South Korea's 169% Semiconductor Export Supports AI Stocks, but Treasury Yields Remain the Key Risk

• South Korea's exports rose 53.2% year-over-year to a record US$87.75 billion in May, led by a 169% increase in semiconductor exports and a 291% increase in computer exports, indicating continued demand for AI infrastructure.
• Samsung Electronics gained nearly 10% after announcing shipments of next-generation HBM memory samples, signalling continued demand for AI server memory and stronger revenue expectations.
• If semiconductor export growth remains above 100% through Q3 2026, AI infrastructure suppliers are likely to continue outperforming the broader market; if interest rates rise and financing costs increase, valuation multiples could fall despite strong demand.
• Retail investors cannot reliably predict when hyperscale AI spending will peak or when companies will cut AI budgets, making short-term market timing difficult.
• A US payrolls report well above the 85,000 consensus forecast on June 5, combined with a sustained rise in the US 10-year Treasury yield toward 5%, would indicate that higher interest rates are becoming a bigger driver of equity valuations than AI spending.

Record Semiconductor Exports Lift AI-Linked Equity Markets

Asian equity markets rose on June 1 as investors responded to evidence of continued AI infrastructure spending. South Korea's KOSPI rose 4.4%, extending the previous week's 8% gain, while Taiwan's benchmark advanced nearly 6%. Japan's Nikkei gained 1.1%. US equity futures rose as investors continued to back companies exposed to AI infrastructure spending.

Investors are responding to evidence that AI spending is expanding from software into semiconductors, servers and data centers. Semiconductor production, memory demand, server deployments and data-center construction are providing measurable evidence of AI investment. Investors can track AI infrastructure spending through export data, memory shipments and data-center investment, providing clearer demand signals than share-price movements alone.

AI Infrastructure Spending Drives Semiconductor Demand & Valuations

As hyperscale technology companies expand AI infrastructure, demand for advanced graphics processors and memory products is increasing. Semiconductor manufacturers are raising production capacity, lifting export volumes and supporting revenue expectations. Investors are assigning higher valuations to companies exposed to AI infrastructure spending.

Samsung began shipping samples of its latest HBM memory chips, providing evidence of continued AI infrastructure demand. The announcement pushed Samsung shares nearly 10% higher, reflecting expectations for continued AI-related demand. South Korea's export data showed the same trend, with semiconductor exports up 169% and computer exports up 291% year-over-year.

Strong export growth is supporting revenue expectations for semiconductor manufacturers and other companies supplying AI infrastructure. As companies increase AI infrastructure spending, semiconductor demand is becoming a larger driver of corporate capital expenditure and equity-market performance.

AI Spending Supports Earnings as Market Concentration Increases

A small number of AI-linked companies are driving a growing share of market returns, increasing concentration risk despite strong AI infrastructure spending. The top 10 companies in the S&P 500 account for roughly 40% of the index, indicating that market gains are concentrated in a small group of stocks.

At Computex, Nvidia CEO Jensen Huang said Taiwan's AI ecosystem is preparing for significant investment in AI infrastructure. Even with strong demand, expanding semiconductor fabrication, data-center capacity and power infrastructure requires significant time and capital. Capacity constraints favour semiconductor manufacturers, infrastructure providers and power suppliers with existing capacity, while limiting near-term gains for companies further down the AI value chain.

Revenue-Backed AI Infrastructure Companies Lead Market Gains

Companies supplying semiconductors, memory products and AI infrastructure are benefiting from increased corporate capital spending. By contrast, sectors dependent on low borrowing costs or consumer demand may deliver slower earnings growth as interest rates remain elevated. Jensen Huang's comments point to planned AI infrastructure spending rather than speculation about future demand.

Semiconductor Exports & Treasury Yields Will Test AI Valuations

Stronger-than-expected US economic data and higher interest rates could weaken the current AI infrastructure trade. Higher interest-rate expectations would be particularly negative for highly valued AI-related stocks.

Semiconductor export growth below 50% year-over-year for several months, payrolls consistently exceeding forecasts or Treasury yields approaching 5% would weaken the case for continued outperformance by AI infrastructure stocks.

South Korea's 169% increase in semiconductor exports indicates continued demand for AI infrastructure. If export growth remains strong and companies continue increasing AI investment, suppliers of semiconductors, memory and AI infrastructure are likely to maintain higher valuations.