After Semiconductor Leverage Liquidation, Will It Rise Again? The core point is “Google CAPEX” and “oil and interest rate risks”

If you simply see this semiconductor rebound as “it fell a lot, so it went up,” you will miss a lot.

The core point is that supply-and-demand pressure has eased as leveraged ETFs and hedge fund positions have been substantially unwound.

Added to this are rising Korean semiconductor exports, the spread of Chinese AI models, big tech earnings season, and data center power demand all tied together at once.

However, while the market looks good if you only look at semiconductor stocks, macroeconomic risks such as rising international oil prices and a rebound in U.S. interest rates are also increasing at the same time.

So the conclusion of this article is simple.

The semiconductor fundamentals are still strong, but whether the current rebound is a new upward cycle or a dead cat bounce still depends on confirming the AI infrastructure investment plans of hyperscalers like Google, Microsoft, and Amazon.

1. Today’s market core news: Semiconductors pulled the entire index up

Based on the original text, the S&P 500 closed up about 0.9%, and the Nasdaq rose about 1.3%.

At the center of the gains were semiconductor and AI hardware-related stocks.

In particular, memory semiconductors, storage, CPUs, and semiconductor equipment names, which had recently seen large declines, staged a strong rebound.

By contrast, large-cap AI semiconductor stocks such as Nvidia and Broadcom, which had fallen less relative to others, saw a more limited rebound.

• Market point 1: Semiconductors have climbed back to major support and resistance zones.

• Market point 2: If they break through strongly here, the chance of returning to an upward trend increases.

• Market point 3: Conversely, if the breakout fails, a short-term pullback, that is, a dead cat bounce, remains possible.

• Market point 4: The next direction is likely to be determined by big tech earnings, especially Alphabet’s capital expenditure announcement.

2. First reason semiconductors rose: The spread of Chinese AI models could actually be a positive for U.S. AI hardware

When a Chinese AI model like Kimi K3 appears, at first glance it may feel like bad news for U.S. AI companies.

That is because if model competition intensifies, margins may narrow for AI service companies and hyperscalers.

But from the perspective of AI hardware and memory semiconductors, the story is different.

The fiercer the AI model competition becomes, the more computing power is required.

In the end, that means more investment is needed in GPUs, HBM, storage, network equipment, and power infrastructure.

In other words, the rise of Chinese AI models gives U.S. big tech a reason to say, “We cannot reduce AI investment.”

In this structure, companies that absorb AI infrastructure investment may actually benefit.

Typical examples include memory semiconductors, HBM, semiconductor equipment, optical communication equipment, and data center power companies.

3. Second reason semiconductors rose: Korean semiconductor exports proved the fundamentals

One of the most important numbers in the original text is Korean semiconductor exports.

It was mentioned that Korean semiconductor exports during the first 20 days of July increased by about 181% year over year.

This figure is not just a matter of expectation, but fundamentals confirmed by actual export data.

It is important that improvement trends are also being detected in key component unit prices such as semiconductor prices, memory demand, and MLCCs.

In the end, it can be said that the recent semiconductor correction was driven more by supply-demand concentration and leverage liquidation than by a collapse in fundamentals.

If the semiconductor cycle itself had turned down, export data and price indicators should have collapsed together.

But now, it looks more like a situation where demand indicators are still alive while investor positioning alone was shaken too aggressively.

4. The most important change: Leveraged ETF and hedge fund deleveraging has progressed substantially

The essence of this semiconductor selloff is that “too many people were betting in the same direction.”

A considerable number of market participants were crowded into long semiconductor positions, and some interpretations suggested the long bias was extremely high.

In such a situation, even good news leaves little room for new buying.

Conversely, even a small piece of bad news can trigger a chain of position liquidations and deepen the decline.

Looking at the JPMorgan and UBS analyses mentioned in the original text, the core point is that supply-demand normalization is progressing.

• A substantial portion of positions related to leveraged ETFs was said to have been exhausted or deleveraged.

• Hedge fund deleveraging was also analyzed as having progressed to about the 50% level.

• UBS assessed that semiconductor and software long positions had returned to April levels.

• It is also important that hedge funds significantly reduced their long positions in momentum stocks and semiconductor stocks on a total market value basis.

This does not mean the market has become completely safe.

