At some point on Tuesday morning, Kim Yong-beom — South Korea’s presidential policy chief and one of the most influential economic thinkers in the country — posted a thought on Facebook. It was not a formal policy announcement. It was not a bill introduced to the National Assembly. It was not even a press conference. It was a social media post, shared by a senior adviser to President Lee Jae-myung, suggesting that South Korea should pay its citizens a “dividend” from taxes on AI profits — a recognition, in Kim’s framing, that the extraordinary wealth being generated by the artificial intelligence boom deserved to be shared with the broader public rather than concentrated among shareholders of chip companies and core engineers.
By the time Asian markets registered the implications of those words, the benchmark Kospi index had sunk as much as 5.1% in a single session. Samsung Electronics and SK Hynix — whose HBM memory chips have made both companies among the biggest beneficiaries of the global AI infrastructure buildout — fell sharply before recovering when Kim clarified that he was proposing to redirect excess tax revenue from the AI boom, not impose a new windfall levy on corporate profits directly. Shares in both companies eventually recouped much of their early losses. But the whipsaw — a 5% intraday drop triggered by a Facebook post that a presidential office spokesperson later characterised as Kim’s personal opinion rather than official policy — tells a story about markets, political risk, and the global politics of AI distribution that extends far beyond Seoul.
What Kim Actually Said and What He Meant
The distinction between “tax AI profits directly” and “redirect excess tax revenue generated by the AI boom” is not merely semantic. It is the difference between a windfall tax on Samsung’s and SK Hynix’s income — which would directly reduce corporate earnings — and a fiscal redistribution policy that uses the higher-than-expected tax revenues that flow to the government when companies are more profitable. The first would hit company earnings and shareholder returns immediately. The second is a broader argument about how governments should allocate the fiscal windfall that an AI boom creates.
Kim’s clarification shifted the proposal from the former category to the latter. He was not, ultimately, proposing to confiscate corporate profits. He was arguing that when the AI boom generates surplus tax revenues beyond what budget models expected, those revenues should be directed toward a “national dividend” rather than absorbed into general government expenditure.
But the analysis Kim presented around the proposal is worth examining in full, because it is the diagnosis that has attracted attention far beyond the immediate question of how to fund a dividend. Kim warned that AI-era profits could gather around a narrow group, writing: “Excess profits in the AI era are, by nature, concentrated.” He said shareholders of memory companies, core engineers, and asset holders may gain the most, while the broader middle class may get only indirect gains — a stronger currency, small fiscal benefits, or mild asset appreciation. The worry, as Kim framed it, is what he called “K-shaped divergence”: South Korea becomes richer as a nation, but the distribution of that wealth is so uneven that many citizens never feel it.
That diagnosis — a country enriched by technology while many of its people fall further behind in relative terms — is not unique to South Korea. It describes a dynamic that economists, political scientists, and AI researchers have been warning about for years across every economy where AI adoption is advancing rapidly.
Why South Korea Is Where This Debate Is Happening
South Korea is not a random location for this conversation. It is one of the most precisely positioned economies in the world to experience both the maximum upside and the maximum distributional anxiety of the AI revolution — and the concentration is not subtle.
Samsung Electronics and SK Hynix together hold approximately 65% to 70% of the global market for High Bandwidth Memory — the specialised chip architecture that is the critical enabling hardware for AI training and inference at scale. Every data centre being built by Microsoft, Google, Amazon, Meta, and the hyperscalers of China and the Gulf runs on HBM. The AI infrastructure arms race that has generated the capital expenditure boom visible in every major technology company’s 2026 earnings report flows, in significant part, through Suwon and Icheon — the South Korean cities where Samsung and SK Hynix operate their most advanced fabs.
The financial benefits of that position have been real and large. SK Hynix’s HBM revenue doubled in 2025. Samsung has been investing hundreds of billions of won in HBM3E and HBM4 capacity. The companies’ shareholders — including large institutional investors, wealthy Koreans, and global funds — have captured the equity upside of that position. The employees who work in the fabs and design centres have captured the salary and bonus upside. The Korean government has captured the corporate tax upside in the form of higher-than-projected revenues.
