Why the Nasdaq Fell: Semiconductor Selloff, AI CapEx Payback Risk, and Korea Market Impact [EN]

* The original Korean post is available here. -> Korean Version

Semiconductors were shaken again. Oil moved lower. Manufacturing cooled, but did not collapse. Meta opened the question of turning AI costs into revenue. The market did not abandon AI. It recalculated the price attached to AI and the time needed to recover that investment.
— System View Daily Market Framework

 

[System View Quick Take]

U.S. stocks slipped on July 1.
The Dow fell 0.03%, the S&P 500 declined 0.22%, and the Nasdaq lost 0.66%.
The Philadelphia Semiconductor Index plunged 6.3%, and technology was the largest burden among S&P 500 sectors.
By contrast, Meta rose 8.8% after reports that it could monetize excess AI computing capacity through a cloud business.
ISM manufacturing PMI slowed to 53.3, but remained in expansion territory.
The prices index fell from 82.1 to 73.0, but cost pressure in electronic components, memory, and semiconductors remained.
Brent fell to $71.57, while WTI dropped to $68.58.
But the U.S. 10-year Treasury yield rose to 4.46% ahead of the employment report.
Today’s core issue is not simply “weak technology stocks.”
The core issue is a simultaneous repricing of AI payback risk, semiconductor valuations, and rate caution — not AI demand itself.

1. U.S. Market Summary: Indexes Were Weak, but Semiconductors Were Much Weaker

The headline decline was small, but internal pressure was concentrated in semiconductors

U.S. stocks slipped on Wednesday, July 1, 2026. The Dow Jones Industrial Average fell 13.96 points, or 0.03%, to close at 52,305.24. The S&P 500 declined 16.13 points, or 0.22%, to 7,483.23. The Nasdaq Composite dropped 173.69 points, or 0.66%, to close at 26,040.03.

At the index level, this was not a major correction. But the internal structure was different. The Philadelphia Semiconductor Index fell 6.3%. The market did not collapse broadly. Instead, the high price attached to AI and semiconductors was shaken again.

This decline is difficult to explain as simple profit-taking. During the previous quarter, the S&P 500 and Nasdaq rose strongly, led by AI and semiconductors. But on the first trading day of July, the market started reviewing that price. This does not mean AI demand has disappeared. More precisely, the market reconsidered how quickly AI demand can turn into earnings and whether the cost structure can be recovered.

Meta’s move showed that distinction. Meta rose sharply after reports that it could sell excess AI computing capacity through a cloud business. The market did not treat AI investment only as a cost. But it started asking more forcefully whether that CapEx can actually be converted into revenue.

[System View Market Brief] U.S. Market Close on July 1

Asset / Index	Close / Move	System View Interpretation
Dow Jones	52,305.24 -0.03%	The index approached an intraday high but slipped slightly by the close. Large-cap technology pressure weighed on the broader market.
S&P 500	7,483.23 -0.22%	Technology and semiconductors led the decline. Sector pressure was larger than the headline index move suggested.
Nasdaq	26,040.03 -0.66%	AI and semiconductor position adjustment directly pressured the growth-heavy index.
Russell 2000	3,012.59 -0.39%	Small caps also weakened. But the core pressure of the session was in semiconductors, not small caps.
Philadelphia Semiconductor Index	-6.3%	This looked more like a repricing of semiconductor valuations and AI payback structures than a collapse in AI demand.
Meta	+8.8%	The market priced the possibility that AI infrastructure could become a revenue source, not just a cost line.
VIX	16 range	Index-level fear was not severe. Risk was concentrated inside semiconductors rather than across the entire market.

2. Today’s Core Variable: Semiconductor Selloff and Meta’s Opposite Move

The market questioned AI investment payback more than AI demand itself

Today’s core variable was the semiconductor selloff. The Philadelphia Semiconductor Index dropped 6.3%. This cannot be summarized simply as weak technology stocks. It was another price reset in the key sector that led the AI rally.

