannot answer cleanly in a memo. “Are they regulated in Singapore?” is exactly that kind of question. The accurate answer today is no, with a headquarters yes, and an alert on file. Memos like that slow capital down. Sometimes that is the intended effect.
What Clear Rules Would Actually Have to Cover
The company said it wants to support clear rules for onchain finance. Fine. Clarity is not a press adjective. It is a list of jobs someone has to do. If I were sketching the minimum, not as a lawyer’s opinion but as a trader who has watched venues get halfway regulated and then stall, I would want answers on a handful of boring topics.
- Who is allowed to offer the interface, and to whom.
- What disclosure a leveraged perpetual owes a retail user before the first click.
- How liquidations are explained in language a non-specialist can survive.
- Which entity, if any, handles complaints when the interface and the contract disagree.
- How market abuse is watched when the book is public but the incentive to spoof does not vanish.
- What happens to open interest if an upgrade, a halt, or a parameter change is required.
The proposed United States structure tries to answer several of those by moving them onto an exchange, a clearinghouse, and a carrying broker. That is the old architecture, pointed at a new matching layer. It may be the only architecture current law can digest. It is also slower, narrower, and less “anyone with a wallet” than the open market. Both things can be true. Speed and permission are a trade, not a moral ranking.
Ebersole’s ten-to-twelve-month band is a reminder that even a friendly reading of existing powers is not a weekend project. Identify authority. Write the framework. Take comments if comments are required. Implement. A disagreement between agencies, or a court case, and the band breaks. Anyone pricing a token or a points program off an imminent retail perpetual license is guessing with extra steps.
HYPE, Attention, and the Temptation to Overread
A headquarters confirmation will get stapled to price talk. That is inevitable, and mostly unhelpful. Regulatory geography can change who is allowed to touch a market. It does not, on its own, change funding rates, depth, or whether a liquidation cascade finds a buyer. I have found that the weeks after a jurisdictional headline are when narrative traders and basis traders briefly pretend they are the same person. They are not.
If anything, the Singapore story is a constraint on storytelling. You cannot honestly sell the venue as a regulated Singapore exchange. You can say a development company is registered there, that the authority does not see the platform as inside its remit, and that an investor alert is on the books without being a ban. That paragraph is less viral. It is also the one a serious allocator can repeat without flinching.
Token holders have a further wrinkle. Governance, fees, and protocol value do not map one-to-one onto the legal status of a labs company. A wrapper that routes institutional or American flow through HIP-3 could increase the importance of the order book without making the token a regulated security or a regulated anything. Equally, a wrapper could capture the customer relationship at the broker layer and leave the protocol as infrastructure with a toll that someone else sets. The September outline did not settle that economics. It named a first protocol and withheld a date. Withholding the date was the responsible part.
A Reader’s Checklist Before the Next Rumor
Rumors will fill the gap, because gaps are where rumors live. The useful habit is a short list you can run in two minutes. I keep something like this when a venue’s geography becomes the story.
- Did the company confirm a fact, or did a screenshot imply one?
- Is the fact about an office, a license, or a customer permission?
- Did a regulator speak, or did “people familiar” speak? Both can matter. They are not the same speaker.
- If a list was updated, does the list’s own description say ban, or does it say warning about a false impression of licensing?
- If a US product is mentioned, is there an approval, or a plan that requires approval?
- Which legal entity would carry your account if you used the regulated version?
- What remains true about the open market if the regulated version never launches?
Run those and the current picture stays stable. Company confirmation on the headquarters. Regulator unaware of a major-jurisdiction license, and inclined to see decentralization as placing the platform outside its box. Alert list entry that the company correctly notes is not a ban. Overseas token regime in the background with a high bar and a past hard deadline. American access sketched through Bitnomial and NinjaTrader, still gated. Open market unchanged in its permissionless claim.
Office confirmed + license absent + alert listed + ban denied + US wrapper proposed = do not collapse into one headline.How Other Financial Centers Tend to Treat This Shape
I am not going to pretend every capital has published a Hyperliquid memo. They have not. The pattern, though, is familiar enough to be worth naming without turning it into fake precision. Centers that want to host builders often separate technology companies from customer-facing financial services. A software employer can rent space, hire, and pay tax. A firm that offers leveraged derivatives to the public is asked for a different permission, sometimes several.
That split is why “we moved the team” and “we are regulated” keep getting confused on social platforms. The first can be finished in a quarter. The second can take a year and still end in a no. Singapore’s overseas token stance shows one version of the no: a general reluctance to license a model whose customers and substance sit somewhere else. The United States discussion shows another version: not a blanket no, but a stack of registrations that existing agencies would have to stretch or sequence.
