Цена в 10 000 долларов: комиссия, кладбище хранилищ и новые технологии, стоящие за всем этим

On April 13, 2026, Hyperliquid raised the fee to create a vault from $100 to $10,000 — a 100x jump, buried in one line of a weekly update. The price tag is the least interesting part. That fee, in Hyperliquid's own words, exists "to gradually deprecate legacy HyperCore vaults." In plain terms: the only kind of vault you can still create today is the kind they're phasing out. Here's the whole picture — and why I'm not paying the toll yet.

The graveyard is a museum of the cheap era

I've written that two of every three Hyperliquid vaults are dead, and that the median user vault holds exactly $100 — the minimum and nothing more. Both true; both shaped by history. For most of the platform's life, opening a vault was free or nearly so — $100 from January 2025, nothing before that. People opened them the way you'd start a free blog: on a whim, by the thousand, then abandoned. The graveyard isn't proof that vaults are doomed. It's proof that cheap things attract litter.

The fee cleared the room

Watch creation collapse the moment the toll appeared: 220 new vaults in March 2026, 83 in April, 10 in May, 1 so far in June — down roughly 95%. A $10,000 fee at the door doesn't stop serious operators; it stops the litter. The Hyperliquid Discord filled with the sound of it — "this used to be free, then $100, now I can't afford to create one." As a filter, it did exactly what filters do.

It's not a price — it's a deterrent

Most coverage treats the $10,000 as a fee. It isn't, really. Here's the team's own announcement, posted April 13:

"The vault creation gas fee has also been increased from 100 USDC to 10k USDC. This is meant to gradually deprecate legacy HyperCore vaults, which have been feature frozen for around 2 years. Vaults can be tokenized in a more configurable way as smart contracts on HyperEVM using precompiles and CoreWriter."

Read that again: the fee exists to push people away from the vaults you can currently create. The original design runs on HyperCore, Hyperliquid's core trading engine, and it's been frozen — no new features — for two years. The $10,000 is a polite eviction notice.

What's replacing them is a different machine

The future is HyperEVM, Hyperliquid's programmable smart-contract layer. Instead of a fixed template, a new vault is a tokenized smart contract with customizable accounting, and it can reach everything the old one can't: spot markets, the permissionless HIP-3 markets, delegation to trading agents, and composability with the rest of on-chain finance. In plain terms, strategies that until now only lived inside centralized-exchange APIs — delta-hedged liquidity provision, on-chain market making, automated basis and funding trades, structured products that rebalance against the live order book — become buildable by anyone, on-chain, with public receipts. For a blog whose whole premise is receipts, this is the most exciting thing happening on Hyperliquid right now.

What the app won't tell you

Open "Create Your Own Vault" in the app today and you get a clean form: a name, a description, a deposit, and the fee flagged in red — "not refunded even if you close the vault." What the screen does not say, anywhere, is that you're buying into the generation Hyperliquid has publicly marked for retirement. The toll is on the screen; the obsolescence is not. That gap — between what an interface shows and what it leaves out — is exactly the kind of thing this blog exists to point at.

Did the fee at least buy better-funded vaults?

A little. In the handful of vaults born since the toll, the median deposit is about $18,000 — up from the old $100 — and two have crossed six figures. But look closer and you find the quiet punchline: several operators paid the full $10,000 and their vaults still sit at the $100 minimum. Ten thousand dollars spent to open a container holding one hundred. They're living proof of the rule this blog keeps repeating: the fee buys a container, not capital. It changed who opens a vault. It did nothing to repeal the law that money follows reputation and distribution — not toll payments.

And the math still doesn't bend

Even the healthier new median makes the point. A vault with $18,000 of outside money, run brilliantly at a 30% annual return, earns its leader about $540 a year in performance fees — payback on the $10k measured in decades. To justify the fee within a year you still need hundreds of thousands in deposits, and the fee hands you none of them. Run your own numbers:

One honest caveat on the calculator: it's the optimistic case. It assumes steady positive returns — no drawdowns — and ignores the high-water mark, the rule that pays the leader no fee until depositors are back above their previous peak. Every losing month pushes payback further out than the blue line shows.

So where does this leave someone thinking of running a vault — including me?

With a sharper plan, not a discouraged one. The only turnkey way to create a vault today is still the legacy one — there's no HyperEVM button in the app; that path means deploying a smart contract, a builder's job. So paying $10,000 right now is the worst of both worlds: full price for frozen technology. The honest move isn't "use the new rails instead" — it's don't pay for either until you can fill it. Build the audience first — this blog, these reviews. Build a public, on-chain track record now, for free, trading my own capital. And let the HyperEVM vaults mature into a real, accessible option in the meantime. When there's an audience and months of public proof, the $10,000 — whatever it's attached to by then — becomes a rounding error. The fee didn't make a vault a worse idea. It made timing it correctly — and building it on the right technology — far more valuable.

Sources: the fee change and the legacy wind-down are from Hyperliquid's official announcement of April 13, 2026; the new-vault capabilities are from Hyperliquid's documentation. All vault figures come from Hyperliquid's public API, snapshotted twice a day and recomputable by anyone. Nothing here is financial advice — it's one trader showing his work before spending his own money.