The surface explanation for semiconductor weakness is valuation pressure. But stopping there is insufficient. What the market started reviewing again was the cost structure of AI investment. AI data centers, HBM, memory, networking equipment, and power infrastructure are revenue for semiconductor companies. But for Big Tech, they are CapEx, depreciation, electricity costs, cooling costs, and debt-funding costs.

Meta’s sharp rise connects directly to this structure. Reports that Meta could sell excess AI computing capacity through a cloud business created the interpretation that AI costs could be recovered as revenue. The market did not reject AI spending itself. It is now demanding clearer proof of when and how AI CapEx can be monetized.

This distinction matters. The same AI news can produce different stock reactions. If AI investment remains only a cost, the stock is pressured. If AI infrastructure can be converted into revenue, the stock can hold up. Today’s market priced exactly that difference.

[System View Core Line]

The market did not abandon AI.
It recalculated the price attached to AI.
For semiconductors, AI is revenue.
For Big Tech, AI is cost.
Today, the market separated stocks by whether that cost can be converted back into revenue.

3. Causal Chain: Why the Nasdaq Fell While Meta Rose

The difference is whether AI CapEx remains a cost or becomes revenue

The causal structure of this session is not simple. The Nasdaq fell. Semiconductors plunged. But Meta rose sharply. These moves are not contradictory. They show that the market is starting to differentiate AI more precisely.

First, valuation pressure increased after the semiconductor rally. During the previous quarter, AI and semiconductor stocks received strong flows. Prices rose first, and the market now wants earnings and cash flow that can justify those prices.

Second, ISM manufacturing remained in expansion, but slowed. Manufacturing PMI fell to 53.3. Orders remained in expansion territory, but momentum slowed. The prices index came down, but remained elevated. In other words, growth and cost conditions were not fully comfortable at the same time.

Third, oil moved lower. This was favorable for inflation. But lower oil alone could not remove overvaluation pressure in semiconductors. The market separated energy-inflation relief from AI valuation repricing.

Fourth, Meta received support from the possibility that AI infrastructure could be turned into external revenue. This created differentiation within the AI theme. The market began assigning different prices to companies that use AI, companies that sell AI infrastructure, and companies that can monetize AI infrastructure externally.

[Causal Chain] Why U.S. Stocks Fell

Stage	Event	Market Interpretation
Stage 1	High prices formed after the Q2 AI and semiconductor rally	Expectations were already high. Even a small concern could trigger a large price adjustment.
Stage 2	SOX plunged 6.3%	The market priced semiconductor valuation pressure and position adjustment before any collapse in AI demand.
Stage 3	Nasdaq declined	Semiconductor weakness directly spilled into the large-cap growth index.
Stage 4	Meta rose 8.8%	Expectations that AI infrastructure costs could be recovered through cloud revenue partially cushioned the broader AI narrative.
Stage 5	Lower oil, Warsh comments, and employment-report wait	Inflation pressure eased in part, but rates and labor confirmation still limited the index upside.

4. Data Check: Manufacturing Expanded, Costs Slowed, but Price Pressure Remained

ISM showed moderate cooling rather than economic collapse

The key data release was ISM manufacturing PMI. The June manufacturing PMI came in at 53.3. It was lower than May’s 54.0, but remained above 50. Manufacturing slowed, but stayed in expansion territory.

New orders fell to 56.0. That remains expansionary, but the pace softened. The production index also slowed to 52.2. In other words, manufacturing is not shutting down. It is moving from a stronger pace into a cooler phase.

The prices index fell to 73.0. That was lower than May’s 82.1, showing the impact of oil stabilization and supply-chain improvement. But 73.0 is still high. Cost pressure has not disappeared. The rate of increase merely slowed.

Electronic components, memory, and semiconductors were still cited as items facing supply shortages. This connects directly to AI. AI investment supports manufacturing demand, but it also pushes up the cost of electronic components and memory. AI is therefore both a growth driver and a cost driver.

