Hyperliquid Dedicated Node Clusters and Unlimited Nodes: Pricing and When to Use Each

Hyperliquid Dedicated Node Clusters and Unlimited Nodes: Pricing and When to Use Each

The free Hyperliquid API has hard limits. Most info endpoints give you about one call a second. One IP address gets 1,000 WebSocket subscriptions. Hyperliquid lists 234 perp markets and 326 spot pairs, so that budget goes fast.

So you need paid access. That is where it gets hard. Ask most providers what a Hyperliquid node costs and you get a contact form.

Dwellir sells Hyperliquid L1 access three ways. Every price is on this page. No credits to convert. No compute units to work out.

Scope: this post covers Hyperliquid L1, also called HyperCore. That means the order book, fills, order statuses, and the info endpoint. HyperEVM JSON-RPC is a separate product. New to the two layers? Start with What Is Hyperliquid.

The Three Options, and What They Cost

Why There Is No Single Dedicated Node

The obvious middle step is to rent one Hyperliquid node of your own. Dwellir does not sell that. The reason shapes everything below.

One node is a ceiling. Hyperliquid's official machine specification for a non-validating node is 16 vCPUs, 128 GB of RAM, and 500 GB of SSD. That box answers a fixed number of requests per second. When your workload passes that number, there is nowhere to go. You cannot add capacity to one node. You can only replace it with a bigger one.

A pool behaves differently. Put the same workload in front of several nodes and the work spreads across them. The pool absorbs growth instead of stopping it at one machine's limit. It absorbs bursts the same way. A spike that would saturate one node spreads across several.

So both paid options put you on a pool. The $1,500 Unlimited Node puts you on Dwellir's pool. The $4,000 cluster gives you a pool of your own. Neither leaves you stuck with one machine's capacity.

Option 1: The API Platform

The API platform is the shared, metered product. It is the right starting point. Most Hyperliquid workloads never need more.

You get an endpoint onto Dwellir's shared Hyperliquid pool. Every response counts as one API credit. Trace and debug methods count as one credit too. Other providers charge a multiple for those.

Past your allowance, Dwellir bills extra responses per million: $5.00 on Developer, $3.00 on Growth, $2.00 on Scale. Above a billion requests a month, ask about pay-as-you-go.

Good fit for:

• Getting started, and prototypes
• Dashboards and portfolio trackers
• Analytics backends with predictable volume
• Anything where you would rather pay $49 than $1,500

The reason to leave is not performance. It is the counter. Once you watch the overage line every month end, a flat price beats a cheap one.

Option 2: Unlimited Node, $1,500 a Month

An Unlimited Node is the same shared pool with the meter switched off. Send as many requests as you like, at whatever rate you like. There is no credit counter, no monthly allowance, and no overage line on the invoice.

Two things make this the right buy for a serious workload.

The bill stops moving. Hyperliquid volume is not steady. A volatile week means more order book changes, more fills, and more polling from your own systems. On a metered plan, that is when the counter runs hot. At a flat $1,500, a quiet August costs the same as a week the market moves 20%.

You can scale inside the pool. This part is easy to miss. Your workload is not pinned to one machine. Growing it does not mean buying a bigger box. The pool holds more capacity than any single node in it, and your traffic spreads across that capacity. Doubling your request rate changes what you send, not what you buy.

The honest caveat: the pool is shared. Other customers send their requests to the same nodes. Most of the time you will not notice. When you do, responses come back a little slower during the busiest minutes of the day. Your requests are queueing behind other people's.

"Unmetered" means nobody is counting your requests. It does not mean the pool has infinite capacity. Those are different promises. Know which one you are buying. The same distinction applies to Dwellir's general Unlimited Nodes.

Good fit for:

• Mid-frequency and signal-driven trading bots
• Indexers and data pipelines running continuously
• Teams whose metered bill has become unpredictable
• Any workload that is growing and does not want to repurchase capacity as it grows

Option 3: Dedicated Node Cluster, $4,000 a Month

A Dedicated Node Cluster is a pool of Hyperliquid nodes run for you alone. Two or more nodes, nobody else's traffic on them.

