Dan Ives, the Wedbush analyst known for his bullish technology calls, has declared the semiconductor sell-off "way overdone" as fears over artificial intelligence spending rattled shares of Nvidia (NVDA), Micron (MU), SK Hynix (SKHY) and SanDisk (SNDK).

The four tickers named in his call span the memory and accelerator supply chain. That breadth matters. When Micron, SK Hynix and SanDisk fall alongside Nvidia, the market is not repricing one vendor. It is repricing the entire AI hardware stack, from HBM memory suppliers to the GPU makers that consume their output.

What triggered the sell-off?

Investor anxiety centers on AI capital expenditure. The worry is straightforward: hyperscalers and enterprises may cut back on data center buildouts, which would hit demand for Nvidia's accelerators and, in turn, the high-bandwidth memory that Micron and SK Hynix supply for those systems.

Ives rejects that reading. His verdict, delivered in a note flagged by TradingView, is that the selling has overshot any realistic deterioration in AI spending. He did not mince words: the sell-off is "way overdone."

That is a direct quotation, and it carries weight precisely because Ives has been among the most persistent defenders of AI infrastructure spending on Wall Street. His position has not wavered through prior corrections in the sector.

Which companies are caught in the slide?

The affected names cover distinct layers of the AI hardware market:

• Nvidia (NVDA) — the dominant supplier of AI accelerators whose data center revenue has anchored the sector's rally
• Micron (MU) — the US memory maker supplying DRAM and high-bandwidth memory for AI systems
• SK Hynix (SKHY) — the Korean memory leader deeply tied to HBM demand
• SanDisk (SNDK) — the newly separated NAND flash specialist

A decline across all four suggests investors treat AI memory and AI compute as a single trade. Ives argues that logic has run ahead of the fundamentals.

Is the analyst right about the fundamentals?

The bull case rests on continuity. AI infrastructure programs at major cloud providers have shown no confirmed signs of cancellation, and memory makers have spent the past year in a supply discipline regime rather than a glut. When a sell-off hits names with intact order books, analysts of Ives's disposition tend to call it noise.

The bear case needs no encouragement from fundamentals at all. Semiconductor stocks carried premium valuations through the AI rally, and premium valuations correct violently on even modest doubt about future spending. A rumor of slower capex can do the work of a confirmed slowdown.

Ives is not alone in viewing the pullback as exaggerated, but he is the loudest voice saying so this week. The counterargument from skeptics is that "overdone" is a timing claim, not a fundamental one — even a correctly priced decline can overshoot on the way down and still leave stocks lower a year out.

What happens next?

The market will test Ives's call against the next round of earnings from the affected companies. Micron's results and guidance will signal whether memory demand is holding; Nvidia's data center numbers remain the sector's bellwether. If AI spending holds firm, the correction should prove temporary; if even one hyperscaler trims its budget, the sell-off will look, in hindsight, like an early warning rather than an overreaction.