After Google reported earnings, its share price fell below its annual moving average and formed a death cross, weakening its medium- to long-term trend, while Facebook is also struggling near its annual moving average. Against the backdrop of Warren Buffett's heavy bet on Google, the position is currently sitting on a book loss of about US$7 billion, or about NT$224 billion. With US heavyweight tech stocks Google, Facebook, Amazon and Intel reporting weak earnings and undergoing sharp corrections, analysts expect four key changes and structural impacts for Taiwan's AI supply chain.
Orders Concentrate Among Four Major Clouds, Taiwan Makers' Structure Stabilizes
First, capital will concentrate in top-tier cloud service giants, making Taiwanese makers' orders more secure. Elon Musk noted more than 200 emerging cloud-computing vendors facing bankruptcy, such as Nebius, CoreWeave and others, which are themselves high-risk customers. After these companies run out of funding, AI computing power will concentrate in the hands of four well-capitalized top-tier cloud providers: Microsoft, Google, Amazon and Facebook. Hon Hai, Quanta, Wistron and other Taiwanese makers are already first-tier contract manufacturers for Microsoft and the other three major cloud providers. After the customer reshuffle, bad-debt risk will fall and long-term profitability will be better protected.
From Training to Inference, In-House Chips Bring Opportunities
Second, the focus of computing power is shifting from training to inference. Over the past two years the market疯狂 purchased chips for large-model training, and as Microsoft, Amazon and other major firms are now being asked by the market to monetize AI applications, AI infrastructure is accelerating toward inference applications. To reduce reliance on expensive NVIDIA chips, North American giants are increasing investment in in-house inference chips, such as Google TPU, which will drive advanced-process outsourcing to TSMC, as well as long-term opportunities for ASIC design service firms such as MediaTek, Alchip and Global Unichip.
Bottlenecks Remain, Spec Upgrades and Capex Stay High
Third, hardware bottlenecks remain and the spec-upgrade trend is unchanged. The global data-center expansion wave has not stopped because of the stock-price correction, with major firms pledging that capital expenditure this year will continue to hit new highs. As NVIDIA's high-power racks enter volume production and iteration this year, manufacturing bottlenecks in water cooling, liquid cooling, high-spec power supplies and high-bandwidth memory remain. Taiwanese makers that can overcome process-yield hurdles and squeeze into the high-end component supply chain will still enjoy significant advantages in long-term average selling prices and gross margins.
Double-Digit Profit Momentum Intact, Valuations Return to Reasonable Levels
Fourth, share prices have decoupled from fundamentals, with long-term valuations returning to reasonable ranges. Institutional investors estimate that, driven by the AI theme, overall corporate earnings in Taiwan's tech supply chain will still post very strong double-digit momentum this year, and the foundation for a medium- to long-term bullish structure remains intact. The current plunge is a deleveraging and overheated-valuation digestion. When share prices return to reasonable or even undervalued levels, it will instead provide entry points for long-term value investors.