Is AI Semiconductor Investing Really Over? Crash History, SOXX vs SMH & the Next Portfolio Strategy

Navy and gold editorial illustration of an AI semiconductor chip with a market pullback and recovery chart
DAYHOOK INVESTING | 10-SECOND BRIEF

Is the AI chip trade over?

Statistical claim

16 of 17 rebounds is a source claim, not a verified backtest
ETF structure

SOXX caps weights; SMH gives its largest winners greater influence
Korean memory

Evaluate estimate revisions and positioning—not headlines alone
Portfolio response

Stage entries and balance hardware with applications

Takeaway: A drawdown is not proof that AI demand has ended; distinguish earnings evidence from price momentum.

Is AI semiconductor investing over? A seven-session slide in the iShares Semiconductor ETF (SOXX), as described in the supplied market commentary, has reignited a familiar debate: is this the end of the AI hardware cycle or an opportunity to rebalance? This article tests the logic rather than assuming either answer.

The case involves three separate questions: whether historical losing streaks predict rebounds, whether Korean memory stocks remain attractive relative to alternatives, and how much concentration investors should accept in semiconductor ETFs.

The seven-day semiconductor sell-off: signal or noise

Seven consecutive down days, if confirmed using adjusted daily closes and a clearly specified cutoff, describe price action rather than a change in the underlying cash flows of chipmakers. The distinction matters: earnings estimates can improve even while multiples contract, and an AI capex boom can coexist with severe stock-price corrections.

  • Check whether the decline is concentrated in GPU leaders, memory, equipment or the whole sector.
  • Compare earnings revisions and guidance with the ETF’s price move.
  • Watch funding costs, export restrictions, AI capex returns and concentration risk.

What the claimed 16-year rebound study actually tells us

The source commentary reports 17 episodes of at least six consecutive SOXX down days from 2010–2026, with 16 positive three-month outcomes (94.1%) and a median three-month return of +11.4%. It also cites an average additional decline of 4.6% and a median maximum drawdown of 9.3%. These figures have not been independently reproduced here; they should be treated as the commentary’s backtest claims, not established probabilities.

REPORTED BACKTEST · NOT INDEPENDENTLY VERIFIED
17

claimed historical streaks
94.1%

claimed 3-month positive frequency
+11.4%

claimed median 3-month return
9.3%

claimed median maximum further decline

Methodology, event dates, dividend adjustments and overlap treatment were not provided.

Before using this as a trading signal, an investor should reconstruct each event using adjusted SOXX daily data, avoid counting overlapping streaks as independent samples, define the three-month holding period consistently and compare results with unconditional three-month returns. Even 16 successes in 17 observations cannot guarantee the next outcome.

Samsung Electronics and SK hynix: reassessing momentum

The supplied commentary describes a decision to exit Samsung Electronics and SK hynix after weaker price momentum, institutional selling and concern about earnings revisions. That is a portfolio manager’s decision, not a universal instruction to sell either stock.

One cited Goldman Sachs Samsung forecast change is stated as “112 trillion won to 106 trillion won” for a single quarter. That unit/period appears questionable and has not been verified, so it is intentionally excluded as a factual earnings forecast. Readers should check the original analyst note and Samsung’s official quarterly disclosures before relying on it.

  • Price momentum: confirm moving-average levels using the same date and adjusted chart series.
  • Earnings momentum: distinguish a one-off estimate revision from sustained HBM, DRAM and foundry earnings trends.
  • Relative evaluation: ask whether a fresh investment today would beat alternatives after taxes, fees and portfolio constraints.

Practical test: “If I held cash today, would I initiate this position at the current price?” It is a useful discipline, but existing positions also require considering tax, diversification and rebalancing costs.

AI semiconductor chip with a market pullback and rebound illustration
Original DayHook Investing illustration. Market path is conceptual, not a price forecast.

SOXX vs SMH: diversification versus concentration

Both ETFs offer diversified semiconductor exposure, but they differ materially in concentration. SOXX follows a capped float-adjusted market-cap methodology. BlackRock’s published methodology highlights an 8% cap for the five largest names and 4% for other names in its reference description, with changes scheduled for September 2026; current index documents should be consulted before assuming fixed weights. iShares methodology supplement.

SMH is more top-heavy. VanEck’s October 31, 2026 fact sheet reported 26 holdings and an approximately 20.19% Nvidia weight at that date. Concentration can boost returns when a leader outperforms and deepen losses when it reverses. VanEck SMH fact sheet.

SOXX

Capped exposure

Designed to limit the influence of the largest holdings.

Trade-off: less single-name concentration, but still highly cyclical.

SMH

Leader-heavy exposure

Greater sensitivity to dominant chip companies such as Nvidia.

Trade-off: potential upside from winners, plus higher concentration risk.

Neither fund is automatically “safer”: factor exposures, index changes, fees, country listings and historical returns must be compared over the same measurement window.

