AI Chip Sell-Off Deepens; Apple Tops Nvidia: July 28

A broad reassessment of AI valuations rattled semiconductor stocks on Tuesday, pushing memory giants to multi-month lows while Apple quietly retook the title of the world’s most valuable public company.

Key points

• SK Hynix fell as much as 13% in Seoul, extending its decline to roughly 45% from a June record high and erasing around $570 billion in market value over recent weeks.
• Samsung dropped as much as 11% as investors cited worries about AI spending sustainability and advances by Chinese chipmakers.
• Nvidia shed 5%, dropping its valuation to roughly $4.77 trillion.
• Apple rose about 1% to approximately $4.95 trillion, reclaiming the top spot from Nvidia.
• SpaceX extended its post-IPO slide, with the stock hitting new all-time lows and the company having shed roughly $1.2 trillion from its peak valuation.

What is driving the memory sell-off?

Two concerns are doing most of the damage. First, investors are questioning whether the pace of AI infrastructure spending by hyperscalers can hold at current levels. That uncertainty hits memory chipmakers hard, since high-bandwidth memory demand is almost entirely an AI buildout story. Second, reports of progress from Chinese semiconductor firms are raising the prospect of future competition in a market that SK Hynix and Samsung have effectively owned.

The sell-off spread to Micron as well, confirming this is a sector-wide repricing rather than company-specific news. Micron is the primary US-listed memory name with direct HBM exposure, so US investors watching Korea’s session had ample warning before their own market opened.

For context, SK Hynix alone has lost around $570 billion in market value from its peak over the past few weeks. That is a significant destruction of wealth concentrated in a short window, and it underscores how much of the Korean chip rally had been built on AI demand assumptions that the market is now stress-testing.

Apple’s quiet ascent past Nvidia

Apple’s return to the top of the market-cap rankings is less a story about Apple doing something dramatic and more a reflection of Nvidia having a bad day. Apple gained roughly 1%, while Nvidia dropped 5%, and the gap opened up. Apple has largely stayed on the sidelines of heavy AI capital expenditure, which analysts have recently framed as an advantage given the scrutiny now landing on companies with massive AI spending commitments.

Whether Apple stays at the top depends heavily on whether the current AI chip sentiment stabilizes. Nvidia’s valuation swings have been wide enough that a single good session could flip the rankings back.

SpaceX slide and the Congress conflict question

SpaceX’s stock continues to fall, hitting new all-time lows on Tuesday after its post-IPO peak. The decline now totals roughly $1.2 trillion from the high. That is a remarkable reversal for a company whose IPO generated enormous excitement just weeks ago.

Adding a political dimension, six members of Congress have purchased SpaceX shares since the stock began trading publicly. CNBC reports that five of the six sit on committees whose jurisdiction touches SpaceX directly, covering defense programs, AI, telecommunications, and securities markets. The purchases have drawn conflict-of-interest scrutiny. This is unlikely to affect SpaceX’s business operations near-term, but it adds reputational noise at a moment when the stock is already under pressure.

Anthropic and the open-source debate

Separately, Anthropic CEO Dario Amodei clarified the company’s position on open AI models, rejecting an outright ban but calling for safety testing requirements. This follows Nvidia CEO Jensen Huang’s public push for open AI development. The debate matters for investors because the open-source question shapes competitive dynamics between closed-model companies like Anthropic and the broader ecosystem. A regulatory environment that restricts open models would benefit closed-model incumbents; one that leaves them unrestricted increases competitive pressure.

On the infrastructure side, Silvaco announced it is integrating Nvidia’s accelerated computing and AI tools into its physics-based simulation platform for semiconductor design, a niche but illustrative example of Nvidia’s compute stack embedding deeper into chip development workflows.

This article is for informational purposes only and does not constitute investment advice.