Hana Securities researcher Lee Young-joo stated on the 24th that the core evaluation criteria for global artificial intelligence (AI) investment cycles are shifting from capital expenditure (Capex) to efficiency and profitability. This shift occurs because AI infrastructure investment has already expanded to hundreds of billions of dollars, making the conversion of massive investments into actual revenue and cash flow the market's primary concern. Lee explained that while markets previously focused on how much hyperscalers would increase AI facility investment, the focus now centers on how efficiently increased investments are being utilized and whether they have secured sufficient economic viability to sustain further investment.
Market Shifts Focus from Investment Scale to Efficiency
Lee Young-joo analyzed in the 'Asset Allocation Window' report that recent volatility in AI-related stocks stems from market doubts about whether hyperscalers' large-scale AI investments can continue. The researcher stated that the market's baseline scenario still anticipates continued AI infrastructure investment by major big tech companies alongside expanding AI demand. Major forecasting institutions predict hyperscaler facility investment will continue to increase based on AI computing demand and cloud service advancement.
Lee emphasized that productivity is the core of the AI investment virtuous cycle. The growth outlook for semiconductor, data center, power, and network companies presumes continuous facility investment by hyperscalers. This means AI services must generate sufficient revenue to enable additional investment.
AI Revenue Exceeds Depreciation Costs for Two Consecutive Quarters
Lee assessed that AI monetization is gradually becoming visible. According to technology research firm Exponential View, AI-related revenue excluding China exceeded the depreciation costs of currently operating data centers and AI semiconductors for two consecutive quarters. While the stage of recovering all costs for new data centers and additional graphics processing unit (GPU) investments has not yet been reached, the analysis finds significance in AI services beginning to absorb existing infrastructure costs.
Hyperscalers Implement AI Monetization Strategies
Microsoft stated that Azure and Copilot demand is contributing to cloud business growth. Alphabet explained that AI product and infrastructure demand is the core driver of Google Cloud growth. Meta is pursuing strategies to increase utilization of existing AI infrastructure, including building new data centers and reviewing external cloud provision of some computing resources.
Lee diagnosed that market interest will ultimately shift from AI investment scale itself to investment productivity. The researcher stated that how many AI services can be provided with the same scale of facility investment and how quickly these can be converted to revenue and cash flow will be the key variables determining the next investment cycle.
![]()
Chart from Hana Securities report
FAQ
What did Lee Young-joo say about AI investment evaluation criteria on the 24th?
Lee Young-joo stated that the core evaluation criteria for global AI investment cycles are shifting from capital expenditure (Capex) to efficiency and profitability, as the market now focuses on whether massive investments can convert into actual revenue and cash flow.
Why are AI-related stocks experiencing increased volatility?
Lee Young-joo analyzed that recent volatility in AI-related stocks stems from market doubts about whether hyperscalers' large-scale AI investments can continue sustainably.
What evidence shows AI monetization is becoming visible?
According to Exponential View, AI-related revenue excluding China exceeded the depreciation costs of currently operating data centers and AI semiconductors for two consecutive quarters, indicating AI services are beginning to absorb existing infrastructure costs.
