Highlights

South Korea’s retail investors, once focused on meme coins, have moved en masse into AI-related semiconductor stocks, especially Samsung Electronics and SK Hynix. Upbit trading volumes plunged by about 80% year-on-year, while Samsung and SK Hynix surged several-fold, pushing KOSPI past 6,000. This shift shows a dramatic reallocation of capital from crypto to AI hardware, driven by HBM demand and retail enthusiasm. Government support and new leveraged ETFs amplify the trend, but concentration risk tied to continued AI demand is a clear vulnerability.

Sentiment Analysis

• 
  Overall sentiment: highly positive toward Korean AI semiconductor stocks, reflected in strong retail buying, rapid price gains, and national pride. The market tone is bullish and optimistic, with many retail investors viewing Samsung and SK Hynix as national champions capable of capturing AI-driven growth. The optimism is visible in large inflows, social media focus shifts, and policy moves to encourage domestic investment.
  


• 
  Risk sentiment: mixed to cautious beneath the surface. While enthusiasm is high, analysts and some industry leaders warn of concentration risk and potential reversal if AI capex weakens. This introduces a tempered caution into the narrative.
  


• 
  Technical sentiment indicator:
  
  
  

  
  75%
  

  
  
  
  The progress bar shows strong positive sentiment intensity (75%), though not absolute — reflecting both the bullish momentum and underlying systemic risks.
  

Article Text

South Korea’s retail investment landscape has changed rapidly. Where Kakao Talk groups once buzzed through the night debating which altcoin would list on Upbit or when an airdrop would arrive, conversations now center on HBM utilization rates at SK Hynix and Samsung’s quarterly revenue outlook. This shift is not anecdotal: trading volume at Upbit has dropped roughly 80% from the same period in 2024, signaling a significant decline in retail crypto activity. Meanwhile, retail investors funneled nearly 597.5 billion won into stocks on the first trading day of the new year, with the lion’s share directed at Samsung Electronics and SK Hynix.

Market performance has matched the narrative. Since early 2025, SK Hynix’s share price jumped roughly sixfold and Samsung Electronics almost quadrupled, producing a combined market capitalization of about $1.32 trillion and accounting for roughly 35% of KOSPI’s total market value. That concentration helped push KOSPI past 6,000 for the first time, a thousand-point rise achieved in just 34 days — the fastest such move in decades. South Korea’s market capitalization has risen, overtaking some major European markets and ranking among the world’s largest exchanges, driven primarily by AI-related demand for high-bandwidth memory (HBM).

The technical demand story underpins much of the enthusiasm. Samsung and SK Hynix together control around 80% of global HBM capacity. With AI workloads hungry for bandwidth and specialized memory, the market has tilted into a seller’s environment where pricing power and capacity constraints translate into outsized gains for producers. Policymakers have also moved to support this domestic investment wave: the financial regulator approved plans to list 2x leveraged ETFs tied to Samsung and SK Hynix, intending to retain retail capital that might otherwise flow overseas. Education courses on leveraged ETFs have surged, reflecting retail desire to amplify exposure.

A useful comparison can be drawn with Taiwan, where retail investors long embraced TSMC. Since the introduction of odd-lot trading in 2020, TSMC’s shareholder base has ballooned — over 900,000 odd-lot holders now participate, allowing small investors to own a piece of a global leader with modest sums. For many Taiwanese, TSMC functions as more than an investment; it is an identity marker and a perceived safe harbor in technology-driven growth. TSMC’s dominance in advanced process nodes — over 90% share in cutting-edge foundry capacity — positions it similarly at the center of AI-related supply chains.

Despite parallels in retail behavior, the origins of these loyalties differ. Taiwanese retail accumulation in TSMC grew organically over years: younger investors routinely set up dollar-cost averaging plans and regard TSMC as a long-term core holding, part of retirement and savings strategies. In Korea, the base of retail investors now buying chips largely migrated from a prior crypto-focused cohort. They lived through meme coin frenzies, dramatic market collapses such as Luna and FTX, and a crypto winter, bringing trading experience and appetite for leverage into equities markets. In some cases, blockchain-based trading platforms have been repurposed to offer perpetual contracts on Korean equities, enabling the same high-leverage behavior once aimed at cryptocurrencies.

Both markets, however, share a significant structural vulnerability: heavy concentration of market value in a few AI beneficiaries. If AI capital expenditure slows or memory and chip cycles move into a downtrend, markets in both countries could experience sharp corrections. The warning is not hypothetical. Industry leaders have cautioned that if AI investments fail to deliver returns, the sector could rapidly reprice, and companies with stretched valuations would be susceptible. Thus, while the current narrative is one of triumph and national pride, it carries an embedded conditionality tied to sustained AI demand.

In short, Korea’s youth appear to have traded the volatility and novelty of crypto for a concentrated bet on national semiconductor champions, blending trading experience with a new sectoral focus. Taiwan’s smaller investors, by contrast, have accumulated exposure to TSMC steadily over time, making their position a long-term, identity-linked holding. Both approaches reflect confidence in AI’s continued ascendancy, but both are also exposed should that premise be challenged. The central question remains whether AI investment will perpetually drive demand — the answer will determine whether these markets thrive or face sharp reversals.

Key Insights Table