In-depth analysis of why AI chip stocks have crashed (valuation bubble, excess capacity, liquidity shock), and provide investment advice: prioritize upstream equipment, industry-leading companies with strong moats, and domestic substitutes; avoid concept small-cap stocks and companies expanding with heavy assets.

Trillions of dollars evaporated overnight—Can chip stocks still be bought?

1. The core reason behind this plunge

‌The fundamental cause is the bursting of valuation bubbles‌

The current AI chip market started in 2023, and by May 2026, the Philadelphia Semiconductor Index had increased by over 220% cumulatively, with the sector’s average price-to-earnings ratio (TTM) reaching 48 times, placing it in the 95th percentile historically, far above traditional tech sectors. Especially leading AI chip companies like Intel, Nvidia, and Western Digital, whose short-term gains have already priced in growth expectations for the next 2-3 years, are vulnerable to profit-taking and profit warnings if fundamentals weaken.

‌Shift in fundamental expectations: from "AI demand explosion" to "overheated capital expenditure"‌

Previously, the market consensus was that AI computing power demand would continue to surge, but recent disclosures from major chip giants have broken this consensus:

Oracle, in a bid to secure AI computing orders, has ramped up capital expenditure to 75% of annual revenue, turning free cash flow negative, with debt surpassing $100 billion. The market is beginning to question the profitability of the "AI infrastructure heavy-asset model";

Storage chip manufacturers Western Digital and Seagate, which previously expanded production due to AI-driven demand and price increases, are now facing downstream data center inventory backlog, with expectations of price cuts to clear inventory causing panic;

Although Intel’s AI chip orders are growing, its foundry business posted a quarterly loss of over $2 billion, and ongoing cash burn raises concerns about profit stability.

In simple terms: the market has shifted from "seeking AI chip market share" to "fearing overcapacity," and expectations have completely reversed.

‌Liquidity and sentiment resonance amplify declines‌

This week, US CPI data slightly exceeded expectations, delaying the Fed’s rate cut expectations. The 10-year US Treasury yield rose back to 4.6%, putting pressure on high-valuation growth stocks; at the same time, many AI chip stocks were included as collateral for margin trading, and their declines triggered forced liquidations, further amplifying the fall, creating a "decline—forced liquidation—further decline" negative feedback loop.

2. Can chip stocks still be invested in later? Clarified into two categories

‌Yes, but absolutely no blind buying—must choose the right niche and targets‌:

‌Long-term investment logic still exists (worth paying attention to)‌

‌Upstream core equipment/materials‌: Whether expanding capacity or competing, AI chips rely on upstream components like lithography machines, photoresists, and silicon wafers. These companies have stable orders, do not directly bear downstream demand fluctuations, and are relatively reasonably valued, such as Applied Materials and TSMC supply chain-related stocks;

‌AI leaders with real barriers‌: For example, Nvidia, although highly valued, holds over 80% market share in GPUs, with the next-generation chips already mass-produced and orders full. Their core barriers remain intact, and after a correction, they could present long-term investment opportunities;

‌Segment leaders benefiting from domestic substitution‌: As the global chip landscape reconfigures, domestic and emerging market chip manufacturers are steadily replacing overseas market share in power semiconductors, automotive chips, and other fields. Growth prospects are strong, and valuations are generally lower than overseas AI chip leaders.

‌Directions to avoid (greater risks than rewards)‌

‌Pure AI concept stocks without core technology‌: These stocks have surged early on but lack fundamental support. After correction, they may experience prolonged declines;

‌AI computing power companies with heavy asset expansion and deteriorating cash flow‌: Models like Oracle that rely on leverage to grab orders are vulnerable if demand slows, risking liquidity issues—best to avoid;

‌Traditional storage chip manufacturers with overcapacity‌: Facing short-term inventory pressure, it’s unlikely they will see a rally within the next 1-2 quarters.

View Original

This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.

2 Likes