Originally published at mrnasdog.com/research/hype by MrNasdog.
This is a MrNasdog Pressure Framework analysis of Hyperliquid (HYPE) on Metric 1 (sell pressure) and Metric 2 (buy pressure). Narrative (Metric 3) is covered separately. Every number is pulled from Hyperliquid's own API. The short version: the protocol buys back more of its own token than its insiders unlock.
The setup
HYPE has a fixed maximum supply of 1 billion tokens. Of that, ~298.7M is circulating, ~241.4M sits in a locked team/contributor vault, ~44.44M has been absorbed by the Assistance Fund, and ~414.7M is reserved for future emissions. Price is around $58. Three features make HYPE unusually clean to score: no protocol inflation, an on-chain Assistance Fund that buys HYPE with 99% of trading fees, and a discretionary ā but stable ā contributor unlock.
Metric 1 ā Sell pressure
Sell pressure measures the predictable selling baked into the design. Walking the six sources for HYPE:
1. Protocol inflation ā zero. HYPE's supply is fixed; nothing is minted on a schedule. This source contributes nothing, and that's verifiable from the supply figures.
2. Vesting unlocks ā Tag A, ~1.0M / 90 days. The locked vault holds 241.4M HYPE. On paper the contributor allocation authorizes up to ~9.92M/month, but Hyperliquid distributes at discretion and has consistently released far less ā roughly ~330K/month. Because that smaller figure repeats month after month, it's a stable pattern, so we treat it as trackable and predictable (Tag A) and project the recent rate forward: ~1.0M over 90 days. The authorized ceiling is not the number; the actual on-chain release is.
3. Treasury releases ā Tag B, ~60.2M. The Hyper Foundation treasury sits at the identified address 0xd57eā¦ and holds ~60.2M HYPE ā verified directly from the official API (~60.1M staked + 0.1M spot). We can read it, but its deployment timing is discretionary, so it's Tag B: trackable, unpredictable. It doesn't enter the bars, but it's a real, watchable overhang.
4. Locked stake unlocks ā skipped. HYPE's unstaking cooldown is 7 days ā under the 90-day threshold, so it's short-term noise the framework deliberately ignores.
5. Bankruptcy estate ā zero. HYPE launched in late 2024; there is no FTX-style estate distributing tokens.
6. Large concentrated holdings ā Tag C, ~238M. With the Foundation now separated out as Tag B, the remaining float (~238M, which is also what aggregators report as "circulating") is spread across ~237,000 wallets, most unidentified. We can see balances but not owners or intentions, so it's untrackable for sell-timing.
Metric 1, Tag A total: ~1.0M HYPE of predictable sell pressure over 90 days ā entirely the contributor vault release.
Metric 2 ā Buy pressure
Buy pressure measures the predictable buying baked into the design. HYPE's is unusually strong and unusually clean.
1. Revenue-backed buyback ā Tag A, ~2.95M / 90 days. Hyperliquid's Assistance Fund receives 99% of all trading fees in USDC and uses them to buy HYPE on the open market, sending the tokens to an address with no private key ā effectively a burn (ratified by a December 2025 validator vote). This is the cleanest Tag A in crypto: a public address, a fixed rule, fully on-chain. Reading its buy fills directly, the fund is buying ~32,723 HYPE per day ā about 2.95M over 90 days ā and it now holds ~44.44M HYPE.
2. Burn mechanisms ā Tag A, negligible. Explicit burns to dead addresses total only ~1,676 HYPE. The real "burn" for HYPE is the Assistance Fund itself.
3. Locked allocations ā context. ~432.0M HYPE is staked. Staking locks supply rather than buying it, so it doesn't enter the buy bar ā but it does mean a large share of supply is illiquid.
4. Protocol-level demand (gas) ā small. HYPE is the gas token; usage-driven demand is real but minor next to the buyback, and is largely captured in the burn figure.
Metric 2, Tag A total: ~2.95M HYPE of predictable buy pressure over 90 days ā almost entirely the Assistance Fund.
The read
On the predictable (Tag A) layer, HYPE takes in ~2.95M HYPE of buying against ~1.0M of unlock selling every 90 days ā a 2.95Ć buy-to-sell ratio, net +1.95M HYPE. The structural consequence is visible in the supply itself: circulating is roughly flat-to-shrinking even as the vault unlocks, because the Assistance Fund removes more than the vault releases. In framework terms, HYPE's structural conditions on the supply side are favorable ā the protocol is a bigger buyer of its own token than its insiders are sellers.
Data & confidence
Source: the Hyperliquid info API (api.hyperliquid.xyz) ā origin-first. High confidence: buyback rate (on-chain buy fills), supply, Assistance-Fund and vault balances (read directly). Medium: the ~330K/month vesting release (discretionary, but a stable pattern). Reconstructed: the 90-days-ago balances ā the chain stores only current state, so a daily snapshot is now recording exact history going forward.
Limitations
Discretionary distribution can change ā the foundation could release more or less than its recent pattern. The buyback scales with trading fees, so a volume collapse would shrink it; the projection assumes status quo. And the untrackable float (~298.7M, Tag C) plus any unidentified foundation treasury sit outside the headline number by design ā not because they don't matter, but because we won't put a number on what we can't verify.
MrNasdog Pressure Framework analysis of HYPE, Metrics 1 & 2. Data + explanation only. Not financial advice. Numbers as of May 2026. By MrNasdog (Zhiyi Song).