This is the tier for a large workload. The argument is not "faster" in the abstract. Dwellir can tune a cluster around one known workload. No shared pool can do that.

The nodes match what you actually do. A shared pool has one configuration, and it has to work for every customer at once. On your own cluster, Dwellir knows your access pattern. Node flags, the feeds written to disk, regions, and load balancing all follow your workload rather than an average.

Nobody else's work lands on your nodes. On a shared pool your neighbours run heavy workloads too. Some look nothing like yours. A neighbour doing something completely different can still take capacity you wanted. Isolation removes that. Your performance then depends on your own traffic and nothing else.

Two nodes means one can go down. Node software needs updates. Hosts need maintenance. A node that goes offline must resync with the network before it is useful again. With more than one node, all of that happens behind a live feed instead of in front of it.

The cluster price also bundles the data services Dwellir sells separately, and removes the usage charges on them:

• info endpoint access, included
• gRPC Streaming API, normally $299/month plus usage, included and unmetered. All seven raw node feeds, plus 14 typed market streams for L2, L4, BBO, trades, candles, and mids
• Order Book Server, normally $199/month plus usage, included and unmetered
• Extended SLA terms available
• Regions of your choice

Worth doing the arithmetic. The step up from the Unlimited Node is $2,500. The gRPC stream and the Order Book Server cost $498 of that at their standalone prices. The bundle also removes their usage charges. The rest buys isolation, tuning, and redundancy.

Good fit for:

• Market makers quoting continuously
• High-frequency strategies and latency-sensitive execution
• Trading firms where performance consistency is the product
• Exchanges, custodians, and anyone with an uptime commitment to their own users
• Teams feeding many downstream consumers from one Hyperliquid feed

Which One Should You Pick?

Three questions get you most of the way there.

Is the metered bill still predictable? If you can forecast next month's invoice and it is under $1,500, stay on the API platform. When you cannot forecast it, or the forecast is above $1,500, the Unlimited Node is cheaper and calmer.

Does shared capacity cost you anything real? If a slow response means a chart redraws late, shared is fine. If it means a worse fill, or a missed one, you already pay for shared capacity. That cost just never reaches the infrastructure invoice. That is the case for a cluster.

Do you need the nodes configured around you? Some workloads have a specific shape, lean on particular feeds, or need a named region. That is a cluster conversation. A shared pool is built for everyone, which means it is built for nobody in particular.

What Other Providers Charge

Here is the honest state of the market in September 2026. The question is narrower than it first looks. Who will sell you dedicated or unmetered access to HyperCore? And will they tell you the price?

Alchemy serves HyperEVM only, not HyperCore, and has no standard dedicated tier. RouteMesh and Uniblock are routing layers rather than node operators. Both pass HyperEVM traffic only. If your workload touches the order book, none of the three is an option.

Read that table carefully, because the last column is the interesting one. Several providers will serve you HyperCore data. Only one publishes what dedicated or unmetered access costs.

QuickNode is the closest competitor on coverage. They expose all seven HyperCore node data streams, and so does Dwellir. Those seven are blocks and signed actions, fills, order statuses, raw book diffs, miscellaneous events, TWAP statuses, and periodic state dumps. Dwellir adds 14 typed market streams for L2, L4, BBO, trades, candles, and mids. On coverage the two are genuinely comparable.

Where they differ is billing. QuickNode meters HyperCore in credits and publishes the rates. Every info method costs 20 credits per request. Streaming costs 10 credits per 0.1 MB of data after filtering. Plans run from $49 to $999 a month. That is a real alternative, priced clearly, which deserves saying. QuickNode does not publish the dedicated tier.

Metering streams by the megabyte is the nuance worth understanding. Busy markets produce more order book changes, which is more data, which is more credits. The bill tracks volatility. You learn what a volatile month cost after it happened. That is the trade the flat $1,500 removes.

