To The Moon

This AI and semiconductor correction is exactly the type of market that tests my discipline. Prices have fallen enough to become tempting, but I don’t think this is the moment to blindly shout “BUY THE DIP!” and throw all my cash into the market.

I’m still cautious. At the same time, I also don’t want fear to stop me from accumulating quality AI and semiconductor exposure when valuations are being reset. So instead of trying to predict the exact bottom, I’ve started collecting gradually, mainly $Direxion Daily Semiconductors Bull 3x Shares(SOXL)$ and $ARM Holdings(ARM)$ , while keeping higher-risk names like $CoreWeave, Inc.(CRWV)$ and $NEBIUS(NBIS)$ on my watchlist.

Buying the AI Dip

🔥 Why Did the AI Trade Suddenly Lose Its Shine?

The recent correction isn't simply investors taking profits. Several concerns have hit the AI narrative at roughly the same time.

One major trigger has been interest rates. A hotter inflation reading increased expectations for higher rates, which matters especially to capital-intensive AI infrastructure businesses that depend heavily on financing. The selloff became closely associated with concerns surrounding CoreWeave's debt load and borrowing costs. CoreWeave's 2031 bond yield had climbed above 16% before easing, while the company's shares lost roughly a quarter of their value during the previous selloff week.

This is important because the market is starting to ask a different question.

AI Trade loses its shine?

For the last few years, investors mainly asked:

“How big can AI become?”

Now the question is increasingly:

“Who is actually going to make enough money from AI to justify all this spending?”

That distinction matters to me.

💰 The AI Capex Boom Is Starting to Meet the Cost of Money

AI infrastructure isn't cheap. GPUs, data centers, networking equipment, power infrastructure and cooling all require enormous upfront investment.

That becomes much more uncomfortable when the businesses financing this expansion are paying expensive interest rates.

AI Capex Boom

CoreWeave has become almost a poster child for this debate. Goldman Sachs recently upgraded the stock to Buy, arguing that tight computing capacity and CoreWeave's performance advantages could support its economics. At the same time, the credit market has been much more cautious about financing highly leveraged AI infrastructure.

That contradiction fascinates me.

AI demand can be strong while an AI company's stock can still be risky.

Strong revenue growth doesn't automatically mean strong free cash flow. And strong demand doesn't eliminate financing, depreciation, capex or refinancing risks.

That's one reason I haven't rushed to bottom-fish CoreWeave.

🇨🇳 Then Came Another Shock: CXMT

The semiconductor selloff also received another unpleasant surprise from China.

$Cxmt Corporation(688825)$ CXMT's progress in DRAM technology raised concerns over future competition in server memory. The company has announced a next-generation DDR5 RDIMM targeted for launch before the end of 2026, contributing to pressure on Micron and other memory stocks.

CXMT

But there's an important detail that prevents me from becoming overly bearish.

The reported CXMT development is more directly relevant to conventional server DRAM. Fundstrat's Kent Fung noted that CXMT's 4F² design does not directly compete with the HBM used in AI servers, arguing that the initial market reaction may have overstated its near-term implications for AI memory.

So I'm watching this carefully rather than assuming China's progress suddenly destroys the entire AI semiconductor thesis.

🤖 My Bigger Concern: Can Big Tech Monetize All This AI Spending?

This is the question I think matters most.

I don't doubt AI adoption itself. What I question is whether the enormous amount of capital being deployed across the ecosystem can generate returns quickly enough to justify today's expectations.

Even after this correction, investors still need to differentiate between companies with strong cash generation and those whose growth requires continuous access to outside capital.

AI Spending Concern

That's why I'm buying selectively rather than aggressively.

A lower share price does not automatically mean lower risk.

Sometimes it simply means the market has finally noticed the risk.

🛒 What Am I Actually Buying?

🟢 SOXL: My Aggressive Semiconductor Dip Play

I've started collecting SOXL, but only gradually.

My reasoning is straightforward: if this is ultimately a correction within a longer semiconductor and AI infrastructure cycle rather than the end of that cycle, semiconductor stocks could rebound strongly when sentiment improves.

But SOXL is a 3x daily leveraged semiconductor ETF, so I don't treat it like a normal long-term ETF. Volatility and daily compounding can work brutally in both directions.

For me, this therefore isn't an “all-in” position.

Small entries. Keep ammunition. Accept volatility.

🟢 ARM: Buying the Architecture Behind the Trend

ARM is another name I've started collecting.

What attracts me is that my thesis doesn't depend solely on one AI accelerator winning the race. I see ARM as exposure to a broader computing ecosystem.

The correction gives me a chance to build the position at a better valuation, but again, I'm not assuming the first dip must be the bottom.

My strategy is accumulation rather than prediction.

SOXL & ARM

👀 CoreWeave & NBIS: Tempting… But I’m Not Brave Enough Yet

This is where my greed and fear are fighting each other.

CoreWeave and Nebius (NBIS) are both on my watchlist, but I haven't convinced myself to seriously bottom-fish either one yet.

CoreWeave has already shown how quickly sentiment can reverse. After its sharp decline, Goldman Sachs upgraded it from Neutral to Buy, calling the pullback an opportunity and pointing to tight compute capacity.

But I'm more concerned about the other side of the equation: capital structure and financing risk. CoreWeave's debt concerns have become one of the focal points of the recent AI correction, and its financing costs illustrate why heavily capital-intensive neocloud models can be vulnerable when rates rise. Nebius interests me for similar reasons, but I currently view these neocloud names as substantially more speculative than simply owning semiconductor exposure. During the recent stress, Nebius had fallen more than 31% over two weeks, illustrating just how violently this theme can move when investors become worried about AI financing.

CRWV & NBIS

The upside could be huge if AI compute demand continues expanding.

The downside could also be painful if financing becomes more expensive or the return on all this AI infrastructure disappoints.

For now, they stay on my watchlist.

🎯 My Strategy: Don’t Catch the Knife. Collect It Piece by Piece.

I don't know whether this is the bottom.

Honestly, nobody does.

My approach is therefore quite simple:

SOXL: collecting cautiously ARM: collecting cautiously CRWV: watching NBIS: watching Cash: definitely keeping some

My strategy

I'm deliberately leaving cash available because another leg lower would not surprise me.

If semiconductor fundamentals remain healthy while prices continue falling primarily because of macro fear and sentiment, I'll become more interested.

But if earnings expectations start deteriorating materially, AI capex gets cut, or credit markets become even more uncomfortable financing AI infrastructure, then a falling price may be warning us about something more fundamental.

🧠 Fear Can Create Opportunity, But It Can Also Be a Warning

This correction has reminded me of something important:

Buying the dip works beautifully until the dip becomes a change in fundamentals.

That's why I'm neither fully greedy nor fully fearful this round.

I'm cautiously greedy.

I'm willing to start accumulating businesses and themes that I still believe have long-term potential, but I'm not willing to assume every AI stock that has fallen 20%, 30% or 40% is automatically cheap.

Sometimes the best opportunity isn't buying the absolute bottom.

Opportunity or Warning?

It's surviving long enough, with enough cash left, to keep buying when everyone else has already run out of ammunition.

So yes, I'm buying this dip.

But I'm buying it slowly, selectively, and with one hand still firmly holding onto my cash. 📉➡️📈

This is my personal investment view and not financial advice.

@TigerStars @Tiger_SG @Tiger_comments @TigerClub