Micron Rebounds 7% Pre-Market, Musk Personally Endorses. Wells Fargo Gives $1,220 Price Target

TradingKey - Following last Friday's broad sell-off, the U.S. storage sector saw a strong rebound in pre-market trading today, with Micron Technology ( MU) gaining more than 7%.

On the news front, Tesla ( TSLA) CEO Elon Musk publicly expressed a bullish view on Micron Technology during an interview at JPMorgan Chase's global headquarters, stating that "the real bottleneck lies in chip manufacturing capacity" and that Micron's current production capacity is still far below actual demand.

In addition, Wells Fargo ( WFC) significantly raised its price target for Micron Technology to $1,220.

Micron Technology CEO Sanjay Mehrotra shared deep insights into the storage industry in the AI era during a podcast last week. This veteran with over 40 years of experience in the semiconductor industry is currently leading Micron through the most rapid growth phase in its history.

"AI is still in its very early stages," is the core judgment Sanjay has repeatedly emphasized. In his view, as large models, Agent AI, and inference applications continue to evolve, AI requires not only more powerful computing power but also stronger "memory capabilities."

Longer context windows, larger model sizes, and the exponential growth of token consumption are all driving the continuous rise in storage demand. He used the term "intelligence itself" to describe storage's new role: "Today, storage is not just about making devices run; it is supporting the 'intelligence' itself within AI, helping artificial intelligence become smarter."

Sanjay believes that the AI race is not only a race for computing power but also a race for storage. He explained that the essence of AI is data, and data cannot exist without storage. As models grow larger and inference demand continues to rise, with AI moving from training to inference and from data centers to the edge, the demand for storage will only increase—requiring larger capacity, higher performance, and lower power consumption.

He specifically mentioned the dependence of token economics on storage: "As token usage grows, context windows become longer, and KV cache requirements increase, the models themselves also become larger. AI needs more than just computing power; it needs the ability to 'remember'."

Structural constraints on the supply side have further intensified the tightness in the storage industry. Sanjay pointed out that the current storage industry is facing a long-term supply constraint rather than a short-term supply-demand mismatch. Advanced storage products consume more wafers, and building new wafer fabs often takes three to four years, with a similarly long subsequent production ramp-up.

More importantly, as technology nodes advance, the growth in storage capacity output per wafer is declining. He predicts that the industry's supply tightness is likely to persist beyond 2026.

Micron has entered into collaborations with companies including NVIDIA ( NVDA ), Google ( GOOGL ), and Amazon ( AMZN ), among a total of six hyperscale customers, to establish long-term supply partnerships, solidifying its strategic position in the high-end AI memory supply chain.