Europe’s markets regime, Britain’s promotions rules, and various Asian licensing gates all pull on different threads: the instrument, the customer’s location, the marketing, the custody. A perpetual that never touches fiat custody can still be a derivatives contract in the eyes of a home regulator. Your home regulator does not lose interest because the matching happens in a contract deployed from a team in Singapore. Residence of the code is not residence of the trader.
So if you are outside both Singapore and the United States, the headquarters news is even less of a permission slip. It tells you where builders sit. It does not tell you that your local rules have been satisfied. Anyone selling access on the back of a city name is asking you to outsource a question you still own.
Institutions Will Read This More Slowly Than Traders
Retail flow can treat an alert as a meme and a headquarters as a flex. An institution has a committee. Committees like named supervisors, audited controls, and a person who can sign a side letter. The Payward outline is aimed at that temperament. Allowlists, a designated contract market style administrator, a clearinghouse, a carrying firm. It is the vocabulary those committees already speak.
Whether they get to speak it about Hyperliquid markets depends on approvals that have not been announced as granted. Until then, the institutional story is a proposal with a credible cast, not a live product. I would rather see that restraint than a soft launch that compliance has to unwind. Unwinding is how good market-structure ideas get a bad reputation.
There is a commercial logic underneath, and it is not subtle. Perpetual liquidity that already exists onchain is attractive if it can be referenced without importing the entire legal character of the open venue. HIP-3 is the bridge in the announcement. Bridges have load limits. The load limit here is the set of contracts an exchange is willing to administer, not the set of markets a wallet can see.
What I Would Watch Over the Next Few Quarters
Predictions are cheap, so I will stick to observables. First, any change in how the company describes its regulatory status. The current description is consistent: unregulated, no claim of local authorization, open to cooperation. A drift toward softer language would be a tell, and not a reassuring one. Second, any update to the investor alert entry, or any separate enforcement language. The June listing was defined by what it is not. A later action would not have to respect that boundary.
Third, hiring. If the Singapore office keeps asking for people who can sit there, the base is operational. If the language shifts to fully remote or to a new city, the headquarters story needs a rewrite. Fourth, the American wrapper. A filed rule, a comment period, an approval, a denial, or a quiet delay all count as information. Silence past the rough ten-to-twelve-month sketch is also information. Fifth, whether other registered venues try a similar HIP-3 style reference. One proposal is an experiment. A second would start to look like a pattern.
I would also watch the product surface that ordinary users actually touch. Permissionless infrastructure can stay permissionless while the popular front end adds gates, geo blocks, or partner logins. Those gates are where jurisdiction sneaks back in, even if the contracts do not move. The alert already named both a foundation site and a trading application. Front ends are not decoration.
A Fair Reading, Without the Halo
Here is the version I would tell a friend who trades, not a version built to win a thread. Hyperliquid’s builders put a registered base in Singapore and said so. The local authority does not treat the platform as something it regulates, and it is not aware of a license in any major jurisdiction. An investor alert went up in June so that people would not invent that license in their heads. The company said the alert is not a ban and not a finding of wrongdoing, and nothing in the public description of the list contradicts that. A separate American plan would borrow the order book, fence the products, and clear them through firms that already live inside a regulatory system. That plan is not live.
None of this makes the market safe, and none of it makes the market a mirage. It makes the market a place where you can know who is not supervising you. That knowledge is uncomfortable. It is also clearer than the alternative, which is a city name doing quiet work it was never assigned.
If you want a single habit to take out of the story, make it this. When a venue tells you where the team sits, ask who is allowed to tell the team to stop. If the answer is “no one, yet, in the licensing sense,” trade that answer, not the skyline. Skylines are easy to photograph. Remits are not. The remit is the part that decides whether a bad week is a market event or also a legal one.
I keep the two stacked in my head on purpose. Stack one: a small team, a real office, fast shipping, an onchain book that other firms now want to reference. Stack two: no known major license, an alert designed to kill a false impression, a home-city rulebook that is chilly toward overseas-only token services, and a US path that still has to earn its stamps. You can respect the first stack. You still have to price the second. Anyone who offers you a blend in which the office cancels the alert is selling a smoother sentence than the facts will hold.
And if the regulated wrapper does arrive, read the customer agreement before you read the celebration. The name on the chart may match. The name on the account will not. That mismatch is not a flaw in the announcement. It is the announcement.
]]>
