Let him be strong: Buy-side strategies for June under the impact of tail events. June 1st marks the day over 20 million people in Shanghai emerged from the Covid lockdown. If in early 2020, someone told you that China's most international metropolis would experience such a lockdown, shutdown, and collective memory, would you have believed them?
This is a tail event.
Not because no one thought of it, but because when people are in a trend, they always overestimate the stability of the current order and underestimate the impact of extreme inflection points on the real world.
The market is the same.
Humans are good at extrapolating recent events into the future but poor at understanding true structural jumps. Semiconductor storage is an example. In 2023, DRAM and NAND were at a cyclical low with a market size under $100 billion. By 2026 or 2027, AI demand and price cycles will push the entire industry to a completely different magnitude. If you ask an analyst to make a linear prediction of a tenfold change at the bottom, many models become meaningless.
So this week's title is that line from Jin Yong:
Let him be strong; the breeze brushes the hillside.
The more chaotic the outside world, the more you need your own internal order.
But "firmness" here is not brainless all-in, not shouting slogans, and not falling in love with a stock. True firmness comes from knowing what to look at, knowing which variables are changing, knowing what expectations are already priced in, and knowing how your position should withstand the impact of tail events.
June focuses on three things:
Numerator, Denominator, and Position Discipline.
I. The Numerator: Is Agentic AI Starting to Create Real Enterprise Productivity?
Valuation looks at the discounted future cash flow; first, look at the numerator. The main theme of the market remains Agentic AI. AI is moving from training to inference, and then to agentic intelligence. It's not just models getting smarter; AI is starting to have purpose, execution, and feedback, truly entering enterprise production processes.
This is just the beginning.
In the past few years, AI infrastructure hardware was the first stage to realize gains—the shovel-sellers. GPUs, networking, memory, servers, power, and liquid cooling were Phase 1. But if the final loop of AI is just consuming tokens without making money for customers or allowing enterprises to deploy AI as productivity, it will eventually end in a mess. If customers don't make money, the economics of AI capex won't close.
So the true second phase is not just "bigger models," but "useful AI."
Usefulness means AI can enter enterprise organizations, understand context, connect data, execute processes, reduce costs, increase revenue, shorten cycles, and reduce human friction.
Here, a huge problem arises: the model itself is just a capability, not organizational productivity.
Tokens are lone heroes. They must be transformed into true organizational productivity through enterprise data, context, software orchestration, security, scheduling, permissions, processes, and feedback mechanisms.
This is why the SaaS worth watching in the future isn't just those connecting to an AI API, but those who can become the "intermediary" for the enterprise AI execution layer. Whoever controls enterprise data and context, whoever can orchestrate AI into real workflows, and whoever can ensure security, permissions, and auditing is who will survive and even be revalued in the Agentic AI era.
This is the AI innovation diffusion I always talk about. Early on, it's high-value white-collar knowledge workers. Later, it's traditional diffusion in enterprise and mass scenarios. There is a chasm in between.
If we can't cross it, AI infrastructure is just a capex carnival. If we do, AI can form an investment return loop. This is the first anchor of the June strategy:
The AI theme is not over, but the market will increasingly distinguish between those just talking about AI and those truly turning AI into enterprise productivity.
II. The Denominator: Oil Prices, Inflation, Employment, and the Fed's Rate Cut Option
The denominator looks at the discount rate. Behind the discount rate are inflation, interest rates, and risk premiums. The most important denominator in June isn't what the Fed says, but whether two conditions can appear simultaneously:
Oil prices go down, and employment weakens.
Since the new Fed chair took office, the inflation observation criteria have changed, placing more emphasis on an inflation index that excludes extreme volatility. Without getting into academic details, the essence is to reduce the interference of short-term shocks on inflation data, giving policy more room for interpretation.
What does this mean? Not that the Fed will definitely cut rates, but that the Fed has gained a "rate cut option."
As long as the oil price center shifts downward and employment data begins to weaken, the rate cut narrative suddenly gains legitimacy, potentially exceeding market expectations.
This is why the US-Iran situation and oil prices are so important. Trump didn't suddenly fall in love with peace. Oil prices have shifted from a diplomatic issue to a common denominator for elections, inflation, fiscal costs, and recession risks.
As the midterms enter the second half, the political demands of oil-producing and oil-consuming states differ. If oil prices remain high for long, the risk of a US economic recession by year-end will rise, and fiscal costs will be pressured by both high interest rates and high oil prices. If the US-Iran situation cools and oil prices drop, it becomes easier for the Fed to exercise this rate cut option.
So the key to the June market isn't just "rate cuts are coming." More accurately:
The rate cut option is valuable again, but whether it is exercised depends on oil and employment.
If June employment data weakens while oil prices fall, the market will first trade on denominator repair. Small-cap growth, software, REITs, consumer discretionaries, and high-duration assets will see a rebound, and the pressure on AI crowding will temporarily ease.
But this isn't a theme switch; rate cuts only lower the denominator, while AI is changing the numerator.
Once the rate cut happens or the option cannot be exercised, capital will still return to core AI assets: semiconductors, memory, servers, CPUs, networking, power, and software that can truly handle AI productivity.
June is not a simple risk-on; it's a denominator window. Many assets can rebound, but few can be continuously revalued.
III. Predicting the Future is Ridiculous, but Valuation Anchors Must Exist
Whether it's the numerator or the denominator, human predictive ability is very limited. I can say I'm one to three months ahead of the market on some themes, but so what?