[Macro Data Check] June ISM Manufacturing Indicators

Indicator	Result	Market Interpretation
ISM Manufacturing PMI	53.3 May: 54.0	Manufacturing slowed, but remained in expansion. This is not an economic-collapse signal.
New Orders	56.0 May: 56.8	Demand remains alive, but momentum slowed. Inventory accumulation and advance-order effects are fading.
Production Index	52.2 May: 54.3	Production momentum weakened, but it has not entered contraction.
Prices Index	73.0 May: 82.1	Price pressure eased, but remains high. This is not enough to fully comfort the Fed.
Employment Index	49.7 May: 48.6	Still in contraction, but improved. Before the employment report, this was not decisive enough for the market.

5. Rates, Dollar, Oil, and Bonds

Oil moved lower, but yields held firm ahead of the employment report

Oil declined. Brent fell to $71.57, while WTI dropped to $68.58. Expectations of progress in U.S.-Iran talks and easing Strait of Hormuz supply concerns reduced the energy-risk premium.

Lower oil is generally supportive for equities. It reduces headline inflation pressure and eases cost pressure for consumers and companies. But today, lower oil did not prevent the semiconductor correction. The market separated energy-inflation relief from AI valuation repricing.

The U.S. 10-year Treasury yield rose to around 4.46%. Selling pressure appeared in bonds ahead of Thursday’s employment report. The manufacturing prices index moved lower, but it remained elevated, and the Fed continued to signal that it would not easily abandon the 2% inflation target.

The dollar is also in a range where it is difficult to weaken decisively. Lower oil can ease some dollar-supportive pressure, but U.S. yields and the wait for the employment report limit dollar downside. In particular, yen weakness and global rate differentials continue to support dollar demand.

[Cross Asset Check] Signals From Oil, Rates, and the Dollar

Variable	Move	Interpretation
Brent	$71.57	Easing Hormuz supply concerns and negotiation expectations reduced the oil-risk premium.
WTI	$68.58	A price below $70 is favorable for energy-inflation pressure.
U.S. 10-Year Yield	4.46%	The employment-report wait and Fed caution limited downside in yields.
Dollar	Mild strength pressure	Even with lower oil, U.S. yields and labor-data caution limited dollar weakness.
Gold	Higher	Oil moved lower, but geopolitics and rate caution did not fully disappear.

6. AI and Semiconductors: This Was Price Repricing, Not Demand Collapse

For semiconductors, AI is revenue. For Big Tech, AI is cost. For Meta, it may become recoverable revenue.

Interpreting today’s semiconductor selloff as an AI demand collapse would be an exaggeration. AI data-center investment, memory demand, and power-infrastructure demand remain alive. The problem is not demand. The problem is price. The market had already attached a high price to AI demand, and it is now asking whether that price is justified.

For semiconductor companies, AI investment is revenue. GPUs, HBM, DRAM, networking chips, power semiconductors, and server components are direct beneficiaries of AI CapEx. But for Big Tech, the same investment is cost. They must buy servers, build data centers, consume power, and absorb depreciation.

That is why the market no longer treats all AI-related stocks as one group. It separates companies that sell AI infrastructure, companies that use AI infrastructure, and companies that can resell AI infrastructure as external revenue. Meta’s rise showed that distinction.

If Meta can sell excess AI computing capacity as a cloud service, AI CapEx becomes not only a cost but also a potential revenue source. But not every Big Tech company can create the same structure. Going forward, the market is likely to focus more on AI investment payback than on the size of AI investment itself.

[AI and Semiconductor Judgment]

The SOX plunge was not an AI demand collapse.
More precisely, it was a repricing of the valuation attached to AI demand.
For semiconductors, AI is revenue.
For Big Tech, AI is cost.
For Meta, the market attached the possibility that AI costs can be turned into revenue.
Today, the market separated those three structures again.

7. Judgment: The Market Did Not Abandon AI. It Reduced the Price of AI.

Conditional risk appetite remained, but semiconductor-specific risk increased

The essence of today’s market is not a weak index. More precisely, it is a price reset inside AI and semiconductors. The Dow and S&P 500 fell only slightly. The Nasdaq was weaker. The SOX plunged. That gap is the key.

