【研报精选】抽丝剥茧看穿 LSD 高阶玩法本质

Foresight NewsPubblicato 2023-03-05Pubblicato ultima volta 2023-04-03

Introduzione

ETH Staking 奖励是相对平等的,那么那些高收益的 LSD 项目是怎么回事?有一句名言:如果你不知道收益从哪来,那你就是收益来源。

本文将从 Frax/Yearn 的 LSD 到杠杆 Staking 再到 re-staking,为你抽丝剥茧,一眼看穿这些 LSD 高阶玩法的本质。

协议资产激励

首先,我们用一张图来分析 Frax/ frxETH 的收益分配体系,用黄色标注出收益来源:

•FXS 代币激励,就是 Frax Finance 的权益直接激励给 frxETH Pool

•Frax 控制的 vlCVX/veCRV 投票间接让 CRV/CVX 激励给 frxETH Pool

•ETH 的 Staking 激励

1/2 无论是直接还是间接,本质上都是 Frax Finance 出的激励,无非是形式和包装不同。所以我们将此图简化一下,就非常明了,收益来源和分配就是:

•Frax Finance 协议资产 / 协议收入激励了 frxETH

•frxETH/sfrxETH 两者的 Staking 收益全部给了 sfrxETH

sfrxETH 吃「两份」 Staking 收益,APR 自然就高了。

同理我们按此分析一下 Yearn 近期预计出品的 yETH:

也把他简化一下,一眼就看出收益来源和分配就是:

•Yearn Finance 协议资产激励流动性从而间接激励了 yETH

•yETH/st-yETH 两者的 sfrxETH 收益全部给了 st-yETH

不过需要注意的是,这里所谓的 「协议资产」并非仅是 Yearn 的国库所有的,可能还包括 Yearn 从 yCRV 产品那里挪过来的投票权。

是的,frxETH/st-yETH 超额收益来源都是协议资产在直接或间接的补贴!

诶,这么看下去岂不是无限套娃了?但

•协议资产是有机会成本的,比如 vlCVX/ veCRV 本身是可以拿来赚取贿赂收益

•协议如果不是空气币,就要赋能,比如收入抽成,达到一定规模后补贴和抽成接近后,那么超额收益便不复存在了

那这么做的意义在哪里?因为预期 LSD 在上海升级后将会迎来爆发,参照下方推文,LSD 是做大了也有护城河的,那么趁此扩大规模抢到门票,到时候便有可观的收益。

超额收益有可持续性吗?质押资产 / 协议资产上升无疑会稀释超额收益,若协议资产由于价格等原因相对 ETH 在上升的话,那么将会延长这一过程,反之亦然,也就是说这个模式具有明显的反身性。因此这种模式虽然本质是补贴,但是精巧复杂的包装也是有必要的,通过更精彩的叙事可以提升协议资产的价值,从而推动正向螺旋。

简化来看的话,收益来源就是协议资产和标的资产两侧,那么两端的资产类别其实也是可以更多样化?bestLSD 这个项目(名字挺土狗)就自称协议资产端和标的资产端放开来打,协议资产从 GMX/GLP 搞到 CRV/CVX/Velo,标的资产从 AMM LP 搞到杠杠 Staking ,总而言之是在两端合力推高收益。

虽然这个玩意儿看上去有点野,不过这也提醒了我们,目前除了持有 CRV/CVX 的协议,持有任何其他生息资产的协议其实也都可以参与角逐,这块的门槛其实很低,将来可能会很卷。So it’s another use case for GLP and gDAI, gud!

息差套利

至于通过借贷协议做杠杆 staking 增大收益,参见下方推文,本质上就是息差套利。虽然看上去存款方有点吃亏被套利,但借款方也承担了额外的 LSD 协议风险和流动性风险。借贷协议充当了将质押利率传导到原生资产上的通道。

再打一份工

Re-staking,所谓第三方借用 ETH 主网安全性,不专业的但简单的形容就是:

•网络、应用需要运作需要节点之类的保证安全

•节点之类的作恶需要付出代价,才能保证安全

•那么就先让他交保证金,作恶就罚没

•保证金也要资本效率,ETH Staking 是主要生息方式,那就用相关资产保证

•Staking 资产的进出、罚没等管理,需要人来执行吧,这就是 re-staking

•你给第三方做节点,再打一份工,第三方自然给你工资才干吧

综上所述,协议资产激励、息差套利以及再打一份工是三种超额收益来源,必须强调的是,复杂的玩法会带来更多的风险,所以有必要评估超额收益是否能够覆盖增加的风险,毕竟无论是什么收益来源,最后都可能成为科学家的收益来源。

LSD 本来涉及到和 Validators 这种线下实体的交互,是相对来说比较「重」的工作,但上面几种高阶玩法,通过包装 LSD 资产来运作,转化成了相对比较「轻」的工作,这里的门槛也显然会比传统 LSD 更低,花样自然也会更多,值得持续观察。

Letture associate

The Semiconductor Industry's Most Obscure Chokepoint Material: It Has Multiplied in Price Unnoticed, and China Just Banned Its Export