AI infrastructure chips, HBM and networking compared with cloud, enterprise AI and search applications
Original DayHook Investing value-chain illustration; not an allocation recommendation.

The two-track AI portfolio: infrastructure and applications

A two-track strategy can balance companies selling the compute infrastructure with platforms seeking to monetize AI. Hardware has tangible shipment and pricing drivers but can suffer order-cycle corrections; software platforms have distribution advantages but still need to demonstrate incremental revenue and margins.

AI VALUE CHAIN · TWO TRACKS
Infrastructure

GPUs → HBM → networking → power and cooling

Applications

Cloud → enterprise tools → search → AI agents

Conceptual diagram; not a performance forecast.

  • Conservative scenario: use broad diversification, cash buffers and predetermined position limits.
  • Balanced scenario: phase into both infrastructure and platforms while monitoring revisions.
  • Aggressive scenario: accept higher concentration only with a clear maximum-loss and rebalance plan.

Bitcoin’s 200-day moving average: a separate risk signal

The supplied commentary also claims that Bitcoin recovered its 200-day moving average before major rallies in eight of nine past cases. Without event dates, market data definitions and an out-of-sample test, the 8/9 statistic is unverified. Bitcoin’s moving-average behavior should not be treated as proof of an impending bull market.

Bitcoin differs from semiconductor equities: it has no earnings stream, trades continuously and can experience severe liquidity-driven drawdowns. A 200-day average may be useful as one trend indicator, but sizing and downside risk matter more than any single technical signal.

GLOBAL 5+5+5 THEMATIC SPOTLIGHT · WATCHLIST, NOT A BUY LIST
US

Nvidia

NVDA
Broadcom

AVGO
Micron Technology

MU
Alphabet

GOOGL
Amazon

AMZN
India

Dixon Technologies

DIXON
Kaynes Technology

KAYNES
CG Power

CGPOWER
Larsen & Toubro

LT
Tata Consultancy Services

TCS
Korea

SK hynix

000660
Samsung Electronics

005930
Hanmi Semiconductor

042700
Samsung Electro-Mechanics

009150
LS ELECTRIC

010120

Indian names are indirect exposure to electronics, data centers or AI services; not all are semiconductor producers.

🔴 Bottom Line: Top Picks 15

The following 15 securities form a thematic research watchlist, not a recommendation to buy. The Indian selections include indirect exposure, while US and Korean selections span both hardware and software.

1. Nvidia (NVDA) · US

WHY · AI accelerator leadership and software ecosystem underpin demand

  • Review earnings quality and management guidance.
  • Compare valuation with peers and history.
  • Monitor the next company-specific catalyst.

RISK · Export controls, hyperscaler spending pauses and valuation compression

  • Test a downside case before entering.
  • Watch liquidity, concentration and macro risks.
  • Set a position-size limit before trading.

Links open research searches, not verified live quotes or consensus targets.

2. Broadcom (AVGO) · US

WHY · Custom AI silicon and networking diversify the infrastructure thesis

  • Review earnings quality and management guidance.
  • Compare valuation with peers and history.
  • Monitor the next company-specific catalyst.

RISK · Customer concentration and execution risk on custom-chip programs

  • Test a downside case before entering.
  • Watch liquidity, concentration and macro risks.
  • Set a position-size limit before trading.

Links open research searches, not verified live quotes or consensus targets.

3. Micron Technology (MU) · US

WHY · HBM and DRAM supply discipline could support memory profitability

  • Review earnings quality and management guidance.
  • Compare valuation with peers and history.
  • Monitor the next company-specific catalyst.

RISK · Memory pricing cycles and costly capacity expansion

  • Test a downside case before entering.
  • Watch liquidity, concentration and macro risks.
  • Set a position-size limit before trading.

Links open research searches, not verified live quotes or consensus targets.

4. Alphabet (GOOGL) · US

WHY · Search, cloud and AI distribution offer application monetization

  • Review earnings quality and management guidance.
  • Compare valuation with peers and history.
  • Monitor the next company-specific catalyst.

RISK · AI search economics, regulation and capital intensity

  • Test a downside case before entering.
  • Watch liquidity, concentration and macro risks.
  • Set a position-size limit before trading.

Links open research searches, not verified live quotes or consensus targets.

5. Amazon (AMZN) · US

WHY · AWS AI services and enterprise distribution can monetize inference

  • Review earnings quality and management guidance.
  • Compare valuation with peers and history.
  • Monitor the next company-specific catalyst.

RISK · Cloud competition, high capex and margin uncertainty

  • Test a downside case before entering.
  • Watch liquidity, concentration and macro risks.
  • Set a position-size limit before trading.

Links open research searches, not verified live quotes or consensus targets.

6. Dixon Technologies (DIXON) · India

WHY · Domestic electronics manufacturing creates indirect hardware exposure

  • Review earnings quality and management guidance.
  • Compare valuation with peers and history.
  • Monitor the next company-specific catalyst.

