To The Moon

The AI Rally Is Being Repriced

The recent sell-off in technology stocks has definitely made investors nervous. When names like Micron, $SanDisk Corp.(SNDK)$  andfall more than 40%–50% from their recent highs, it is easy to ask: Is the AI boom finally coming to an end?

Personally, I don't think so. I see this more as a major reset in expectations, valuations and positioning after an exceptionally strong AI rally. The key question is no longer simply whether AI will grow, but which companies can turn massive AI spending into sustainable revenue, earnings and free cash flow.

From"Buy AI" to "Prove It”

The market has changed significantly. Earlier in the AI rally, almost anything related to semiconductors, data centers or AI infrastructure could rise simply because investors wanted exposure to the theme.

Today, the bar is much higher. Investors want actual orders, revenue growth and profits. That explains why a company can report strong results and still see its stock fall sharply.

Micron is a good example. Long-term demand for memory, HBM and AI infrastructure remains strong, but after a huge rally, investors start asking whether future growth can exceed what is already priced in. When expectations become extreme, "good" is no longer good enough.

Valuation Compression

I Don't Think the AI Bubble Is Bursting

That is why I don't view the current correction as simply an "AI bubble bursting." I see it as the AI trade entering its next stage.

NVIDIA CEO Jensen Huang has argued that a semiconductor bubble is unlikely to emerge in the next several years because the industry remains constrained by supply and the overall semiconductor market could expand dramatically.

A correction in semiconductor stocks does not necessarily mean semiconductor demand is disappearing. Stock prices can fall sharply even while the underlying industry continues to grow. What changes is the valuation investors are willing to pay for that growth.

I Don't Think the AI Bubble Is Bursting

My Preferred AI Exposure

Personally, I remain constructive on AI but much more selective.

$NVIDIA(NVDA)$  remains one of my highest-conviction names because its advantage goes beyond GPUs. Its CUDA ecosystem, software, networking and rapid product development create a powerful moat. SpaceX's decision to use NVIDIA's latest architecture for its AI computing plans is another example of strong customer demand.

However, even great companies can become poor investments at excessive valuations. I therefore prefer accumulating gradually rather than chasing rallies.

I also like $Taiwan Semiconductor Manufacturing(TSM)$  as a long-term AI infrastructure play. NVIDIA, AMD and other advanced chip companies depend heavily on leading-edge manufacturing, putting TSMC at the center of the ecosystem. It gives investors exposure to AI expansion without having to predict which individual accelerator architecture will ultimately win.

AI Ecosystem Heatmap

Memory Is a Different Story

Memory is more complicated.

I remain bullish on$Micron Technology(MU)$  over the long term because AI servers require significantly more memory, especially HBM and advanced DRAM. But memory is highly cyclical, and prices can move quickly when supply catches up with demand.

Therefore, I prefer to accumulate MU gradually rather than trying to call the exact bottom. If the stock falls further while the long-term fundamentals remain intact, I would rather have cash available to average down.

Why I'm Careful With Leverage

I am also cautious with leveraged products. Leveraged ETFs can deliver powerful rebounds after major sell-offs, but they magnify volatility and can suffer from compounding and volatility decay.

I am willing to use leverage selectively, but I don't want it to force me into short-term decisions. Position sizing and keeping cash available remain important to me, especially in a market where daily moves can be extreme.

Buy the Dip — But Don't Buy Everything

So, am I buying the dip or running for the exit?

I am buying selectively, but definitely not blindly.

The biggest risk to the AI thesis isn't necessarily that AI fails. It is that companies spend enormous amounts on AI infrastructure without generating sufficient economic returns. If hyperscalers eventually reduce AI capital expenditure because returns disappoint, the entire semiconductor ecosystem could face another valuation reset.

For now, however, I believe we are still in the middle of a major technological transformation rather than at the end of the AI story.

The Next Phase Will Be About Winners and Losers

The easy phase — where almost every AI stock could rise together — may be over. The next phase will be about execution, profitability and competitive advantage.

My investment philosophy remains simple: consistency over noise.

I don't need to predict the exact bottom. If I can gradually accumulate high-quality companies during periods of excessive pessimism, manage my position sizes and stay invested through volatility, I believe the long-term opportunity remains attractive.

Winners or laggards?

AI is not dead — but the era of buying anything with “AI” in its name and expecting it to go up probably is.

The next winners will be the companies that prove, quarter after quarter, that AI is not just a story, but a real business.

Summary

@Tiger_comments  @TigerStars  @TigerClub  @Tiger_SG