I've been following SOXL closely. It plunged from over $100 to around $40 in just a few trading days—a drop of more than 60%.
Honestly, this sell-off was still worse than I expected.
I've been bearish on semiconductors for a while, believing that AI capital spending is excessive, chip valuations are too high, and we may now be in the latter half of this semiconductor cycle.
I still haven't changed my views on this.
But I went long on SOXL at $43, with 8x leverage.
Why?
Because I think this downturn is no longer purely about fundamentals. Political maneuvering, a rush for the exits, and market panic have all piled on, making the short-term sell-off seem clearly excessive.
I don't think the semiconductor industry's long-term correction is over, but I think it's worth taking a chance on an oversold rebound here.
One: AI is practically becoming a tool in U.S. political battles
One important backdrop to this U.S. stock market crash is that, after the midterm elections, the market suddenly amplified policy disputes over the construction of AI data centers.
The Democratic camp had already proposed increasing scrutiny of AI data center construction, citing concerns that the many data centers would consume electricity and drive up household bills, as well as issues involving water resources and energy supply.
Trump's camp has once again blamed the Democrats for the stock market plunge.
Reading these news stories, I feel that AI is no longer just a matter of technological development—it's increasingly becoming a tool in America's domestic political power struggle.
But I think the market is reading too much into it.
What kind of industry is AI now? I think it's become a matter of national destiny.
The United States and China are both competing for technological advantages in AI. In that context, how could the United States voluntarily give up on AI development just because it's worried about rising electricity bills?
Would it really hand China the technological edge in the competition for the decades to come?
I don't believe it.
Of course, restricting data center construction and restricting AI development are not the same thing.
In the future, the United States could certainly tighten approvals and raise energy requirements, making companies more cautious about investing.
But that's entirely different from the AI industry stopping its development.
So I think the market has rapidly priced potential future policy constraints into stock prices.
Policy risks are real, but stock prices have fallen much faster over the past few days than the fundamentals have changed.
That's why I think there's an oversold opportunity.
Two: But why am I still bearish on semiconductors over the long term?
I've talked about this view many times before.
I've never doubted that AI has value.
The issue is that AI having value doesn't mean the market can build computing capacity recklessly, regardless of cost.
Companies like Google, Microsoft, and Meta are spending so much each year on data centers. How much profit will they actually generate in the future? Can their returns on investment justify today's capital expenditures?
I think there's a great deal of uncertainty here.
And semiconductors are a cyclical industry by nature.
In the past, the market's valuations of companies like Nvidia, Micron, and SK Hynix had already priced in expectations of sustained high growth for several years to come.
If growth slows even slightly in the future, valuations could fall sharply—even if these companies remain profitable.
So I've always believed that AI will develop over the long term, but semiconductors may not keep rising according to the same logic as over the past two years.
These two things are not contradictory.
The market may indeed be repricing semiconductors right now, but repricing doesn't mean stock prices should fall endlessly in just a few days.
I'm bearish on medium- and long-term valuations and the cycle; that doesn't mean I think prices should fall every day.
Three: When it fell to around $40, I actually started thinking about going long
SOXL fell all the way from around $107 to $60, then $50, and finally just above $40. After-hours, it even briefly hit around $38.
I think that was an extraordinarily steep drop.
SOXL is a daily 3x leveraged ETF, so consecutive declines magnify losses, and it also suffers from volatility decay.
When people see this kind of price action, they wonder whether the entire AI industry is about to collapse.
But I think exactly the opposite.
When everyone starts interpreting every negative development in the worst possible light, I look into whether there might be a rebound opportunity.
To be clear, I don't think $40 is necessarily the bottom.
I also don't know how low it could go.
It could be $38, $30, or even lower.
But I think short-term panic is already running very high. Even if prices fall further, there could still be a substantial rebound along the way.
And that's the rebound I'm trading.
I don't need SOXL to climb back to $100.
It doesn't even need to return to a bull market.
A rebound from the $40s to the $50s would already be a substantial opportunity for someone like me, trading with a small amount of capital and leverage.
Four: I went long at $43 with a stop at $38—why was I willing to buy?
This time, I didn't jump in with a large position to make a bet.
I used only one-tenth of my margin, with 8x leverage. I opened a long position around $43, with a stop at $38.
For now, my first take-profit target is $55, and my second is $65.
$65 is an extended target in the event of a strong rebound; I won't stubbornly wait for it.
If it reaches $55 and the rebound is clearly losing steam, I'll consider taking some profits rather than giving back all my unrealized gains just to make a little more.
