Multicoin 合伙人谈投资决策:如何寻求投资的最优解?

Foresight ResearchPublished on 2022-04-12Last updated on 2022-04-12

Abstract

在整个投资的过程中,这种简单概念的确认非常重要。

本文将探讨在投资过程中几乎是最重要的那个环节——「决策」。

这里的投资决策不止包含个人的投资选择,也包含Multicoin从机构视角进行投资前会考虑的一系列问题,以及一些注意事项。

如果用最简洁的文字概括一下,我觉着是:

简单>复杂

简洁>冗长

那么,如何在「决策」时做到简单呢?

即,通过对单一变量进行优化。

在生活中你会做出很多决策——你应该住在哪里?你应该开什么车?你应该吃什么?等等。

理论上有无数的变量需要优化。

处理「决策」的一种方法是列出这些变量的清单,并为每个变量的重要性创建某种加权标尺,然后尝试沿着若干维度评估每个潜在的决定可能带来的结果,得出加权分数,并以此寻求最优解。

这种方法听起来很理性,但据我所了解,对于大多数的决策来说,某一个单一的变量可能比所有其他变量加起来都要重要。

这就是为什么「决策」的关键不是建立复杂的效用模型,而是要清楚地知道你需要优化其中哪一个变量。

举个简单的例子,我基本上只买一个品牌的T恤,我知道它很合身,这样我就不用考虑穿什么,只需要根据天气或心情来搭配颜色就可以了,这可以减少在一些琐碎的事情上的精力花费。而我午餐吃沙拉和蛋白质,不吃碳水化合物,因为碳水化合物让我在下午犯困,这可以最大限度地提高一天的工作效率。

你也可以把这个框架扩展到投资上。

我们的投资备忘录的第一部分被称为「摘要」,而这部分只允许用三句话以内的篇幅来进行概括,如果你无法用三句话精准地阐述投资该项目的理由,那么我们就不会进入到下一步的讨论环节之中。

举几个例子来看:

Helium:从根本上降低了建立物理WAP网络的成本结构;

Solana:技术的可扩展性让生态可扩展;

Graph:GraphQL可用于抽象检索,而高效的分散读取对大型应用很重要;

Fractal:NFT游戏市场潜力巨大,团队实力足够强;

LayerZero:可以有效优化当下跨链非常复杂的状况;

Arweave:永久存储是一种全新的基础技术,可以释放出我们无法想象的用例;

Superfluid:解决支付问题的最佳解决方案;

Serum:CLOB是处理价格发现问题的最佳解决方案;

Ceramic:可靠的去中心化数据库,可以为dTwitter或dReddit等应用提供动力;

Fluence:dAWSLambda;

UXD:创建完全抵押的去中心化稳定币的最有效的资本运作方式。

当然,以上这些总结都是极简的概括,并没有包含投资决策中的更多细节。但事实上,在整个投资的过程中,这种简单概念的确认非常重要。

每当你在互联网上读到一篇文章,或与朋友交谈时,你就可以将这篇文章/谈话与你现有的整个投资组合进行实时交叉对比。而想要快速处理海量信息的唯一方法,就是将每项投资提炼成极简的概括性描述。

而这就是上面这些案例被刻意简化到极致的原因。

这些例子中最有趣的是Arweave,因为在做出投资决策时,实际上违反了我提到的「理论的清晰性」。在2019年6月我们第一次投资该项目时,该项目的描述还停留在「全新的存储解决方案!未来可期。」

当然我提这个例子并不是要推翻前文阐述的观点,只是说在真正的投资过程中,可能确实会有一些特例让我们无法精准地进行概括性总结,而如果你要是想继续推进这笔投资,你就必须要承担相应的风险,但这时候你一定要能够搞清楚风险究竟存在在哪里。

Related Reads

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.

marsbit36m ago

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

marsbit36m ago

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."

marsbit50m ago

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

marsbit50m ago

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.

marsbit55m ago

ChangXin Technology: A Cyclical Stock Standing Atop the Cycle Peak

marsbit55m ago

Trading

Spot
活动图片