Jeff Sun

Crypto value investors ，dedicate to finding Compounds.

[Jeff Sun @sunyefei]

What's been happening in crypto over the past month
(Aug 27 – Sep 25, 2026 — vs August)

One-line take: risk assets still paid, but the driver flipped from “sentiment thaw” to “digesting highs + rate re-pricing.”

Charts (attached, in order):
1) Price: Aug vs last ~30d (BTC / HYPE / JUP / SIGN)
2) HYPE fundamentals
3) HYPE HIP-3 mix
4) JUP fundamentals

——

1) Bitcoin & the tape

August: BTC roughly $63k → $79k (~+25%). Mood flipped fear → greed.
Last ~30 days: ~$80k → ~$85k (~+5%), trough ~$76k (Sep 16), peak ~$86k (Sep 22). Higher price, flatter slope.

Same window: short- and long-end Treasuries and effective fed funds rose by tens of bps from the start of the period. The dollar stopped falling and bounced. Money got more expensive.

Take: the uptrend isn’t invalidated — just don’t extrapolate August’s slope. Watch digestion around key moving averages, and whether rates keep grinding higher.

——

2) Why only three names besides BTC

BTC is the base layer for liquidity and risk appetite. Outside that, the list stays short on purpose — only names where you can explain how they make money and keep reconciling the books:

• Hyperliquid (HYPE) — on-chain trading venue: volume, fees, buybacks
• Jupiter (JUP) — Solana trading front door + multi-product stack: traffic & core revenue
• Sign (SIGN) — identity / protocol infra: earlier stage, thinner public data (satellite, not the main course)

Narratives are plentiful. Checkable business logic is not.

——
3) Hyperliquid (HYPE)

Price (CoinGecko): Aug ~+62% ($52→$84). Last ~30d ~+11% ($85→$94). Big move was August; this month looks like digestion at highs — still ahead of BTC’s ~+5%.

Fundamentals (last ~30d sums vs prior ~30d; TVL = point-in-time Aug 26 → Sep 25):
• TVL $6.84B → $7.64B (+12%)
• DEX volume $3.32B → $5.01B (+51%)
• Fees ~$63.3M → ~$72.4M (+14%)
• Revenue ~$47.4M → ~$56.4M (+19%)
• Assistance-fund buybacks ~$44.45M → ~$44.51M (~flat). Still running; no step-up.

Structure matters more than headlines: HIP-3–related 24h notional ~$0.31B → ~$2.43B; share of the venue ~4.4% → ~24.3%. Venue 24h notional ~$7.1B → ~$10B. Stablecoin “bridge” size roughly +55%.

Newsflow was noisy this month (access/compliance, prediction markets, position color). Weight that below the table. Meme-pad / energy-contract chatter = lead, not conclusion.

Call: price digesting, but 30d fees/volume/TVL still thicker than the prior window; buybacks stable; growth isn’t only old perps — a new line’s share stepped up. Revisit if we later see “price still firm while 30d fee & volume sums turn negative.”

——

4) Jupiter (JUP)

Price: Aug ~+17% ($0.193→$0.225). Last ~30d ~+32% ($0.242→$0.320) — stronger than BTC here, and stronger than JUP’s own August.

Fundamentals (same lens):
• TVL $2.30B → $2.62B (+14%)
• DEX volume $0.89B → $1.44B (+62%)
• Fees ~$15.3M → ~$17.0M (+11%)
• Revenue ~$4.9M → ~$6.4M (+32%)

Weekly cuts: lending / perps / jupSOL and other core lines lifted TVL; prediction markets stay tiny and soft — core hot, experiments cold. No new buyback that clearly drove the tape. Deep news almost empty.

Call: “Solana trading heat getting priced.” Usage and revenue expanded; revenue grew faster than fees — not a soft mix. Without fresh catalysts, don’t read this as a tokenomics shock. If JUP sells off hard ahead of BTC, check Solana traffic and perp activity first.