When Intel was at $70, I said we must wait for earnings verification. When MU was near $500, I said reduce positions first.
Seeing the right direction doesn't mean you can abandon discipline. Companies are not static functions, markets are not static models, and CEOs are not static variables. This is the hardest part of investing.
You cannot precisely predict all dynamic changes in advance. A company's internal development involves human initiative, management's strategic learning, supply chains, inventory, customer demand, technical routes, and capital market feedback.
Even CEOs are learning dynamically. Intel is a classic example. It was hard to predict a year ago that at this stage of AI development, CPUs would be re-emphasized by the market in inference and Agentic AI architectures. Previously, everyone focused only on GPUs, only to find that large-scale agent operation requires not just computing power, but also CPUs, memory, networking, storage, and scheduling.
Even management might emphasize inventory management one quarter, only for a change in demand structure to turn what was seen as a burden into a source of revenue and gross margin. This is the real world of business operations.
So, does valuation matter? Of course.
But valuation isn't static fortune-telling. Valuation is an anchor, a coordinate system for constantly updating judgments when facing new information. Discounted future cash flow is not about calculating an eternal target price, but about answering several questions:
What does the market believe now?
How much has been paid for this belief?
Is new data verifying or overturning the original hypothesis?
Is the duration of the company's competitive advantage being extended or compressed?
Does management have the ability to reactivate dormant assets?
Behind this is the value of people, the value of the CEO, and the value of corporate creativity.
The AI era makes it easy to think everything can be modeled. In fact, truly valuable companies still depend on how people organize technology, call upon resources, and turn complex systems into executable productivity.
A great CEO is not a footnote in a financial report but a core variable in the path of corporate value creation.
IV. Investing is Not About Predicting Everything, but Preparation
Since the future is unpredictable, what should investors do? The answer is not to lie flat, nor to bet everything.
The answer is preparation.
You need core positions and technical positions. You need to know that cash is also a position. You need to know when to hold, when to audit, when to trim during a carnival, and when to add during a panic.
If you are bullish on the AI theme, don't get washed out by a few points of volatility. But if a ticker is at a high, before earnings, amid market hype, and extreme option sentiment, you cannot pretend the risk doesn't exist.
Core positions handle the big trends; if the trend doesn't change, ignore the wind and rain. Technical positions handle volatility based on your time, energy, and risk tolerance. Cash positions wait for mispricing after tail events.
This is buy-side discipline.
Pay special attention in June: if expectations for lower inflation and interest rates rise, some laggard assets will have a window for a catch-up rally. This isn't for you to abandon the main theme to chase every rebound, but to appropriately allocate to assets suppressed by the denominator whose fundamentals are not broken.
What's truly important is distinguishing: What is a rebound? What is a revaluation? What is a short squeeze? What is the main theme? What is just liquidity repair?
The easiest mistake to make in the market is treating every rise as a new theme and every fall as the end.
V. The Most Important Task for Subscribers This Week
New subscribers, please make sure to read the "Subscriber Guide" first.
Full version: Subscriber Release and Subscriber Knowledge Payment Wealth Creation Guide
Those are not community rules, but investment mentalities.
This subscription is not an ordering counter or a private investment advisor. I will not answer individual requests on what price to buy a stock, how to allocate positions, target prices, or whether to stop loss.
What you should really do is learn to read the main themes, verify hypotheses, and understand position discipline.
In my homepage article column, two articles are particularly important:
The first is "The Guide for Those Who Missed Out".
It talks about how to get skin in the game when you agree with a long-term theme but missed the first stage, how to enter in batches, and how to avoid missing big cycles due to perfectionism.
The second is "Investor's Guide to Investing and Position Management Facing Event Shocks".
This one is more suitable for now. It discusses how to manage core, technical, and cash positions under the impact of tail events and how to maintain actionability amidst uncertainty.
June is not a prediction problem; June is a discipline problem.
You don't have to guess every data point correctly; you just need to know, after the data comes out, whether it changes the numerator, the denominator, or is just market noise.
VI. Finally: In the AI Era, the Most Expensive Thing is Not Tokens, but Judgment Responsibility
Finally, a more revolutionary prediction. In the AI era, output is getting cheaper. There are more tokens, more articles, more summaries, and more opinions.
But what is truly scarce is not the number of words. What is truly scarce is judgment, trade-offs, responsibility, and a reviewable chain of causality.
Jensen Huang's Taipei speech repeatedly emphasized that the direction is essentially "useful AI." AI is ultimately not for producing more text or consuming more tokens, but for producing better results.
Investment content is the same.
Value doesn't come from who writes more or who generates faster. True value is: in a complex market, who can compress variables into judgments, who can provide boundaries in uncertainty, and who can review their own mistakes while sticking to their main theme at critical moments.
I wrote these words one by one, then optimized the layout and format via AI.
This isn't because manual input is sacred, but because behind these words are trade-offs, risks, judgments, and the traces of continuous verification and correction in the weekly series over the past few months.
AI can generate content. But true buy-side judgment must be responsible for the results. The ultimate value of AI is not token output, but the final result.
This is true for enterprises, for investing, and for content.
A word is worth a thousand pieces of gold, not because the word is expensive, but because the judgment is.
Educational sharing for more granular valuation and investment strategies. See subscription sharing for the first updates.
All rights reserved. Please contact me for authorization if you wish to clip or move content.