But the forced liquidation pressure that caused the sharp drop may have eased substantially.

So this rebound can be seen more as a recovery from supply-demand pressure relief than as a simple technical bounce.

5. Variables that will determine the future direction of semiconductors: Google earnings and hyperscaler CAPEX

The event semiconductor investors should focus on most right now is the earnings of big tech firms like Alphabet, Microsoft, Amazon, and Meta.

In particular, the core point is capital expenditure, or CAPEX, rather than revenue or EPS.

Building AI data centers, securing GPUs and TPUs, and expanding power infrastructure require massive CAPEX.

Therefore, if hyperscalers say they will continue increasing AI infrastructure investment, the semiconductor rally can gain strength again.

Conversely, if CAPEX comes in below Wall Street expectations, the market may immediately become skeptical.

As concerns arise that “AI investment may have already peaked,” semiconductor stocks could face another correction.

In Alphabet’s case, the original text mentioned that a computing capacity shortage issue still exists.

It also raised concerns about Google’s AI competitiveness in relation to Gemini model performance and launch delays.

If Google is indeed facing a computing power shortage, that could actually become grounds for future CAPEX expansion.

6. Semiconductor sector flow: The names that had fallen more rebounded harder

In this rebound, later-stage semiconductors and surrounding infrastructure companies that had fallen more moved more strongly than the large leading names.

• Memory and storage: They rebounded quickly because the recent correction had been severe.

• Micron: A positive view from Bank of America highlighted its undervaluation appeal.

• SK Hynix: Increased call option trading and hedging-related net buying were mentioned, and some short-covering is also possible.

• Intel, AMD, ARM: CPU-related stocks also joined the semiconductor rebound.

• TSMC: It rebounded amid mention of possible future price increases.

• Semiconductor equipment stocks: They generally showed a solid trend amid expectations that AI infrastructure investment will continue.

The important point here is that buying was not confined to a single stock like Nvidia.

Interest is spreading again across the entire AI hardware value chain.

7. Neoclouds and data center power: Hidden beneficiaries of AI infrastructure investment

Another area that stood out in this move is neoclouds and data center power.

Nebius rebounded sharply amid reports of Nvidia increasing its stake.

CoreWeave also posted a strong rebound after recent declines.

These companies are classified as infrastructure firms that can directly benefit as AI computing demand increases.

In the data center power sector, Bloom Energy drew attention.

The original text included that JPMorgan raised its target price.

The key point is the SOFC fuel cell solution.

Gas turbines can take years just to secure supply, and nuclear power requires even more time.

By contrast, some fuel cell solutions stand out because they can secure power supply in a relatively short period of time.

AI data centers are no longer an industry where chips alone are enough.

Power, cooling, optical communications, networks, server racks, and storage can all become bottlenecks.

That is why data center power and optical communication equipment are areas that must continue to be watched in the AI infrastructure investment cycle.

8. Why optical communications and network equipment should be viewed again

Optical communication equipment stocks such as Lumentum and Coherent had seen large declines from their highs recently, but they showed a strong recovery in this rebound.

Inside AI data centers, GPUs are not the only important element.

Optical communications, switches, and network equipment are essential for connecting countless GPUs and exchanging data quickly.

As AI models grow larger, bottlenecks can shift from compute chips to networks and power.

In this earnings season, you should look not only at semiconductors themselves, but also at orders and guidance for optical communication equipment, network equipment, and semiconductor test equipment.

If demand remains healthy in this area, it can be interpreted as a sign that AI infrastructure investment is still expanding.

9. Macro risk: The Red Sea and oil are shaking the market again

The mood improved if you only look at semiconductors, but in the macro area, burdens are increasing.

The original text mentioned that due to the Houthi rebel issue in Yemen, oil tankers avoided the Bab el-Mandeb Strait and took detours.

In that case, transportation distance increases significantly and logistics and insurance costs rise.

As a result, the trend showed Brent crude and WTI prices rising.

What matters here is that international oil prices are not just an energy sector issue.

When oil rises, transportation costs, raw material costs, and production costs all rise together.

This then feeds back into inflationary pressure.

If inflation does not come under control, the central bank will find it difficult to rush into rate cuts.

10. A formula you must remember: Oil → inflation → rates

The easiest formula for understanding the market right now is “oil → inflation → rates.”

