Hyperliquid Weekly — Fees Flat at $17.89M, HYPE Down 6.3%

2026-10-12 · 2026-w42

The BuyHype Bayesian model shows WAIT at 58% up-probability now versus WAIT at 57% a week ago, a one-point shift that leaves the bias essentially unchanged despite HYPE falling 6.3% to $87.26. The biggest reason it held: protocol fees were nearly flat, at $17.89M last 7d versus $17.93M prior 7d, a decline of only 0.2%. That revenue stability contrasts with softer chain metrics and keeps the model from turning more cautious, though it is not enough to move to a buy signal.

Fees remain the core support. The $17.89M in protocol fees over the last 7d came alongside $4.51B in core perps 24h volume and $12.53B in core perps open interest. Those are current activity levels, not weekly sums, but they show Hyperliquid's perps venue continues to transact at scale. News this week included a headline on crypto trading growth and Hyperliquid gains, while another focused on how Hyperliquid Strategies ties its story to one token. The model's 58% up-probability reflects that revenue durability, not a directional call.

Chain-level liquidity weakened. Chain DeFi TVL fell 5.4% WoW to $1.11B, and stablecoins on Hyperliquid declined 2.6% WoW to $7.23B. Both moves are modest but point in the same direction: less capital parked on-chain even as perps volume remained large. A headline about an address holding a $49.3M long position in Ether at $2,562.30 per ETH is a reminder that leveraged exposure can cut both ways. If TVL and stablecoin balances keep slipping, the fee base could eventually feel it.

Supply dynamics are mixed. The Assistance Fund holds 47.97M HYPE and net bought back +0.17M HYPE over 7d, a small but positive absorption. Against that, the unstaking queue shows 7.40M HYPE completing within 7 days, a much larger near-term potential supply. The buyback does not fully offset that queue on its face. This is a key reason the verdict stays WAIT rather than shifting higher despite stable fees.

HIP-3 builder markets are now a visible second engine. There are 10 DEXs, with combined 24h volume of $997.76M; top builder XYZ alone did $944.59M. The top markets were xyz:SKHX at $114.95M, xyz:BRENTOIL at $113.08M, xyz:XYZ100 at $110.38M, xyz:SP500 at $92.44M, and xyz:CL at $87.31M. That concentration in one builder is a strength in volume terms and a dependency risk if activity rotates away.

The most consequential news was operational. Multiple headlines reported that a trader netted $1.28M on Hyperliquid-based Papertrade through alleged price manipulation with 1,000x leverage, and Papertrade tightened limits afterward. The repeated coverage matters because it raises questions about market integrity and risk controls on venues built around Hyperliquid prices. Other headlines placed Hyperliquid in crypto top 10 discussions while also noting falls. Neither changes the core fee picture, but they shape sentiment.

Next week, watch whether the unstaking queue of 7.40M HYPE completing within 7 days is absorbed without a further drop in TVL or stablecoins, and whether core perps volume holds near $4.51B. This is data commentary, not financial advice.

Data recap & AI read

HYPE closed the week down 9.1 percent, sliding from an open of $86.34 to $78.50 after printing a high of $87.68 and a low of $76.17, and the single clearest driver in the data is the divergence between improving protocol revenue and a market that de-risked anyway, with TVL easing 2.5 percent to $6.94B over the same stretch.

The revenue line is the strongest counterweight to the price action: protocol revenue came in at roughly $16.07M this week against $14.18M the prior week, a 13.3 percent increase. That is real usage expanding while the token fell, which is the classic setup where price lags fundamentals rather than leading them, though one week of revenue does not confirm a trend.

On supply, the Assistance Fund grew from 47.96M to 48.12M HYPE, a net addition of 0.16M HYPE of buyback absorption. Set that against an average unstaking queue of 3.28M HYPE and the imbalance is obvious: potential sell-side supply from unstaking is roughly twenty times the weekly fund accumulation, so the buyback is a cushion, not an offset.

Positioning looks stretched rather than washed out. Average open interest held at $11.52B with average funding at 11.0 percent APR, meaning longs were still paying to hold exposure into a down week, and perp volume of $953.27B shows heavy turnover without a capitulation signal. Crowded, paid-for longs are the bearish read here; the bullish read is that leverage cleared without a deeper break below $76.17.

The AI verdict moved from 48 percent to 56 percent up-probability even as price fell, which is a mild vote of confidence in the fundamentals-over-price interpretation, and it is the kind of shift worth noting rather than trading on.

Net read for a buyer: the revenue trend and rising model confidence argue for accumulating on weakness, while the unstaking overhang and still-elevated funding argue for patience and smaller sizing until supply flows and positioning normalize. This is data commentary, not financial advice.

See the live AI buy / wait / avoid signal →

Past issues