"Helium: The Silent Choke Point in Semiconductors" Helium, despite being the universe's second most abundant element, is critically scarce on Earth. It cannot be synthesized artificially and escapes Earth's gravity once released, making it a non-renewable resource. Yet, it is indispensable for advanced semiconductor manufacturing, requiring ultra-high purity (99.9999%) for processes like wafer etching, leak detection, and thermal management in lithography tools, with no viable substitutes. This low-profile resource, whose price has doubled unnoticed, suddenly gained attention when China imposed an export ban on July 10th. The ban is a response to a severe global shortage triggered months earlier. In March, missile attacks damaged the world's largest helium production hub in Qatar's Ras Laffan, halting a third of global supply with repairs expected to take 3-5 years. Concurrently, Russia tightened export controls, and the US sold off its federal helium reserve. Approximately 200 specialized transport containers, each worth ~$1 million, were stranded with perishable cargo. Global spot prices skyrocketed, and chipmakers like TSMC warned of potential impacts. China, which imports over 84% of its helium (primarily from Qatar and Russia), faces a severe squeeze. Despite being helium-poor, China has developed domestic extraction technology over five years, processing byproduct gases from LNG plants to produce ultra-pure helium. Domestic output, while still only covering less than 20% of demand, enabled the recent export ban aimed at securing domestic supply for its own expanding chip industry. The industry now faces "Helium Shortage 5.0." With key global sources constrained for years, this invisible gas—once wasted on party balloons—has become a critical, geopolitically charged bottleneck for chips, healthcare (MRI machines), and beyond, forcing the world to finally treat it as the finite strategic resource it is.

marsbit32 min fa

The Semiconductor Industry's Most Obscure Chokepoint Material: It Has Multiplied in Price Unnoticed, and China Just Banned Its Export

marsbit32 min fa

Early Investor of Unitree Bets on a Brain-Computer Interface Company

"Wabo Technologies, a brain-computer interface (BCI) startup, has secured new funding from Vertex Ventures (under Temasek), Hongtai Fund, Chengdu Future Industry Fund, and Huifengda Capital. Vertex Ventures is known for its early bet on Unitree Robotics. Inspired by the sci-fi film "Avatar," founder Zhang Xingzhi has pursued human-machine interaction for over a decade. He established Wabo in late 2025, coinciding with China's push for BCI as a future industry. Unlike companies focused solely on medical hardware, Wabo positions itself as a model company. Its core mission is to develop a "Human Neural Intent Model" that decodes EEG, EMG, and behavioral signals to understand a user's goals, confirmations, corrections, and stop intentions before physical action, translating them into standardized control signals for machines, robots, or smart devices. The young, full-stack team, backed by academic collaborators from Tianjin University and Fudan University, quickly attracted investors. The company had previously raised a seed round from Ceyuan Capital. Investors were impressed by Wabo's systematic capabilities in data collection, signal processing, and engineering, seeing potential beyond a single hardware product toward a reusable layer of human intent interpretation. Applications span medical rehabilitation, smart homes, embodied AI, and even space exploration, with Wabo partnering with Tianjin University on pioneering "space BCI" research. Wabo plans a dual-headquarters strategy, with Shanghai focusing on high-end R&D and model development, and Chengdu handling hardware, product development, and supply chain. The BCI sector in China is heating up, with over 30 financing deals in early 2026. Zhang believes future competition will center on who can build a sustainable data-model-scenario loop. He predicts industry consolidation within three years, with only companies that master this闭环 surviving. The vision from "Avatar" is inching closer to reality."

marsbit47 min fa

Early Investor of Unitree Bets on a Brain-Computer Interface Company

marsbit47 min fa

ChangXin Technology: A Cyclical Stock Standing Atop the Cycle Peak

Changxin Technology: A Cyclical Stock at the Peak On July 27, 2026, Changxin Technology topped the A-share market with a market capitalization of 3.28 trillion yuan, surging 465% on its first trading day. The company, which lost 16.3 billion yuan in 2023, reported an estimated net profit of 50-57 billion yuan for the first half of 2026. Its dramatic reversal mirrors the volatile DRAM (Dynamic Random Access Memory) cycle. The DRAM industry is inherently cyclical, with booms and busts every 3-4 years. This is due to product standardization and a significant time lag in supply adjustment. When prices rise, manufacturers expand capacity, but new production takes 2-3 years to come online, often leading to oversupply and price crashes when demand cools. Changxin's performance perfectly tracks this cycle. It recorded deep losses in 2023-2024 during the industry downturn, turned its first annual profit in 2025, and saw profits skyrocket in Q1 2026. This surge is primarily price-driven. The AI boom has led major players like Samsung and SK Hynix to shift 70-80% of new capacity to high-margin HBM (High Bandwidth Memory), creating a severe shortage and price explosion in general-purpose DRAM markets where Changxin competes. However, a massive global capacity expansion is underway. The top three manufacturers have announced nearly $70 billion in capital expenditure for 2026. Changxin itself plans to expand from three to seven 12-inch wafer fabs. This investment will translate into significant new supply in 2-3 years. While DRAM prices are expected to remain high through 2026-2027, price growth is already slowing, and a potential downturn is forecast for around 2028 as new capacity ramps up. A key challenge for Changxin is catching up in the critical HBM segment. While it has delivered HBM3 samples, leaders are already mass-producing more advanced HBM3E. Success in HBM is crucial for gaining true cyclical resilience. In conclusion, Changxin is a commendable company that has broken foreign monopolies in DRAM. Its long-term growth narrative—driven by import substitution and AI—is valid. Yet, its current valuation of 5-6x forward P/E, typical for a cyclical stock at its peak, suggests much future growth is already priced in. AI may extend the current cycle but cannot eliminate the industry's inherent volatility. For investors, the critical question is preparedness for the inevitable downturn when the cycle turns.

marsbit52 min fa

ChangXin Technology: A Cyclical Stock Standing Atop the Cycle Peak

marsbit52 min fa

Trading

Spot
活动图片