The market did not abandon all risk assets. Oil moved lower, and manufacturing remained in expansion. Meta rose on expectations of AI cloud monetization. This combination means risk appetite did not fully break.

But caution appeared in semiconductors. Even if AI demand remains alive, prices can correct when valuations are too high. Even if AI CapEx keeps rising, the market reduces multiples if that spending cannot be recovered through revenue and cash flow.

The Fed variable remains as well. The manufacturing prices index fell, but remained high. The 10-year yield rose ahead of the employment report. The market liked lower oil, but it did not restore an unconditional premium to technology stocks before checking rates and labor data.

Therefore, the core sentence of today’s Daily is this.

The market did not abandon AI. It repriced AI CapEx payback risk and semiconductor valuations.

[System View Judgment]

U.S. stocks fell slightly today.
But semiconductors were shaken severely.
This was less about broad risk-off and more about valuation repricing inside AI and semiconductors.
Lower oil eased the inflation burden, but it did not prevent semiconductor valuation compression.
Meta’s rise shows that the market has not abandoned AI.
But the market is now asking more strictly whether AI investment can be recovered, not just how large that investment is.
The next checkpoints are the employment report, the U.S. 10-year yield, whether the SOX rebounds, and AI CapEx guidance from major companies.

8. Impact on Korea: Semiconductor Flows Face Pressure, While Lower Oil Provides a Buffer

Korea should focus first on the SOX plunge and AI-cost repricing, not the small U.S. index decline

The core issue was not simply the decline in U.S. equities. The index-level decline was limited, but the Philadelphia Semiconductor Index moved sharply lower. Therefore, for Korea, the first thing to watch is not the Dow or S&P 500 decline. The key is how the U.S. semiconductor selloff spills over into Korean memory leaders and foreign flows.

Korean semiconductors face two signals at the same time. First, the SOX plunge is a short-term flow burden. The high price attached to AI and semiconductor rallies was adjusted again, and this is psychologically negative for Samsung Electronics and SK Hynix. If foreign investors reduced U.S. semiconductor exposure, some risk reduction could also appear in Korean large-cap semiconductors.

Second, memory demand itself has not disappeared. AI data centers, HBM, server DRAM, and high-performance memory remain part of the structural case for Korean semiconductors. The problem is price, not demand. The market is now asking how long the HBM and AI memory premium can last, and whether Big Tech’s AI CapEx can actually be recovered as revenue.

Lower oil is a buffer for Korea. If Brent and WTI fall, Korea’s energy-import burden, airline and transportation costs, and some chemical input costs decline. This is also basically favorable for the won. But the rise in the U.S. 10-year yield is a burden. Even if lower oil helps the won, higher U.S. yields and dollar-strength pressure can limit foreign equity flows.

Therefore, Korea’s market can be summarized in one line. Semiconductors are affected by price repricing, not demand collapse. Lower oil reduces cost pressure. Higher rates cap risk assets. These three forces are operating at the same time.

[Korea Market Impact] Implications for Korea

Korean Market Variable	Direction	System View Interpretation
Large-Cap Semiconductors	Short-term flow burden	The SOX plunge is a burden for foreign flows into Samsung Electronics and SK Hynix. But it is too early to call this an AI memory-demand collapse.
KOSPI	Selective flow	Because of Korea’s semiconductor weight, index upside may be capped. Lower oil and manufacturing expansion can partly buffer downside.
KOSDAQ and Growth Stocks	Pressure	Nasdaq weakness and higher U.S. 10-year yields translate into discount-rate pressure for high-multiple growth stocks.
USD/KRW	Mixed	Lower oil is favorable for the won. But higher U.S. yields and the employment-report wait limit dollar downside.
Airlines and Transportation	Supportive	WTI below $70 strengthens expectations of fuel-cost relief.
Refining and Energy	Pressure	Lower oil is a short-term burden for energy-sensitive names. Renewed geopolitical risk will decide their defensiveness.
Autos, Shipbuilding, Defense	Sector differentiation	Sector reactions may diverge depending on oil and FX direction. If semiconductor flows weaken, index leadership becomes limited.