——
5) Sign (SIGN)

Price: Aug ~+29% (~$0.0068→$0.0087). Last ~30d ~+67% (~$0.0081→$0.0135). High beta both windows.

Fundamentals: basically price + market cap only. No TVL, fees, revenue, users, or a stable progress digest.

Call: shows risk appetite spilling into small-cap narratives — not a proven business inflection. Satellite observation until public, checkable metrics appear.

—

6) Bottom line

1. August = thaw + BTC/leaders leading. Last 30 days = high-level gear-shift + rates lifting. Slope and leadership both changed.
2. HYPE: digestion in price; 30d TVL/volume/fees/revenue still above prior; buybacks ~$44.5M flat; HIP-3 share jumped. JUP: price more aligned with volume & revenue expansion. SIGN: upside both months, no reconcile-able fundamentals.
3. Macro still answers to BTC and rates. The three satellites reconcile: fees–buybacks–new product; Solana traffic & core revenue; whether SIGN ever fills in public metrics.

Sources: CoinGecko prices; TVL/fees/volume from public DefiLlama + Hyperliquid-related series (rounded). Comparison = last-30d sums vs prior ~30d (or dated point-in-time). Public research summary — not investment advice.

[Jeff Sun @sunyefei]

@benjamincowen 以下是MVR系统间接和近期趋势预测结果

[Jeff Sun @sunyefei]

为了不再凭“感觉”炒币，我做了一个 BTC 趋势量化MVR系统（Minimum Viable Research）。目的很简单：不求精准预测顶/底部，只求在仓位调整时有个客观依据。用 2024 至今的数据回测，对趋势拐点的判断准确率达到 81%。 目前系统对 BTC 的最新趋势判断是：拒绝给出“买入”建议，但也尚未确认“全面熊市”。

[Jeff Sun @sunyefei]

@benjamincowen 我会持续完善这个系统，向优秀的量化前辈们学习，包括Arthur Hayes、Glassnode等。 目前的 81% 准确率基于 2024 年至今的震荡市数据，未来在极端行情下表现如何还需要时间验证。 如果你也想摆脱“凭感觉”交易，欢迎关注。我会每月更新 MVR 系统的信号读数，并与其他知名分析师的结论做对比。

[Jeff Sun @sunyefei]

@MaxCryptoSpace Jupiter收购大量优秀项目和团队是对的，也是同为明日新星hyperliquild所欠缺的。科技型初创公司就应该用高额盈利不断拓宽护城河。至于币价上不去，那是回购通缩比不过持续ASR和团队解锁造成的通胀结果，不回购肯定跌更多。这��基本的经济学常识，创始人不可能不懂，我觉得是在为今年的周年庆造势铺垫吧

[Jeff Sun @sunyefei]

@MeteoraCN @hongkongdoll 看了一天的新手训练营，恨没有早点研究DLMM

[Jeff Sun @sunyefei]

I’ve taken the orange pill by locking in my $SIGN claim with a bonus—ready to HODL strong! Let’s ride this wave together. 🧡🧡 #SIGN #OrangeDynasty https://t.co/v7fYMlb6Pf

[Jeff Sun @sunyefei]

@off_thetarget 月底来回头看，这预测准确率大概30%

[Jeff Sun @sunyefei]

@6666huo @BTCdayu 会爆的，如果可转债无法偿还的话

[Jeff Sun @sunyefei]

@mdzzi 空军已待命，请首长指示

[Jeff Sun @sunyefei]

长期主义者社区，nice! https://t.co/YZW8361YLU

[Jeff Sun @sunyefei]

@TingHu888 如果项目收益不回馈币东，那为什么要发币，纯割���菜？

[Jeff Sun @sunyefei]

@jiujinshan2022 那你空它了没

[Jeff Sun @sunyefei]

@0thTachi In addition, how come the top 10 holdings of $askj account for more than 80%? Such a holding structure makes it difficult for people not to suspect that it is a conspiracy