Over the last two months, HYPE fell from the $45–50 range to ~$20.
That move wasn’t random, and it wasn’t just “market conditions”.
It was the result of three very specific sources of sell pressure, all of which were visible on-chain and are now either resolved or close to exhaustion.
This article breaks down what actually happened and why the setup going forward is structurally different from what it was two months ago.
1. Team Unlocks: The Supply Everyone Mispriced
One of the biggest sources of confusion around HYPE has been team unlocks.
Most trackers still show ~9.9M HYPE unlocking per month, which led many participants to assume a constant ~$200M monthly sell overhang. That assumption turned out to be wrong.
>Unlocked does not mean distributed.
>Distributed does not mean sold.
>And sold does not mean sold on the open market.
Looking at the first two unlock months (data tracked publicly by @qwantifyio):
Month 1
- Unlocked: 9.92M HYPE
- Actually distributed: 2.6M
- Staked back: ~1.2M
- Sitting in spot: ~300k
- Sold OTC: ~819k
- Sold to market: ~290k
Month 2
- Unlocked: 9.92M HYPE
- Actually distributed: ~1.125M
- Staked back: ~375k
- Sitting in spot: ~50k
- Sold OTC: ~700k
- Sold to market: 0
In both cases, roughly 7-10% of the headline unlock number actually resulted in direct or indirect sell pressure.
If this pattern holds, team unlocks are not a sustained shock. They are a gradually declining trickle, not a cliff.
The only real caveat is execution: if part of this supply is sold OTC to entities like PURR DAT (via Flowdesk), it reduces how much they need to buy on the open market. That affects where demand shows up, not whether demand exists. We will get back to this point later.
Net effect: misunderstood, now largely priced in.
2. Leverage Reset: Longs Had to Die
HYPE entered Q4 with a structurally unhealthy derivatives setup.
On Hyperliquid itself, long positioning dominated. That discouraged new buyers and incentivized others to front-run liquidations. And they were right.
What followed was not subtle:
- Millions of dollars in long liquidations across venues
- Additional forced selling from money-market positions, where users lent HYPE, borrowed USDC and bought more HYPE
These money-market liquidations do not show cleanly on public heatmaps, but their effect was real. Even without mass liquidations, many users had to sell HYPE preemptively to repay loans or add collateral.
Today, the leverage regime is quite different:
- Although there is still more than $150M in longs to liquidate at $15, the majority of aggressive longs are gone
- Liquidation profiles are far more balanced across Binance, OKX and Hyperliquid
- Reflexive downside pressure from leverage has largely burned itself out
This is not bullish by itself.
It is necessary for anything bullish to happen later.
3. The Tornado Cluster and the Anonymous CEX buyer
This was the single most damaging factor for price action, not because of their sales, but because many participants front-ran them and either sold, shorted, or simply refused to bid.
A cluster of 16 addresses, originally funded via Tornado Cash, accumulated roughly 4.4M HYPE at an average price of ~$8.8.
This cluster has been publicly tracked and documented by accounts like @mlmabc (https://x.com/mlmabc/status/1979669770237628678) and dashboards such as
@qwantifyio (https://www.qwantify.io/app/entities/05e5413a-fd86-4050-bb16-8572658d8518).
Starting in early January, this entity executed a highly mechanical liquidation strategy:
- Roughly one wallet unstaked per day
- Immediate TWAP selling on unstake
- No visible attempt to optimize execution
In total, this represented >$80M of supply that, under normal conditions, should have pushed HYPE well below $10.
It didn’t.
And that’s where the story changes.
As the Tornado Cash-funded cluster began selling aggressively on HyperCore, a very clear pattern emerged.
Almost immediately, Wintermute started arbitraging the flow:
- The Tornado cluster sold on HyperCore
- Wintermute bought on HyperCore
- Wintermute sent HYPE to Bybit (0xe401A6A38024d8f5aB88f1B08cad476cCaCA45E8)
- Wintermute sold to an anonymous buyer on Bybit
This type of flow is not new. Wintermute has been running similar arbitrage loops across multiple assets for over a year on Hyperliquid.
What changed was the intensity.
Once the Tornado cluster started liquidating its positions, this flow accelerated sharply.
Crucially, Wintermute was not acting as a directional buyer.
Its role was to transfer inventory from on-chain sellers to a large and persistent off-chain buyer, absorbing supply that would otherwise have overwhelmed HyperCore liquidity.
Over the last 30 days alone, Wintermute arbitraged over than $70M worth of HYPE, exceeding even the Assistance Fund’s net buying during the same period.
