I keep coming back to the same quiet problem in crypto trading. Everyone talks about speed. Far fewer people talk about whether the book you are looking at is even complete. On September 24, DoubleZero said it had put five Hyperliquid data feeds on Edge, sent over dedicated fiber, so subscribers can watch the live order book and related markets without stitching public updates together by hand. That sounds dry. It is not. For anyone who prices risk across dozens of contracts at once, a missed tick is not a trivia item. It is the difference between a fill and a shrug.
Why This Launch Matters More Than Another Press Note
Hyperliquid has become one of those venues people mention in the same breath as serious volume, not just as a novelty chain. DoubleZero’s move treats that volume as something that should travel on infrastructure built for market data, not as a leftover from a public websocket. In my experience, that shift is how a venue stops looking like a side project and starts looking like a desk can actually lean on it.
The package is not one firehose with a pretty label. Four feeds cover native perpetual futures and the trade markets that sit under the HIP-3 framework, including contracts tied to commodities. A fifth feed carries order intents pulled from mempool transactions. Together they are meant to give firms a sequenced view they can drop into pricing, hedging, and internal models.
DoubleZero framed the rollout as a paid subscription available now. Each plan includes two IP addresses per region. That detail is small and very telling. This is not a free ticker widget. It is a commercial data path.
What Edge Is Actually Selling
Edge is a read-only distribution layer. It does not enter orders. It does not execute. It publishes. A co-founder described that role earlier when talking about another venue on the same network: the platform delivers information once, then the fiber layer fans it out to connected subscribers. Hyperliquid is now the third source on that map, after validator-style data from another major chain and a prediction-market order book that opened in August.
I find that distinction useful because marketing language likes to blur data and trading. Here the line is clean. Firms still decide whether to trade. They still own the strategy. What they buy is a cleaner path to the book.
When traditional markets are closed, oil and other assets still trade, they just trade somewhere else.
– DoubleZero co-founder, launch remarks
That line is doing a lot of work. Perpetual contracts do not clock out at 5 p.m. in Chicago. Funding payments keep the price honest against a reference while the position can stay open. If you care about crude, gold, or silver outside the hours of a listed pit, you already know the book does not go to sleep. The question is whether your feed does.
The Five Feeds, Without The Brochure Gloss
Let me put the product in plain language. You are not buying a single “Hyperliquid channel.” You are buying a set of streams that cover different layers of the same venue.
- Native perpetual futures that live on Hyperliquid itself
- HIP-3 markets launched by outside teams, including commodity-linked contracts
- A dedicated stream of order intents visible in mempool activity
- Sequenced updates designed so firms do not rebuild the book from public snapshots
- Delivery over a private fiber path rather than a best-effort public socket
The mempool feed is the one I would watch most closely if I sat on a quant desk. Public trade prints tell you what already happened. Intents hint at what might happen next. They are not a crystal ball. They are a second camera angle. Sometimes that angle is noise. Sometimes it is the only reason a model does not walk into a wall.
Partners named in the launch include validator operators and ecosystem groups working with Hyperliquid infrastructure. The announcement listed three firms as collaborators without spelling out a separate technical job for each. That is common in these rollouts. The useful takeaway is simpler: the feed is not being assembled from a lonely public endpoint in isolation. It is being assembled with people who already sit close to the validator set.
Why Public APIs Stopped Being Good Enough
Anyone who has kept a live book knows the choreography. You open sockets. You request snapshots. You apply deltas. You reconcile when a sequence number jumps. You write a watchdog because the socket will die at the worst moment. Then the venue changes rate limits or depth, and your tidy little reconstructor starts lying to you in small ways that compound.
DoubleZero’s pitch is that those public endpoints have already become thinner. Update frequency dropped. Depth dropped. Firms that track many contracts at once were doing extra work just to stay current. Edge claims to replace that work with a sequenced commercial feed. The company calls it its fastest commercially available Hyperliquid path. Fair warning: the launch note did not publish a latency bake-off. I would want those numbers before I retired a homemade reconstructor. Still, the product thesis is easy to follow. Stop rebuilding. Start receiving.
Perhaps the most interesting aspect is not speed in the abstract. It is sequence integrity. Automated systems care about order. A delayed bid is bad. A bid that arrives after the trade that should have sat behind it is worse. You can paper over a few milliseconds. You cannot paper over a book that is internally inconsistent.