What the broader Korean middle class — the office workers, small business owners, teachers, and service sector employees who make up the majority of the country’s workforce — has captured is considerably less direct. Their lives are not obviously better because SK Hynix’s quarterly profit beat expectations. Their wages have not risen proportionally. Their housing costs, education costs, and the general cost of living have continued to rise independent of the chip boom. The “K-shaped divergence” Kim is describing is not a theoretical risk. It is a lived experience for a large proportion of the population.
The Global Policy Context
Kim’s proposal sits inside a wider global debate. In April 2026, Elon Musk again backed the idea of “universal high income,” saying it could be funded by an “increase in the money supply” triggered by AI and robotics. OpenAI CEO Sam Altman has also backed Universal Basic Income and a Public Wealth Fund idea in the AI economy. As of May 2026, over 92,000 tech workers have been laid off in the first five months of the year, with companies shifting more money into AI infrastructure.
The specific policy models Kim referenced are notably varied, reflecting the principle-over-specifics nature of the proposal at this stage. He listed possible vehicles including youth entrepreneurship asset accounts, rural basic income programmes, artist support funds, stronger pensions for the elderly, and education accounts for the AI age. That menu is more useful as a statement of distributional intent than as a costed policy programme. What Kim is establishing is not a specific mechanism — he explicitly acknowledged this is “still a principle, not a fixed cash policy” — but a frame for thinking about who should benefit from South Korea’s AI wealth.
Franklin Templeton’s Christy Tan captured the broader regional resonance of the proposal: “Asian economies do want to provide a signal of ownership in the shared future that involves digitalization and AI.” That framing — “signal of ownership” — is analytically precise. The citizen dividend idea is not primarily a welfare programme. It is a political and social claim that the gains from AI should not be treated as the exclusive property of the companies that manufacture the hardware and the investors who own them. It is an argument that the workforce, the citizens, and the social infrastructure of a country contribute to the conditions that make a tech boom possible — through education systems, through regulatory frameworks, through the consumption that sustains the companies — and that those contributions warrant a share of the outcome.
The Market’s Overreaction and Its Signal
The Kospi’s 5.1% intraday plunge was, by any analytical standard, an overreaction to a social media post articulating a principle rather than a policy. The clarification that Kim was discussing excess tax revenue rather than a windfall levy resolved the immediate investor concern. Samsung and SK Hynix recovered most of their losses. The index stabilised.
But the speed and violence of the market’s initial reaction is itself informative. It tells you that investors in Korean equities are acutely sensitive to any signal that the extraordinary profitability of AI-era chip companies might be subject to redistribution pressures — whether through direct taxation, windfall levies, or fiscal mechanisms that channel corporate tax revenues toward citizen dividends rather than infrastructure spending or debt reduction.
That sensitivity is rational. If South Korea implements a meaningful redistribution programme funded by AI-generated tax revenues, the marginal corporate investment in Korean chip infrastructure may decline at the margin, as after-tax returns come under pressure relative to alternative locations. The policy debate Kim has opened is not cost-free for the companies whose current trajectory underpins the Korean economic miracle of the 2020s.
What Kim is asking, fundamentally, is a question that every economy where AI is concentrating wealth is going to have to answer eventually: who owns the AI dividend? The market moved 5% on a Facebook post because it knew, even before the clarification, that the answer to that question is not yet settled — and that whoever settles it will be deciding the financial architecture of the next decade.
Written by Shalin Soni, CMA specializing in financial analysis, global markets, and corporate strategy, with hands-on experience in financial planning and analytical decision-making.
ALSO READ
• Americans Evacuated From Hantavirus Cruise Arrive in Nebraska as Two Cases Confirmed
• AirTrunk to Invest $3 Billion in Malaysia Data Centers as AI Demand Surges
• London Stocks Slip as Investors Weigh Mixed Earnings and Central Bank Outlook
Source: Based on Bloomberg and publicly available financial information.