30D Hype buyers. Data tracked via @0xbloquegenesis bot powered by @infinitefieldx data
If you’re curious about who was buying HYPE on HyperCore over the last 30 days in the chart above, this is what I found:
- Top 3: @ResolvLabs, running a delta-neutral position
- Top 4: @Auros_global, acting as a market maker with no directional position
- Top 5, 6 and 7: anonymous buyers with directional exposure (at least on HyperCore)
- Top 8: likely @SilkBtc, with a delta-neutral position
- Top 9: likely @ManifoldTrading (or its founder Jae Chung), market-making activity but seems likely directional
- Top 10: likely @Fern82L, an early Hyperliquid supporter, appearing to be directional
On top of this, @mlmabc identified a separate cluster that likely belongs to PURR DAT:
Since his message, a large number of new addresses have appeared following the same pattern (staking 700-900 HYPE), suggesting that PURR DAT keeps accumulating.
Execution for this entity was handled by Flowdesk, and the HYPE originated from Bybit, making it very likely that the anonymous CEX buyer absorbing Wintermute’s flow was PURR.
And this was not the only source of sell pressure.
Alongside the Tornado cluster, Continue Capital emerged as another significant seller, offloading approximately 1.3M HYPE (~$28M) over roughly two weeks (they still hold almost 800k HYPE in stake, but so far they're still staked).
These sales followed a similarly steady execution pattern, with no visible attempt to optimize price.
On top of this, our dear Trove team also sold their 500k HYPE they bought a couple of months ago.
Despite these combined sales, the price held.
This situation was not ideal, but it is important to frame it correctly.
Having this cluster selling at $20–25 is materially better than having them sell at $50+, because in that scenario PURR DAT would have had to spend significantly more capital to absorb the same amount of supply.
At this point, this selling pressure is done (or almost).
There are probably no more large sellers active in the market, at least for now.
P.S.: At the time of publication, the cluster still holds around 300,000 HYPE. By the time you read this, that balance may be lower or fully depleted.
4. So What Changes Now?
With the main sources of forced supply resolved, the question shifts from “who is selling?” to “what remains to absorb?”
A. PURR DAT Remaining Firepower
Based on public estimates, PURR DAT still has significant capital available even after absorbing the Tornado cluster and Continue Capital’s selling.
MLM calculated that they spent $67.6M on the cluster he found. In recent days, new addresses following the same pattern have appeared:
- 0x5f46cba327079feb5a46799ab329f36974a89f5e: 117k HYPE
- 0x69e914280fd1356b2abaef78439d112976bab985: 124k HYPE
- 0x05bc24d10cbbae0677f284e6f280ebd27a0ed761: 103k HYPE
- 0x7eacdc425e99eaf52ce75ce9b5fc03dce7e2f901: 85K HYPE
- 0x0ad0878cc9d8ce470886bc4a47c2ba5186980eec: 100K HYPE
Assuming an average price of $21, these new HYPE holdings account for $11M, bringing the total to $78.6M.
To calculate their remaining firepower:
- Let's add a buffer and assume they spent $90M buying HYPE
- Let's also assume they will spend $30M buying back their shares this year
This leaves roughly $170M in cash available to buy HYPE.
This does not imply aggressive spot buying or a guaranteed upside. It implies that residual sell pressure (such as the remaining team unlocks) can be absorbed rather than amplified.
B. Perps Market Share
Despite absolute volumes being below ATH, Hyperliquid’s perps market share versus CEXs is trending up again.
Open interest has already surpassed previous relative ATHs against venues like Bybit.
Market share leads revenue. Revenue leads Assistance Fund flows.
C. HIP-3 Volumes
HIP-3 volumes have accelerated meaningfully following the launch of
@markets_xyz, bringing new exotic markets such as Oil and US Bonds. Recent weekdays exceeded $1B in daily HIP-3 volume (data from @asxn_r):
TradeXYZ remains the leading contributor, generating >$10M in annualized revenue.
D. Assistance Fund Behavior
The Assistance Fund is currently underwater, with an average cost around $23.6.
Historically, this has aligned with a local bottom.
Recent data shows:
- Consistent buying above 60k HYPE/day
- Several days exceeding 100k HYPE/day
It is important to note that this does not imply an uptrend. Rather, it means the Assistance Fund is burning millions of HYPE per day, and that burn rate increases as the price declines.
E. Portfolio Margin
Portfolio margin will unlock the ability to use a broader set of assets as collateral simultaneously.
This materially improves capital efficiency and makes delta-neutral strategies far more efficient, allowing sophisticated traders and market makers to deploy more size with the same capital base.
The likely result is structurally higher open interest, leading to higher volumes, higher revenues, and higher HYPE buybacks.
5. Closing Thought
Two months ago, HYPE was pricing:
- Unknown unlock behavior
- Excess leverage
- Large sellers
Today:
- Unlock reality is understood
- Leverage has reset
- The Tornado cluster and other big entities are done (or almost)
- Large sales are likely to be absorbed for a while
That doesn’t mean price must go up tomorrow.
It means the structural reasons it went down are no longer present.
The HYPE/BTC trend appears to be turning.
Hyperliquid