Commodity Perpetuals Change The Audience
Crypto pairs are the obvious customer. Commodity-linked perpetuals are the sneakier one. HIP-3 lets outside teams list markets on Hyperliquid. Some of those markets point at oil, gold, and silver. They do not expire like a classic dated future. They float, and funding is the tether.
That structure pulls in a different kind of user. Energy desks. Metals hedges. Funds that already think in basis and roll and suddenly notice that the weekend is no longer a dead zone. A recent industry research pack put one HIP-3 deployer’s second-quarter volume in the low hundreds of billions, with a sharp quarter-on-quarter jump, and estimated that the same deployer represented the vast majority of HIP-3 activity in the period. Those figures were research estimates, not audited filings. Treat them as direction, not gospel. Direction still matters. The tape is no longer only coins.
Hyperliquid as a whole was described as clearing more than six hundred billion dollars of volume in the same quarter. Even if you haircut the headline, the operational point stands. A venue that size produces a data problem. Someone will sell a solution to that problem. DoubleZero just raised its hand.
| Feed Layer | What A Desk Actually Gets | Why It Matters |
| Native perps | Core crypto perpetual book and prints | Baseline pricing and hedge maps |
| HIP-3 markets | Outside-listed contracts, including commodities | Cross-asset books that trade around the clock |
| Mempool intents | Pre-book order intention flow | Early signal, with plenty of noise |
| Sequenced Edge path | Ordered updates over dedicated fiber | Less reconstruction, fewer silent gaps |
How This Sits Next To Kalshi And Solana On The Same Rail
Context helps. Edge did not appear overnight for this one listing. Validator data came first. Then a regulated prediction venue opened Level 1 and Level 2 sports and crypto-linked contracts to subscribers. Level 1 is top of book and prints. Level 2 is depth. Hyperliquid now adds a third publisher to that same “publish once, deliver many” model.
I’ve found that multi-venue data platforms either become plumbing or they become noise. Plumbing wins when the format is boring in the best way: consistent, sequenced, documented, billed like a utility. Noise wins when every new venue arrives with a different schema and a different apology. DoubleZero is clearly aiming at plumbing. Whether the Hyperliquid schema stays stable over the next year will decide if desks treat this as core market data or as a pilot that never leaves the lab.
Regulation Is Nearby, But The Feed Is Not A License
There is a U.S. policy thread running under some of these commodity markets. Groups tied to the venue and to a HIP-3 deployer have asked the futures regulator to consider compliant perpetual structures linked to major energy benchmarks. The filing argued that continuous contracts could help firms manage weekend exposure. The agency has been chewing on design questions: reference prices, surveillance, margin, customer protections. Nothing in the DoubleZero launch equals an approval to offer those products to U.S. customers.
That needs to be said twice, because data vendors love the halo of policy news. A subscription to a book feed does not authorize you to trade a contract in a jurisdiction where you cannot trade it. It also does not settle the debate about how a perpetual should be supervised. It just shows the book to people who already have a reason to watch it.
The same announcement nodded at a political signal from the White House about a possible compliant path. It did not name an approval, a start date, or a change in who can access the platform today. I would keep that sentence in the “atmosphere” column, not the “product” column.
What Trading Firms Will Actually Do With This
Strip away the fiber poetry and you get a short list of jobs.
- Keep a live book across many tickers without a homemade reconstructor.
- Mark internal inventory when traditional pits are closed.
- Feed signals into market-making quotes and hedge ratios.
- Compare mempool intents against eventual book updates.
- Run research on HIP-3 commodity flow without polling a public socket all night.
None of that is glamorous. All of it is expensive when you do it badly. A delayed or dropped update can change whether a strategy gets filled. DoubleZero is explicit about that risk. The company supplies the path. The firm still owns the decision.
There is a cultural point here that rarely makes it into launch copy. Market data used to be a stock-exchange product with a binder of fees and a legal team. Crypto copied the public internet instead. That was fine when size was small. It gets messy when volume looks institutional and the public API starts rationing depth. Edge is an attempt to drag crypto market data back toward the older model: publish once, charge for access, deliver on private pipes, sound a bit like the incumbents in New York and Chicago.
