Veteran Wall Street firm sends Micron stock a blunt warning

For Micron Technology (MU), the problem is not the possible elimination of the AI boom.

Instead, it’s whether the best days are now over, with much of the potential already priced in.

In a latest move, Citigroup cut its price target on Micron to $425 from $510 after spot prices for a mainstream DDR5 16GB DRAM product fell about 6% since Micron reported earnings on March 18. Citi maintained its Buy rating, but the message is landing. What is the message?

Essentially, one of Wall Street‘s favorite AI memory stocks can become unsettled when short-term pricing shifts in the opposite direction.

That matters because Micron just delivered an exemplary quarter, which most chipmakers will want coming out of earnings season.

Revenue nearly tripled from a year earlier to $23.86 billion, non-GAAP earnings reached $12.20 a share, and the company guided fiscal third-quarter revenue to about $33.5 billion, far above where Wall Street had been.

CEO Sanjay Mehrotra said Micron set records across revenue, gross margin, earnings per share and free cash flow.

However, the prices of memory stocks do not trade on revenue alone.

Instead, the focus is on a mix of pricing, supply and fear.

And now the fear is that a new efficiency push from Google (GOOGL) could now make a huge difference. The new efficiency push is making some AI workloads less memory-hungry just as Micron is ramping spending and investors are beginning to inquire how long the memory crunch will last.

Shares were trading at about $353.31 on April 1, up roughly 4.6% on the day, but the bounce shows how erratic the share price movement is lately. Micron is still one of the clearest ways to play the AI-driven demand for advanced memory.

Yet, it remains a classic reminder for investors that are looking to play the markets about what happens when Wall Street starts worrying about what happens after the shortage.

The easiest mistake investors can make here is to read Citi’s target cut as a sign that Micron’s business is suddenly over.

More AI Stocks:

That is not what the numbers indicate.

Micron’s fiscal second quarter was enormous.

• Revenue jumped to $23.86 billion from $8.05 billion a year earlier.

• Non-GAAP net income climbed to $14.02 billion, or $12.20 per share.

• The company also reported $11.90 billion in operating cash flow and $6.9 billion in adjusted free cash flow.

• Then it gave fiscal third-quarter guidance for about $33.5 billion in revenue and around $19.15 in non-GAAP diluted earnings per share.

Those numbers tell of a very robust demand cycle currently at work.

The reason is simple: AI servers need a lot of memory, especially the high-performance chips that are in use in data centers. Micron is one of only three major suppliers of high-bandwidth memory chips, along with Samsung Electronics and SK Hynix. The strong demand from tech giants building AI infrastructure helped the company make record profits in the last quarter.

Micron is acting like that demand is real and durable.

Micron increased its fiscal 2026 capital spending plan by $5 billion, taking total planned spending to more than $25 billion, with spending set to go up again as construction ramps.

Most of that extra money is going toward expanding manufacturing, like buying a new fab in Taiwan and paying more for construction because of expected future growth in output. Investors may be scared by that kind of spending because it brings up the old memory-cycle question: if everyone builds more, will prices stay high?

Related: Apple just got a brutal iPhone 18 warning

However, Citi is not revealing the bullish case.

The bank maintained its earnings estimates for Micron in place and said memory suppliers are now discussing three- to five-year strategic agreements with hyperscale customers.

Those deals could include promises to buy a certain amount, prepayments, and price changes based on how the market is doing. That matters because long-term contracts can help keep the business stable even when spot prices drop and traders start to panic.

This is why the stock story is more complicated that the headline will have you know.

Micron’s business is still thriving. Its customer base still appears hungry for memory. But the stock will no longer trade on just memory momentum alone.

Investors want to see proof that the AI memory boom can last through lower spot prices, a bigger spending plan, and new worries that smarter software could make hardware less intense at the edges.

Related: Nvidia CEO makes bombshell call on AI’s next big thing

The conversation repeatedly circles back to Google’s TurboQuant.

Google Research said on March 24 that TurboQuant enables “massive compression” for large language models and vector search engines. Google said that high-dimensional vectors can use up a lot of memory and slow down the key-value cache, which it called a fast memory layer that helps systems avoid doing the same work over and over. Some investors think that the memory opportunity gets smaller right away if those bottlenecks can be fixed.

But Citi is arguing the fear is to simplified.

The bank said that making things more efficient can lower the cost per query, but that cheaper technology often leads to more use, not less. Investors have seen this kind of logic in tech many times before: when something gets cheaper and easier to use, demand can grow enough to more than make up for the efficiency gain. Citi also said that KV cache is still important because as models make more tokens, each new token has to pay attention to the ones that came before it, which keeps memory and compute needs high.

The near-term question is not whether AI is still in need of memory.

It does.

The real question is whether spot weakness will remain restricted or broaden to contract pricing.

If Micron and its big customers sign long-term contracts that lock in volume and support prices, this decline in stock price may look more like a reset than a warning sign. If lower spot prices start to affect bigger contract talks, Wall Street will probably start to ask harder questions about whether Micron’s huge margins are closer to a peak than investors want to believe.

That’s why Citi’s move is important, even though the bank kept its Buy rating.

It doesn’t mean that Micron is broken. It says that Micron has more to prove now. The stock needs to prove that its AI-powered strength can hold up in a market that is more skeptical. It has to show that spending more on capital is a smart way to grow, not the first step toward having too much of something. And it has to show that software tricks like TurboQuant that make things run faster don’t make the memory shortage story a shorter-lived trade.

The simple version for everyday readers is this: Micron is selling a product that AI still needs, but investors are worried that the market might be getting too good too fast.

The message is clearer for people who invest in the stock market: Micron is still one of the best AI memory stocks on the board, but it’s also the kind of stock that can drop a lot when traders stop asking how big the boom is and start asking how long it lasts. The company’s most recent results were great. The next part is about showing that the pricing story is just as strong as the revenue story.

• Citigroup cut its target on Micron to $425 from $510 after mainstream DDR5 16GB DRAM spot prices fell about 6% since Micron’s earnings report.

• Citi kept its Buy rating and maintained earnings forecasts, signaling concern about near-term pricing rather than a collapse in the long-term thesis.

• Micron’s fiscal second quarter was still huge, with $23.86 billion in revenue and guidance for about $33.5 billion in fiscal third-quarter revenue.

• Reuters reported that Micron lifted fiscal 2026 capital spending by $5 billion to more than $25 billion, which has raised fresh worries about future supply growth.

• Google’s TurboQuant has added a new fear that some AI workloads could become less memory-intensive, though Citi argues cheaper AI can ultimately drive more demand.

• The next key signal for investors is whether spot-market softness spreads into longer-term contract pricing.

Related: Micron CEO drops a bombshell after Micron’s huge earnings beat

This story was originally published by TheStreet on Apr 2, 2026, where it first appeared in the Investing section. Add TheStreet as a Preferred Source by clicking here.