[Korea Market Core Judgment]

Korea should watch the SOX plunge before the small U.S. index decline.
Semiconductors are being hit by price and positioning repricing, not confirmed demand collapse.
Lower oil is favorable for Korea’s cost structure.
But higher U.S. yields and the employment-report wait can limit foreign flows.
Today’s core variables are foreign semiconductor flows, won stability, and the U.S. 10-year yield.

9. Today’s Checkpoints

Semiconductor flows, U.S. yields, and stable oil all matter at the same time

The first checkpoint is foreign flow into large-cap semiconductors. Because the U.S. SOX fell sharply, Korea needs to watch Samsung Electronics and SK Hynix first. The key is not simply whether their share prices fall. The key is whether foreign investors actually reduce memory leaders, or whether they defend them after a short-term correction.

The second checkpoint is the U.S. 10-year yield. The fact that yields rose despite lower oil means the market has not abandoned employment-report and Fed caution. If yields rise further, the rebound potential for the Nasdaq and growth stocks becomes limited. This is also a burden for Korea’s KOSDAQ and high-multiple growth stocks.

The third checkpoint is oil. Lower Brent and WTI are favorable for Korea from a cost perspective. Airlines, transportation, some consumer sectors, and chemicals can receive short-term support. But if lower oil starts to be interpreted as a growth-slowdown signal, materials and energy sectors can face greater pressure.

The fourth checkpoint is USD/KRW. Lower oil is favorable for the won. But if higher U.S. yields and dollar strength appear at the same time, the benefit from lower oil weakens. Foreign flows are sensitive to the won’s direction.

The fifth checkpoint is AI CapEx news. Meta benefited from expectations that AI computing capacity could become cloud revenue. Semiconductors, however, plunged. The market now cares more about AI investment payback than AI investment size. Korean semiconductors are not free from that standard.

[Today’s Checkpoints] Reference Lines to Watch

Checkpoint	Reference	System View Interpretation
Samsung Electronics and SK Hynix	Foreign net buying / selling	Check whether the SOX plunge spills into Korean memory-leader flows.
U.S. 10-Year Yield	Whether it holds in the mid-4.4% range	If yields move higher, rebound potential for the Nasdaq and Korean growth stocks remains limited.
Brent and WTI	WTI below $70	Stable oil is favorable for Korea’s cost structure and the won.
USD/KRW	Won stability	Watch whether the lower-oil benefit supports the won or is offset by higher U.S. yields.
AI CapEx News	Monetization potential	AI payback now matters more than AI spending size. Meta and semiconductors showed opposite reactions.
U.S. Employment Report	Payrolls, wages, unemployment rate	Strong jobs raise rate pressure. Weak jobs raise slowdown concerns. The market wants a middle outcome.

10. Secondary Issue: Korea’s Semiconductor Export Momentum Collides With U.S. Semiconductor Repricing

Fundamentals are strong, but much of that strength has already been priced

Looking at Korean semiconductors only through the U.S. SOX plunge would be insufficient. Korean semiconductor exports and AI memory demand remain strong pillars. HBM, server DRAM, and high-performance memory still support the earnings logic for Korea’s large semiconductor companies.

But stock prices do not move on fundamentals alone. Prices move based on the gap between expectations already reflected and expectations newly added. Even if AI memory demand is strong, additional upside requires stronger evidence if the market has already priced much of that strength.

This is where the U.S. SOX plunge matters. The U.S. market is not repricing AI demand itself as much as the price attached to AI, Big Tech’s AI CapEx payback risk, and semiconductor valuations. That repricing can spill over into Korean semiconductors.

Therefore, today’s key point for Korean semiconductors is not just whether the stocks rise or fall. The important question is whether foreign investors sell Samsung Electronics and SK Hynix indiscriminately, or differentiate them based on HBM exposure and memory pricing power. Indiscriminate selling would mean position liquidation. Differentiation would mean the structural demand story still remains inside the market.

