AI Revolution, Is It Really Profitable? Why Even Big Tech Can't Earn Money 'AI Subscription Fees Will Ultimately Keep Falling' - Dr. Yoo Shin-ik
AI Revolution, Is It Really Profitable? Why Even Big Tech Can't Earn Money "AI Subscription Fees Will Ultimately Keep Falling" | Dr. Yoo Shin-ik
Спикеры
Yoo Shin-ik
— Экономист, KB Bank WM Star Advisory Group
Dr. Yoo Shin-ik argues that the AI revolution is turning into a cost-efficiency trap. Massive capex is shifting to debt funding as free cash flow turns negative, while Chinese competition forces AI subscription fees lower. He warns that Korean memory stocks' cheapness hides risk from hyperscaler credit trouble and Chinese overcapacity, and recommends rotating into defensive sectors like consumer goods, utilities, and chemicals.
- AI revolution is becoming a cost-efficiency game, limiting profit potential.
- Hyperscalers' capex increasingly relies on debt, with FCF turning negative and credit risk rising.
- AI subscription fees structurally decline as Chinese low-cost competition drives commoditization.
- Korean memory makers Samsung and SK Hynix face headwinds from hyperscaler credit issues and Chinese memory expansion.
- Big tech earnings beats likely to be sold as markets focus on funding needs.
- Investors should rotate to defensive sectors (consumer goods, utilities, chemicals) and reduce AI growth exposure.
- Asset allocation and risk management are key in this environment.
Идеи
Rotate to defensive consumer, utilities, chemicals.
In an environment where AI growth stocks face margin and credibility issues, global investors are rotating to safe havens. The speaker recommends building a core portfolio of defensive sectors such as consumer goods, utilities, and chemicals, and using only residual funds for growth stocks. These defensive sectors provide stability and income as the AI cost-efficiency battle plays out.
Nvidia's credit guarantees risk future losses.
Nvidia, as the primary AI hardware seller, has been providing credit guarantees to buyers to finance GPU purchases. This creates a web of contingent liabilities. If buyers struggle with sales due to pricing pressure and low demand for expensive AI services, Nvidia's guarantees could result in losses, and the market may increasingly price in that credit risk. This unique exposure makes Nvidia vulnerable as the cost-effectiveness game intensifies.
Korean memory stocks riskier than they appear.
Samsung Electronics and SK Hynix appear cheap, but the top-of-the-chain hyperscalers are wobbling due to debt and credit concerns, which will cap upside for memory. Additionally, Chinese memory makers like Changxin and Yangji are aggressively expanding and bringing low-cost DRAM to the market, intensifying the cost-efficiency competition. The combination of shaky demand drivers and overcapacity risk means buying solely on valuation is dangerous; the semiconductor stocks face repeated circuit breakers and elevated volatility.
Debt-funded AI capex erodes profitability.
Hyperscalers and big tech are increasingly funding their AI capex with debt, as free cash flow turns negative. Their revenue growth, while fast, has been driven by massive investment that yields diminishing returns per dollar spent. AI subscription pricing is under pressure from Chinese competition, which forces a cost-effectiveness game. This leads to margin squeeze and credit risk, with CDS spreads and bond spreads widening. Upcoming earnings may beat but the market will focus on rising capex and negative FCF, causing post-earnings sell-offs. Investors should avoid or reduce exposure.
Далее
This 815 Money Talk (815머니톡) video, published July 30, 2026, features Yoo Shin-ik discussing Korean Consumer Goods Sector, Korean chemicals sector, Korean Utilities Sector, NVDA, 005930.KS, 000660.KS, GOOGL, AAPL, MSFT, AMZN. 4 trade ideas extracted by AI with direction and confidence scoring.
Speakers: Yoo Shin-ik · Tickers: Korean Consumer Goods Sector, Korean chemicals sector, Korean Utilities Sector, NVDA, 005930.KS, 000660.KS, GOOGL, AAPL, MSFT, AMZN