The Uncomfortable Questions Desks Should Ask Before They Pay
I like the product shape. I still want a few answers that the announcement left on the table.
First, latency. Calling a feed the fastest commercial option is a claim, not a measurement. Firms will run their own clocks. They always do. If the gap versus a well-tuned public stack is tiny, the sale becomes about reliability and depth, not about microseconds.
Second, completeness. Does every HIP-3 market land in the four market-data feeds, or only a subset? Commodity names get the headlines. Long-tail listings are where reconstructors usually break.
Third, failover. Two IPs per region is a start. What happens when a region blinks? Market data platforms live and die on the boring Tuesday when a fiber cut and a software deploy arrive in the same hour.
Fourth, schema drift. Hyperliquid has changed public interfaces before. A commercial feed is only valuable if the vendor absorbs those changes instead of shipping them to every client at 2 a.m.
Fifth, the mempool feed’s false-positive rate. Intents cancel. Intents spoof. Intents sit. A raw stream without quality notes can train a model to chase ghosts. I would want documentation that is almost painfully plain about what an intent is and what it is not.
A Closer Look At HIP-3 And Why The Label Keeps Coming Up
HIP-3 is the permission structure that lets outside teams create perpetual markets on Hyperliquid rather than waiting for a single listing committee to invent every ticker. That is a governance choice with market-structure consequences. More listing freedom usually means more tape, more odd contracts, and more work for anyone who wants a unified book.
It also means the venue is no longer only “the Hyperliquid book.” It is a host. Some of the most interesting flow may sit on markets a third party designed. If your data stack only knows native perps, you are watching half the room. DoubleZero’s decision to include those markets in the same commercial bundle is, to my eye, the part that makes this launch more than a simple API reseller story.
Think about gold at 3 a.m. on a Sunday. A dated future on a traditional exchange is dark. A perpetual can still be alive, still paying or receiving funding, still showing a stack of bids. If that stack is only visible through a fragile public socket, the “always on” promise is mostly branding. If it arrives sequenced on fiber, the promise starts to look operational.
Who This Is For, And Who Can Ignore It
Retail traders staring at a mobile chart do not need this. They will never notice the difference between a rebuilt public book and a commercial feed. Market makers, statistical arbitrage books, and cross-asset hedges will notice. So will research teams that want historical integrity rather than a folder of websocket logs with holes in them.
There is also a middle group: funds that are not high-frequency but still run automated hedges overnight. They do not need the last microsecond. They need the book to be there when London is asleep and Houston still cares about crude. That group is easy to overlook in crypto coverage. It is also the group most likely to pay a clean invoice for data and then never talk about it in public.
If your strategy is discretionary and slow, save the money. If your strategy assumes the book is a reliable object, budget for the object.
The Fiber Story Without The Romance
Dedicated fiber sounds cinematic. In practice it is a routing choice. Public internet paths wander. They congest. They take different routes at 10 a.m. and 10 p.m. A private path is supposed to be dull. Dull is the point. DoubleZero says the Hyperliquid view is assembled with validator operators and partners, including presence in Tokyo. Geography matters for a venue whose validators are not sitting in a single matching-engine building in New Jersey.
I would not treat “fiber” as a synonym for “fastest possible.” I would treat it as a synonym for “less random.” Consistency beats a lucky fast hop that disappears on the next packet.
What a desk should measure, in order: 1. Gap rate versus an independent book build 2. Sequence breaks per hour 3. Depth available at each level 4. Time from venue event to local clock 5. Behavior during known venue upgrades
If a vendor cannot talk calmly about those five items, the brochure can wait.
Volume, Funding, And The Weekend Problem
Perpetual design is easy to describe and hard to live with. The contract does not expire, so the market needs a pressure valve. Funding is that valve. When the perpetual trades rich to the reference, one side pays the other. Over time the gap is supposed to shrink. That mechanism is why commodity names can keep trading when a dated future is closed. It is also why data quality on weekends is not a luxury. Weekend prints set Monday’s inventory marks for anyone who stayed in the position.
