By any measure, South Korea should have been one of the safest bets in the artificial intelligence revolution. And for a good part of this year, the bets paid off early investors. The country has the world’s largest memory chip manufacturers that dominate high-bandwidth memory (HBM), a critical component powering Nvidia’s AI GPUs. Add to that, world-class engineering talent and a government that actively backs advanced tech, and the story was perfect. Until it wasn’t.
In the last 40 days, over $2 trillion in market value just evaporated from the Korean stock market. That’s about 40% of the total value gone!
KOSPI has suffered one of its worst monthly declines on record, despite Samsung Electronics and SK Hynix – companies posting enormous profits on the back of booming memory chip demand – have seen their shares crash.
Investors bled as expectations seemed to have simply run too far ahead of reality. Meanwhile, regulators have clamped down on leveraged ETFs after speculative retail money magnified first the boom and then the bust.
This suggests that even if a country genuinely sits at the centre of the AI supply chain, stock prices are often detached from business fundamentals.
THE AI PYRAMID
AI isn’t a single industry. The AI economy resembles a pyramid. At the bottom are semiconductors, power infrastructure and cloud computing. Above that sit data centres. Then come foundation models. Only at the very top are applications that consumers use – think ChatGPT, Gemini or Claude.
Economic profits are being captured at the bottom of this pyramid and valuations soar for the top tier. Once again, America and China dominate here.
The United States is home to Nvidia, AMD, Broadcom, Microsoft, Amazon, Alphabet, Meta and OpenAI. China controls much of the manufacturing ecosystem and electronics supply chains despite restrictions. Huawei, Alibaba, Tencent, Baidu and DeepSeek have demonstrated that China is capable of even developing frontier AI ecosystems despite geopolitical pressure, and arguably by using US tech unscrupulously.
South Korea, in this mix, became indispensable because of memory chips, like Taiwan became for semiconductors.
India meanwhile is straddling between non-AI semiconductors, data centres and a handful of end-user language models demonstrated by startups like Sarvam AI.
KOREA CRASH AIN’T AI BUST
It would be a mistake to conclude that artificial intelligence is a bubble and South Korea’s stock market crash is a sign of a bigger bust in the making. Demand for AI compute continues to rise. Data centres are still being built. Enterprise adoption expands even though real cost saving for companies remains a question mark.
What collapsed in Korea were expectations. Markets had begun pricing AI companies as though growth would continue unabated for years. History suggests that’s not how technology revolutions work. At the turn of the millennium, the internet bubble burst before reshaping commerce.
South Korea has simply become the first major casualty of excessive optimism in AI.
INDIA’S M(AI)SSED BUS
India has no shortage of investor optimism. Quarterly earnings reports mentioning AI get rewarded. Data centre announcements attract premium headlines. In the startup space, AI models command higher valuations, even when stress brews behind the scenes (read Krutrim).
Investors who play the AI theme should ask a simple question: where exactly will India’s AI profits come from?
India does not manufacture advanced AI chips. It does not produce HBM memory. India has almost no large-scale GPU technology. Frontier AI models remain significantly smaller than those developed in the US and China.
Instead, India excels in implementation. Indian companies deploy AI. Consulting firms integrate AI. IT services customise AI. Banks, insurers and retailers implement agentic AI. But the core business for any of the abovementioned cohorts isn’t AI. Implementation is not the same as owning the platform. Companies laying the AI highway capture more value than those travelling on it.
LESSONS FOR INDIAN INVESTORS
Perhaps the biggest lesson from Seoul is not to park all money on a handful of technology bets. South Korea has reminded markets that even companies producing record profits can suffer brutal corrections when expectations become impossible to satisfy.
Diversification, even geographical, protects against extreme volatility.
AI is almost certainly real, and productivity gains will likely reshape the global economy over the next decade. But that does not automatically justify every valuation carrying an AI label.
Indian investors would do well to remember this before assuming every AI-themed stock is the next multibagger.