RISK · Valuation and customer-concentration risks; not a pure AI chip producer

  • Test a downside case before entering.
  • Watch liquidity, concentration and macro risks.
  • Set a position-size limit before trading.

Links open research searches, not verified live quotes or consensus targets.

7. Kaynes Technology (KAYNES) · India

WHY · Electronics manufacturing and semiconductor ambitions add optionality

  • Review earnings quality and management guidance.
  • Compare valuation with peers and history.
  • Monitor the next company-specific catalyst.

RISK · Execution, funding and project-timeline uncertainty

  • Test a downside case before entering.
  • Watch liquidity, concentration and macro risks.
  • Set a position-size limit before trading.

Links open research searches, not verified live quotes or consensus targets.

8. CG Power (CGPOWER) · India

WHY · Semiconductor assembly/test ambitions and industrial equipment exposure

  • Review earnings quality and management guidance.
  • Compare valuation with peers and history.
  • Monitor the next company-specific catalyst.

RISK · Long gestation and capex execution; indirect AI exposure

  • Test a downside case before entering.
  • Watch liquidity, concentration and macro risks.
  • Set a position-size limit before trading.

Links open research searches, not verified live quotes or consensus targets.

9. Larsen & Toubro (LT) · India

WHY · Data-center infrastructure and engineering exposure to digital capex

  • Review earnings quality and management guidance.
  • Compare valuation with peers and history.
  • Monitor the next company-specific catalyst.

RISK · Project execution, cyclicality and capital allocation

  • Test a downside case before entering.
  • Watch liquidity, concentration and macro risks.
  • Set a position-size limit before trading.

Links open research searches, not verified live quotes or consensus targets.

10. Tata Consultancy Services (TCS) · India

WHY · Enterprise AI adoption offers software-services monetization exposure

  • Review earnings quality and management guidance.
  • Compare valuation with peers and history.
  • Monitor the next company-specific catalyst.

RISK · AI automation may pressure legacy services pricing

  • Test a downside case before entering.
  • Watch liquidity, concentration and macro risks.
  • Set a position-size limit before trading.

Links open research searches, not verified live quotes or consensus targets.

11. SK hynix (000660) · Korea

WHY · HBM positioning links directly to advanced AI accelerator demand

  • Review earnings quality and management guidance.
  • Compare valuation with peers and history.
  • Monitor the next company-specific catalyst.

RISK · HBM competition, memory cycle and foreign institutional flows

  • Test a downside case before entering.
  • Watch liquidity, concentration and macro risks.
  • Set a position-size limit before trading.

Links open research searches, not verified live quotes or consensus targets.

12. Samsung Electronics (005930) · Korea

WHY · Memory scale, foundry optionality and HBM recovery potential

  • Review earnings quality and management guidance.
  • Compare valuation with peers and history.
  • Monitor the next company-specific catalyst.

RISK · Execution gap in premium HBM and foundry profitability

  • Test a downside case before entering.
  • Watch liquidity, concentration and macro risks.
  • Set a position-size limit before trading.

Links open research searches, not verified live quotes or consensus targets.

13. Hanmi Semiconductor (042700) · Korea

WHY · Advanced packaging equipment is linked to HBM investment

  • Review earnings quality and management guidance.
  • Compare valuation with peers and history.
  • Monitor the next company-specific catalyst.

RISK · Customer concentration and order-cycle volatility

  • Test a downside case before entering.
  • Watch liquidity, concentration and macro risks.
  • Set a position-size limit before trading.

Links open research searches, not verified live quotes or consensus targets.

14. Samsung Electro-Mechanics (009150) · Korea

WHY · High-end substrates and components support compute hardware

  • Review earnings quality and management guidance.
  • Compare valuation with peers and history.
  • Monitor the next company-specific catalyst.

RISK · End-market demand swings and substrate competition

  • Test a downside case before entering.
  • Watch liquidity, concentration and macro risks.
  • Set a position-size limit before trading.

Links open research searches, not verified live quotes or consensus targets.

15. LS ELECTRIC (010120) · Korea

WHY · Grid and power equipment can benefit from data-center buildout

  • Review earnings quality and management guidance.
  • Compare valuation with peers and history.
  • Monitor the next company-specific catalyst.

RISK · Project timing, competition and industrial-cycle exposure

  • Test a downside case before entering.
  • Watch liquidity, concentration and macro risks.
  • Set a position-size limit before trading.

Links open research searches, not verified live quotes or consensus targets.

Sources, limitations and related reading

ETF index and holdings: iShares SOXX index methodology (September 2026); VanEck SMH October 2026 fact sheet. The historical streak and Bitcoin statistics are attributed to the user-provided commentary and are not independently verified. The commentary timestamps were supplied without a primary video transcript or data table.

Related: Will Semiconductor Stocks Bounce Back? Recovery Timelines.

Educational information only; not personalized financial advice. Prices, holdings and analyst expectations change. Check official disclosures and your own risk tolerance.


Discover more from Dayhook Investing

Subscribe to get the latest posts sent to your email.