When international oil prices rise, inflation is stimulated again.

When inflation rises, expectations for U.S. rate cuts weaken.

When rates rise, growth stocks and technology valuations come under pressure.

The original text mentioned that the U.S. 10-year Treasury yield rose to around 4.63%.

If it breaks the previous high and approaches 5%, the stock market could face greater pressure.

In particular, if changes in the yield curve between short and long maturities also appear, the market may worry at the same time about both economic slowdown and inflation pressure.

11. Two Hs to watch going forward: Hormuz and Hyperscaler

When looking at the market ahead, it helps to remember two Hs.

• The first H is Hormuz.

  Geopolitical risks around the Strait of Hormuz and the Red Sea–Bab el-Mandeb area can pressure oil prices.

  When energy prices swing, inflation and interest rates move together, so they have a direct impact on equities.

• The second H is Hyperscaler.

  How much hyperscalers like Google, Microsoft, Amazon, and Meta increase AI infrastructure investment and data center CAPEX is the core of the semiconductor rally.

  If their CAPEX exceeds expectations, strong momentum can return to AI semiconductors and the data center value chain.

12. The core point that is often missed in other news: The rebound is driven more by “position normalization” than by “good news”

Most market commentary explains it simply as “semiconductors fell a lot, so they rebounded.”

But more important is how much investor positioning has been cleaned up.

The market does not move only on fundamentals.

When supply-demand becomes excessively one-sided, even good companies can fall sharply.

Conversely, once leverage liquidation ends, stocks can rebound strongly even without a special positive catalyst.

This rebound has more to do with the change in supply-demand, specifically that forced selling pressure has eased, than with the structural growth story of the semiconductor industry.

So now, instead of blindly chasing the move, it is necessary to check CAPEX, interest rates, and oil.

Only if all three move favorably at the same time is the semiconductor rally more likely to last longer.

13. Investment strategy: Volatility management comes before leverage

The most realistic advice in the original text is to lower portfolio volatility.

Even if semiconductor fundamentals are good, it is hard to make sound decisions if you are holding volatility that you cannot withstand.

In particular, 2x leveraged ETFs or heavy concentration in a single stock can boost short-term returns, but they can also severely shake your mentality and account fluctuations.

On days of sharp gains, FOMO can easily set in.

But a one-day rally does not suddenly improve a company’s fundamentals.

Conversely, a one-day drop does not mean the industry’s structural growth is over.

What matters is not the price, but whether the investment basis you set still holds.

• If you have no cash, you can consider adjusting some exposure during rebound phases.

• If your portfolio volatility is too high, you should recheck the number of stocks and position sizes.

• Leverage increases profits in bull markets, but in corrections it can undermine judgment.

• If you are investing for 20 or 30 years, survival-oriented habits matter more than short-term returns.

14. If you summarize the current market in one sentence

The semiconductor fundamentals are still alive, but for the supply-demand rebound to turn into a new upward cycle, big tech CAPEX and U.S. rate stability must be confirmed.

The AI industry continues to grow, and bottlenecks in data center power and optical communications are creating new investment opportunities.

However, in periods when international oil prices and interest rates jump again, even the best growth stocks can become more volatile.

Right now, confirmation matters more than aggression.

In particular, we need to see whether Google’s earnings show signs of expanded AI infrastructure investment, and whether the U.S. 10-year yield continues to rise.

< Summary >

The core point of the semiconductor rebound is the easing of leverage ETF and hedge fund deleveraging rather than an improvement in fundamentals.

Rising Korean semiconductor exports and demand for AI infrastructure investment remain positive.

If CAPEX expansion is confirmed in big tech earnings, including Google, the semiconductor rally could continue.

Conversely, if CAPEX comes in below expectations, it may end up being only a dead cat bounce.

Rising international oil prices are a market risk because they feed into inflation and U.S. rate pressure.

Going forward, the most important things to watch are Hormuz and Hyperscaler, that is, geopolitical oil risks and hyperscaler investment plans.

The investment strategy should prioritize portfolio volatility management and cash allocation adjustments over increasing leverage.

[Related Articles…]

• Semiconductor Supercycle and AI Infrastructure Investment Points
• How U.S. Interest Rates and International Oil Prices Affect Stocks