HypeRPC calls itself Hyperliquid's premier RPC and node infrastructure. They sell both shared and dedicated access. Their pricing is hard to establish. The public rate card moved. The institutional tier promises "unlimited data throughput" with no number attached and a Contact Sales button. Hyperdash acquired HypeRPC in 2026. Treat any figure you find as stale and ask them directly for current terms in writing.

Chainstack is the provider most likely to confuse a comparison, so be precise here. They do serve HyperCore on their shared plans. They also publish a dedicated node price built from an open formula. That formula is a $199/month Pro plan, plus compute from $0.50 an hour, plus storage at $0.01 per 20 GB per hour. Those dedicated figures cover HyperEVM nodes. Chainstack does not sell a dedicated HyperCore node or cluster. So those numbers are not an alternative to anything in this post.

Chainstack did write the sharpest summary of this problem.

Most providers in this category sell some version of "dedicated node." Very few tell you what it costs before you talk to anyone.

That is the gap this post is trying to close.

What About Running Your Own Node?

It is a real option, and for some teams it is the right one.

The spec is the easy part. You need 16 vCPUs, 128 GB of RAM, 500 GB of SSD, and Ubuntu 24.04. Ports 4001 and 4002 must be reachable from the internet. Tokyo is the recommended location, because that is where most of the network sits.

The disk is the part that surprises people. With default settings a Hyperliquid node writes roughly 100 GB of logs a day. That fills the 500 GB in under a week. The official repository tells you to archive or remove the older files yourself. Turn on the flags for fills, order statuses, and raw book changes and the number grows again. Storage on a Hyperliquid node is a job, not a purchase.

Then there is the ceiling problem from earlier. One self-hosted node has the same hard limit as one rented node. A workload that grows means building the pool and the load balancing yourself.

The real commitment is a substantial server, heavy daily log output, Tokyo placement, and 24/7 operations. That includes keeping stable peers. If you already run all that, peering is usually the piece that hurts. Both Dwellir and QuickNode sell a peering slot at $500 a month. If you do not, buying access is cheaper than it looks. The server bill is the only part of self-hosting that reaches an invoice, and it is the smallest part of the real cost.

Questions to Ask Before You Buy

Ask any provider, including Dwellir, the same five things. Get the answers in writing.

1. Is this one node or a pool of nodes? If it is one node, what happens when my workload outgrows it?
2. How many nodes, and what happens to my traffic while one of them is being updated?
3. Is "unlimited" a flat price, or a flat price up to a rate limit? If there is a limit, what is the number?
4. Which data services are included at that price, and which are separate line items with usage charges on top?
5. What SLA applies, and is it the same SLA as your headline plans?

Almost nobody asks question one. It decides whether you are buying capacity or buying a ceiling. Question three catches the most people. What Is an Unlimited RPC Node covers that category in more detail, including the three different products sold under the word "unlimited."

The Bottom Line

Start on the API platform. At $49 to $999 a month with one credit per response, it covers more Hyperliquid workloads than teams expect. Do not buy a flat rate before the metered bill tells you to.

Move to the $1,500 Unlimited Node when the counter becomes the problem. Two things change. The invoice stops tracking market volatility. Your workload stops depending on a request budget. You are on a pool, so growth is something you send rather than something you buy.

Move to the $4,000 Dedicated Node Cluster when shared capacity costs you more than $2,500 a month in outcomes you never see on an invoice. On that tier Dwellir tunes the nodes for your access pattern. No other customer can take capacity you wanted. A second node means updates happen behind a live feed. For a market maker or a trading firm running continuously, that consistency is the product.

Before you buy anything, measure two numbers. The first is your peak request rate. The second is what a slow response costs you. The first tells you whether shared capacity is enough. The second tells you whether to care.

Every price in this post is public. Dwellir's Hyperliquid pricing page lists the full line-up. The Hyperliquid docs cover the endpoints and data feeds. If you landed on the API platform or an Unlimited Node, create an account and point a workload at it today. If you need a cluster sized around your regions and SLA terms, contact the Dwellir team.