The current rally in technology and semiconductor stocks is being driven by three overlapping forces: a semiconductor mania, a space mania, and an options mania. These forces can keep stocks climbing, but they also create exaggerated moves and concentrated risks. The disciplined response is not to predict tops or bottoms, but to manage exposure systematically — through a method best described as dynamic hedging.
The Biggest Risks That Could Interrupt the Momentum
Several distinct risks threaten the current upward momentum.
The Federal Reserve and Kevin Warsh. The most significant risk "came from nowhere" after the Fed meeting involving Kevin Warsh. President Trump picked Warsh and is widely understood to want rate cuts, with Warsh presumed to be dovish. In practice, however, Warsh came across as very hawkish. The reasoning behind that judgment is rooted in the Fed's dual mandate: full employment and price stability. Fed chairs, observed over many years, have always mentioned both halves of that mandate. What was noteworthy about Warsh's remarks is that he never once mentioned full employment — he ignored it entirely — while mentioning price stability several times. He also effectively called the Fed "on the carpet," speaking as though he were not part of the institution, criticizing them by saying that for five years they had stated a 2% inflation target, that inflation has been running above that, and that they had done nothing about it. If Warsh behaves like Paul Volcker (referred to as "Paul Walker" in conversation) — an aggressive inflation fighter — that is a serious risk, because long-dated assets like technology stocks are extremely sensitive to interest rates. There is some doubt that he actually will behave that way, but the possibility itself is a risk.
New memory supply from China. A risk many people have not picked up on concerns the dramatic declines in stocks like Micron. New memory supply is coming online from China, where producers are ramping up. Unlike manufacturing Nvidia GPUs, manufacturing memory is not that difficult. As a result, more supply is arriving — and the market is not pricing this in. The same dynamic is expected to play out in disk drives. The market simply is not thinking about this incoming supply.
The shift from "token max" to "token min." The market keeps repeating the assumption that hyperscalers are spending enormous sums and will somehow be able to monetize that spending. But the situation is shifting quickly. Only about a month earlier, large corporations were operating under a "token max" mentality — telling employees to use AI as much as possible. They then realized the cost was too high. Now the mantra has flipped to "token min" — cut the cost of AI. As companies examine their usage, they are concluding that they do not need the expensive frontier models from Anthropic, OpenAI, and Google for everything they do. For many tasks, cheaper Chinese models — such as some version of DeepSeek — are sufficient. For certain tasks, the cost difference between a cheap Chinese model and a model from Anthropic is on the order of eight to one. This cost migration undermines the monetization story behind heavy AI spending.
Extremely positive sentiment. Proprietary indicators show that sentiment has been extremely positive. Sentiment is never a good timing signal — it cannot tell you when something will turn — but it functions as a condition that tells you things can go wrong. Stretched positive sentiment is therefore a warning, not a trigger.
Despite all these risks, the long-term thesis remains intact. A call made back in 2022 held that money is to be made in AI all the way to 2030, but that the path in between would be treacherous. That call still stands.
How to Position: Diversification and a Two-Bucket Framework
A central question is how to position given these risks: should an investor simply be tactical, accept that volatility is "the game you signed up for," and have the stomach to handle redder days — or should the portfolio contain some anti-beta or lower-beta component? The answer is that there is a better way.
It begins with wide diversification. The portfolio is heavy in tech and very heavy in semiconductors, but it also holds names like Walmart, along with small caps and micro caps. Geographically, it holds both Korea and India. The result is a broadly diversified portfolio rather than a concentrated bet on a single theme.
The core game plan is straightforward, and the puzzle is why more investors do not follow it. Think of the portfolio as a set of buckets:
- The strategic bucket is very long-term.
- The shorter-term bucket takes advantage of technical situations.
The power of the strategic, long-term approach is illustrated by Micron, held from $21.77 to roughly $1,000 — the payoff of sticking with positions over years. (Korea, by contrast, sits in the more tactical category, currently around $200 after being bought at 48.60.) These are offered as examples among many.
Dynamic Hedging: The Long-Term Defense
For the long-term portfolio, the defensive technique is dynamic hedging: when conditions change, hedging begins. Crucially, it is not an all-in, all-out approach. It is a slow, gradual process driven by systematic, quantitative, algorithmic signals rather than gut feeling.
Concrete examples illustrate the cadence:
- Hedges were added substantially one day before the Iran war.
- The market bottomed on March 30th; on April 1st and 2nd, profits were taken on those hedges.
- Hedges were then started again, with more added on June 17th.
- Semiconductor positions are also being hedged.
This dynamic hedging is operationalized through a tool called the "protection band," which is designed to make the process easy to execute.
Trader-On Positions: The Tactical Offense
On the technical side, the strategy uses what are called "trader-on" positions. For example, if Micron gets hit hard, it is bought. The key discipline is separation: the long-term position in a company is traded completely separately from the trader-on position. The long-term position is multi-year, while the trader-on position may last anywhere from a few weeks to a month or two, governed by tight stops and closer targets. This dual approach — long-term holdings handled independently from short-term trades in the same name — has been used for 20 years and "works beautifully."
Mania Versus Bubble: A Crucial Distinction
A vital distinction separates a mania from a bubble. Having lived through the late-1990s bubbles, the defining feature of a true bubble is that valuations become totally detached from reality.
Right now, valuations are not detached from reality. Earnings are increasing. Looking at the P/E ratios of names like Western Digital or Seagate, these stocks are not that expensive. The only genuine uncertainty is whether current earnings represent peak earnings now, or whether the peak comes a year or two years from now. By any reasonable measure, this is not a bubble — and one can say so confidently precisely because the earnings can be examined directly.
What we do have is a mania. The market has been "somewhat overdone." A mania is psychologically driven — fueled by FOMO, by momentum (the "MOMO" trade), and by heavy buying of short-dated options. Depending on which direction the market is moving, this options-driven, sentiment-driven activity creates exaggerated moves in either direction. So this is definitively a mania, but, because earnings remain intact, it is not a bubble. The distinction matters because it shapes whether the right response is to flee or to hedge and stay engaged.
Finding Opportunity and Avoiding Complacency
Where might investors be getting too complacent, and how does one find under-discovered, under-appreciated names? The answer flows from the same framework: distinguish mania from bubble, lean on earnings to gauge whether valuations are truly detached, and use the tactical trader-on bucket to capitalize when good names — like Micron after a sharp drop — get oversold.
Final Guidance: How to Keep Your Head
For investors exposed to these volatile days, the closing advice is direct and serves as the summary of the entire approach:
1. Learn to dynamically hedge your long-term positions.
2. Separate out short-term tactical positions from long-term strategic positions.
3. Combine the two synergistically.
Do this, and the promise is twofold: you will sleep every night, and you will beat the S&P 500 every year.