Traditional venues solved overnight data a generation ago, with official feeds and official clocks. Crypto improvised. Improvisation works until the notionals look like real markets. Then someone builds a paid pipe and starts comparing themselves to the old exchanges. That comparison will annoy purists. It will sound familiar to anyone who has bought a professional data seat in equities or listed futures.
Partnerships, Validators, And The Messy Middle
Launch announcements love partner lists. This one named ecosystem firms working around Hyperliquid and said validator operators helped shape the feed. A chief executive from one of those firms said the group was participating alongside two others. Roles were not split on the page. Fine. I care less about the seating chart than about whether the feed can survive a validator set change without going dark.
Decentralized matching and professional market data are awkward roommates. One culture wants many independent operators. The other culture wants a single authoritative sequence. Edge is an attempt to keep both stories standing. The operators stay in the picture. The subscribers get one ordered stream. If that compromise holds, more venues will copy it. If it slips, firms will go back to building their own reconstructors and complaining on group chats.
Pricing Reality And The Two-IP Detail
The foundation said prices live on the company site and that each subscription includes two IP addresses per region. That is how professional data has been sold for a long time: bind the feed to known endpoints, limit leakage, charge for extra seats. It is also a hint about the intended buyer. You do not write IP limits for a casual charting app.
Will the fee be worth it? Only against the cost of the team that currently babysits sockets. If that team is one engineer and a prayer, the invoice can look high. If that team is four people and a graveyard pager, the invoice can look cheap. I have watched firms underprice their own operational pain for years. Data vendors live on that habit.
What Comes After A Third Venue On Edge
Once you have three publishers on one distribution layer, the product stops being a one-off hookup. It becomes a category. The next questions write themselves. Which venue is fourth? How standardized are the schemas? Can a single subscriber process handle Hyperliquid depth, prediction-market depth, and chain-level validator events without three different parsers?
Standardization is where these platforms either mature or stall. Desks do not want a unique snowflake for every source. They want a small family of message types and a clock they can trust. If Edge can keep adding venues without turning into a junk drawer, it has a business. If every new feed needs a new integration project, it has a series of press notes.
I also expect competitors. Any time a venue’s public API gets stingier, a private feed appears. Some will be raw colocated pipes. Some will be reconstructed products with extra analytics bolted on. DoubleZero’s angle is the network itself: one publish, many subscribers, fiber in the middle. That is a distribution story more than a venue story. Distribution stories can travel.
A Practical Way To Think About Risk
Market data risk is easy to ignore because it does not show up as a red number in a profit-and-loss file. It shows up as a quote that should not have been there. It shows up as a hedge that was a second late. It shows up as a backtest that looked brilliant because the research file skipped the outage.
A sequenced commercial feed reduces one class of that risk. It does not reduce model risk. It does not reduce leverage risk. It does not make a bad commodity perpetual into a good one. Keep the humility. Buy better tape if you need better tape. Do not confuse the tape with the trade.
The product supplies the data path; firms still use their own systems to decide whether and how to trade.
That sentence from the launch is the right sentence. Frame it on the wall if you have to. Data vendors drift toward sounding like they are selling edge. They are selling input.
Putting The Day In Perspective
September 24 will not be remembered as the day crypto invented market data. It may be remembered as another brick in a slower shift: venues with real volume starting to treat the book as a product that can be packaged, sequenced, and sold the way older markets sold it. Hyperliquid supplies the tape. DoubleZero supplies the truck. Subscribers supply the strategy. That division of labor is ordinary in traditional finance. It is still slightly new in this corner of the market, which is why the announcement feels larger than a routine integration note.
Will every desk subscribe tomorrow? Of course not. Some will wait for latency charts. Some will wait for a fourth venue. Some will keep their reconstructors because they trust code they wrote at midnight more than a vendor they met in a blog post. That skepticism is healthy. It is also how you end up maintaining five sockets and a wiki page titled “do not touch.”
My own bias is simple. If you are already making markets, hedging commodities through the weekend, or comparing intent flow with eventual prints, a dedicated book feed is worth a serious trial. If you are still learning what a perpetual funding rate is, close this tab and go read something more basic. The market does not need more people paying for pipes they cannot use.
The feeds are live. The prices are posted. The commodity names will keep trading when the listed world is dark. The rest is operational discipline, which has never been the fun part of this industry and has always been the part that decides who is still standing after the next crowded weekend.