My trading logic this time is actually very simple:
I think it's worth taking a chance on a rebound right now, but I don't think I can accurately pick the exact bottom.
If it hits my $38 stop, then I'll take the loss.
Based on this position, if it normally executes at $38, I'll accept the loss.
I'm not going to keep adding to my position and holding on just to prove that my judgment was right.
That's also what I learned from my earlier SOXL trades.
Previously, I shorted at $98 with a stop at $102. The rebound stopped me out, and only then did SOXL begin its plunge.
I got the direction right, but still lost money on the trade.
So judging the market is one thing; figuring out how to participate with limited capital is another.
Five: If SOXL keeps falling to $30, I might buy again
Some people might find this strange.
You went long at $43 and got stopped out at $38. Why would you dare buy again if it falls to $30?
Because I don't think $43 is the bottom.
What I'm looking for is a potential short-term recovery after an extreme sell-off.
Suppose SOXL falls to $30.
I won't add to my position just because it's cheaper.
First, I'll see whether the market is stabilizing, whether SOXX and major semiconductor stocks have also stopped falling, and whether buyers are actually stepping in.
If they are, I might go long again with a small position.
If I'm wrong, I'll stop out again.
These are two completely separate trades—not a case of losing on the first trade and using the second to lower my cost basis and stubbornly hold on until I break even.
Of course, I won't keep trying indefinitely, either. Even though my capital is small, consecutive stop-outs can still eat into it, so I need to calculate the risk on each trade and how many losses the account can withstand.
For me, a low price alone is not a reason to buy.
A steep price decline, signs that buyers are stepping in, and potential gains that clearly outweigh the risk of a stop-out—that's why I'm willing to take a bet.
Six: I don't think this is a market where the only option is to short
Semiconductors have been falling steadily lately, and I was mostly looking for short trades before.
But after this much of a drop, I think it's time to change my trading approach.
When SOXL was at $70, I didn't like chasing it lower because it was already close to its previous low, making the risk/reward unattractive.
When it fell to around $40, the situation changed.
Chasing it lower could certainly make money, but you'd have to accept the risk of a sharp rebound at any time.
I chose to go long with a small amount of capital to take a chance on a recovery after this kind of extreme panic.
But there's another distinction here that we shouldn't get wrong.
A rise from $38 to $43 doesn't mean a bottom is confirmed; a rise from $43 to $50 doesn't mean semiconductors are back in a bull market.
A rebound could simply be short covering, or a return of funds that had fled in panic.
What really matters is whether prices can hold their gains and whether the semiconductor sector strengthens in tandem.
If it rebounds to $50 or $55 and clearly runs out of steam, I'll take profits.
I won't refuse to go long just because I'm bearish over the long term.
And I won't suddenly decide that semiconductor fundamentals are fine just because I've gone long.
Seven: To sum up what I really think
I still think AI is an extremely important industry for the future, but the high valuations, high capital spending, and cyclical risks in this semiconductor cycle are far from resolved.
Over the long term, I'm still bearish.
But I think the plunge over the past few days has clearly been driven by too much political sentiment and panic.
The United States won't give up the AI race because of partisan infighting.
Restricting certain data center projects doesn't mean the AI industry will stop developing.
And the more sharply the market falls, the more likely we are to see a rebound after concentrated liquidations and a temporary easing of panic.
That's why I'm willing to try with a very small amount of capital.
If I'm right, I'll capture part of the rebound.
If I'm wrong, I'll take the loss and stop out.
If the market keeps plunging, I'll wait for new signs of stabilization and then reassess whether it's worth entering.
I'm not fixated on buying at the bottom or selling at the top.
I'm increasingly convinced that for small investors hoping to turn a little into a lot, it's not about going all in every time, or betting that you're definitely right.
It's about looking for opportunities where the loss is manageable if you're wrong, but the potential gain is significantly greater if you're right.
There's another very important distinction:
I'm trading a short-term oversold rebound in SOXL, not investing for the long term. SOXL is already a 3x leveraged ETF, and adding leverage through derivatives makes the risk very high. A stop may also fail to execute as planned due to gaps or slippage.
In this kind of market, surviving matters more than catching every move up or down.
My view is clear: medium- and long-term valuation risks in semiconductors remain, but I think this current panic-driven plunge has created a rebound opportunity worth testing with a small position.
I don't care whether this is the ultimate bottom.
I just want to make money on the part I understand and can afford to risk.
This article documents only my personal trades and market views; it does not constitute investment advice.