[Secondary Issue Judgment]

Korea’s semiconductor fundamentals are difficult to call broken at this stage.
The problem is price, not demand.
The U.S. market is recalculating AI CapEx payback risk and semiconductor valuations.
In Korea, the key is whether foreign investors reduce memory leaders indiscriminately, or defend them because of HBM and server DRAM logic.

11. Conclusion Summary

The essence of today’s Korean market is the collision between semiconductor repricing and lower-oil relief

U.S. stocks slipped on July 1. The headline index decline was small. But semiconductors were different. The Philadelphia Semiconductor Index fell sharply, and the high price attached to AI and semiconductors was adjusted again.

This decline is difficult to read as an AI demand collapse. Meta’s strong gain on expectations that AI computing capacity could be turned into cloud revenue showed that the market has not abandoned AI. Instead, the market separated AI investment payback potential by company.

For Korea, semiconductor flow pressure is likely to appear first. Samsung Electronics and SK Hynix may be affected by the U.S. SOX plunge. But memory demand and the HBM premium have not disappeared. The key is whether foreign flows are indiscriminate selling or selective adjustment.

Lower oil is a buffer for Korea. If WTI stays below $70, Korea’s energy-import burden and airline and transportation costs are supported. But higher U.S. 10-year yields and the wait for the employment report can pressure growth stocks and the won.

Conclusion: Today’s Korean market is likely to reflect semiconductor repricing more than the small U.S. index decline. Lower oil is a buffer for cost structures and the won. The key variables are foreign semiconductor flows, the U.S. 10-year yield, and USD/KRW.

12. Key Questions

Is the U.S. semiconductor plunge immediately negative for Korean semiconductors?

It is a short-term flow burden. But it is too early to call it an AI demand collapse. The key is whether foreign investors sell Samsung Electronics and SK Hynix indiscriminately, or defend them based on memory pricing power.

Is the AI trade over?

No. Meta’s rise shows that the market has not abandoned AI. But the market is now asking more strictly about AI investment payback, not just AI investment size.

Is lower oil positive for Korea?

Basically, yes. Lower oil can reduce energy-import costs, airline and transportation costs, and import-price pressure. But if lower oil starts to be interpreted as a growth-slowdown signal, materials and energy stocks can face pressure.

Why does USD/KRW matter?

Lower oil is favorable for the won, but higher U.S. yields support the dollar. The won needs to stabilize for foreign flows to work more favorably in Korean equities.

What should Korean investors watch first today?

Foreign flows into Samsung Electronics and SK Hynix, USD/KRW, the U.S. 10-year yield, and whether WTI stays below $70. Semiconductor flows and rates need to be watched together.

13. Related System View Reports

[Related System View Reports]

U.S. Semiconductor Selloff, Nasdaq Weakness, and Korea Market Impact
Micron Earnings, AI Memory, and Korea Semiconductor Impact
U.S. PCE, Apple Price Hikes, and Memory Price Impact on Korean Semiconductors
U.S. Long-Term Yields and Nasdaq Duration Risk
Oil, Yields, and AI Rally Risk

14. Sources and References

[Sources and References]

Reuters, “Wall Street ends lower as tech shares slip,” July 1, 2026.
Reuters, “U.S. manufacturing activity eases in June; prices paid by factories remain elevated,” July 1, 2026.
Reuters, “Oil prices fall 1% to 4-month lows as progress in US-Iran talks cools supply concerns,” July 1, 2026.
MarketWatch, “Investors sell 10-year Treasury, yield edges up ahead of jobs report,” July 1, 2026.
Reuters, “Morning Bid: Kicking off H2,” July 1, 2026.
Institute for Supply Management, Manufacturing PMI, June 2026.
Federal Reserve, public calendar and policy materials.

15. Disclaimer

This article is a macroeconomic and market interpretation based on publicly available data and market reports. It is not a recommendation to buy or sell any specific stock, ETF, bond, commodity, derivative, or financial product. All investment decisions and outcomes are the sole responsibility of the investor. Market data and forecasts are based on information available at the time of writing and may change depending on macroeconomic conditions, interest rates, oil prices, geopolitical variables, corporate earnings, semiconductor supply and demand, and exchange-rate movements.