The AI Chip Bubble: Why South Korea & Taiwan Are In the Danger Zone | Michael Fritzell | Asian Century Stocks

Episode Summary

Executive Summary: The episode argues that Asia’s equity markets are diverging sharply: Korea and Taiwan are soaring on the AI/memory-chip trade, but much of the move looks speculative and vulnerable to a future supply response from China. Michael Fritzel favors cheaper, governance-improving markets and small caps—especially Korea and parts of Southeast Asia—while warning that China remains policy-risky and that Japan’s earlier rally may be maturing.

Main Topics: Korea’s AI-driven rally and bubble risk (Priority: 5/5): South Korean equities, especially SK Hynix and Samsung Electronics, have surged on memory-chip optimism. Fritzel argues valuations are stretched, the move is speculative, and Chinese supply coming online from 2027 could pressure pricing. Taiwan and broader Asia’s AI speculation (Priority: 4/5): Taiwan’s market strength is attributed to TSMC and AI-related enthusiasm, but the speaker says retail speculation is rising sharply, with violent stock moves and social-media hype signaling a late-cycle atmosphere. Japan’s reform-led bull market (Priority: 4/5): Japan’s rally is framed as driven by a weak yen, governance reforms, buybacks, reduced cross-shareholdings, and improved capital allocation. He believes the run may be nearing exhaustion relative to Korea’s opportunity set. China’s structural and policy headwinds (Priority: 5/5): China is portrayed as a difficult investment case due to property-sector collapse, US tariffs, and unpredictable state intervention across education, brokerage, tech, and lending. The party’s reassertion of control undermines private-sector confidence. Southeast Asia and Korea as value opportunities (Priority: 4/5): Beyond headline tech names, the speaker favors cheap, high-quality stocks in Korea, Thailand, the Philippines, and select infrastructure names. He highlights insider buying, low valuations, and reform tailwinds. Global capital, geopolitics, and trade fragmentation (Priority: 3/5): The discussion links Asian market performance to deglobalization, a forming new Cold War, energy shocks, and shifting capital flows. Singapore, Malaysia, Vietnam, and others are positioned as potential beneficiaries or intermediaries. Consumer brands and cultural exports (Priority: 3/5): Korean cosmetics, food, and entertainment IP are discussed as export themes, but the speaker says finding investable ideas is harder than the cultural story suggests. He cites Samyang Foods, AmorePacific, and collaboration-driven demand.

Key Arguments: South Korea looks bubble-like because forward net profit estimates for its memory giants imply they could become the world’s most profitable companies despite being commodity producers. The memory-chip cycle is not necessarily over, but Chinese competitors are improving yields and will likely add meaningful supply from 2027 onward, limiting upside. Taiwan’s AI rally is more sustainable than Korea’s because TSMC remains a bottleneck and has not materially increased supply, supporting pricing power. Retail speculation in Korea and Taiwan is unusually intense, with index-level gains, social-media frenzy, and violent reactions to theme mentions resembling prior bubble periods. Japan’s bull market has been helped by a weak yen, corporate governance reforms, buybacks, and reduced cross-shareholdings, but the trade may be later-stage now. Korea may now be the better stock-picker’s market because reforms are reducing governance discounts and small caps remain cheap despite broad enthusiasm in large-cap tech. China remains unattractive because policy can abruptly change, private businesses face weak protections, and the property bust plus tariffs have damaged growth and investor trust. Southeast Asia offers pockets of value, especially in Thailand and the Philippines, but lacks the momentum and catalyst visibility currently present in Korea. Foreign investors tend to concentrate narrowly in Asia, often buying only TSMC in Taiwan or the biggest Korean tech names, while missing broader value opportunities. For long-term capital, the recommendation is to own cheap, cash-generative businesses in undervalued Asian markets rather than chase expensive US or crowded Asian growth names.

Data Points: iShares MSCI Korea ETF 1-year return: 255% - Illustrates the scale of South Korea’s rally over the prior year. Taiwan market 1-year return: 112% - Used to compare Taiwan’s strong performance with Korea’s even larger surge. SK Hynix / Samsung Electronics 2028 net profit estimates: $200B / $250B - Cited as evidence that market expectations for Korean memory chip firms are extreme. KOSPI composition: 50–60% memory chip stocks - Shows how concentrated the South Korean market has become. Japanese yen level: ~160 per USD - Weak yen cited as a major tailwind for Japan’s stock market and exporters. Japan market level: Back to 1989 levels - Indicates how far the Nikkei has recovered from its long post-bubble slump. China residential new starts decline: Down 70%+ since 2021 - Evidence of a severe property-market contraction in China. Korean inheritance tax: 50% - One of the governance/disincentive issues discouraging Korean family firms from maximizing market value. Korean credit score stock valuation: Single-digit P/E; around 9–10x - NICE Information Service is presented as a cheap, high-quality Korean value idea. Japanese software valuation: 2–3x EV/sales - Examples of underappreciated growth opportunities in Japan’s domestic software sector. Thailand paint company valuation: 8x P/E - Toa Paint is cited as a cheap Southeast Asian idea. Korean index concentration: Cospi 50–60% memory chip stocks - Signals how narrow the rally is and why bubble concerns are elevated. Korean equities / governance reform year: 2025 - Directors became personally liable for related-party transactions, improving minority shareholder protections. Korea market multiple for some names: 10x book or higher - Used to argue commodity memory names have become too expensive.

Pivotal Quotes: "I think it is a bubble because specifically South Korea, I do think it is a bubble because the estimates for forward profits for SK Hynix and Samsung Electronics, we're now talking 200 billion US dollars in net profit estimates for 2028." — Michael Fritzel: His core thesis on Korea’s memory-chip rally and why it looks speculative. "If you can track inventories and track prices daily, that means they're standardized enough that I mean this is a commodity." — Michael Fritzel: Explaining why commodity DRAM differs from harder-to-replicate logic-chip businesses. "The party wants to maintain absolute control. So, that's problematic for private sector companies that have too much control." — Michael Fritzel: Summarizing his bearish view on China’s investability.

Implications: Listeners should treat Korea’s AI/memory boom as potentially late-cycle and favor select value, governance reform, and insider-buying situations over crowded large-cap momentum. China remains high-risk, while Japan and Southeast Asia offer more nuanced opportunities.