Memory chip stocks see wild swings as AI inference gains challenge semiconductor market rally

What's behind the volatility in memory chip stocks this week?

News broke Wednesday that Meta Platforms plans to start a cloud computing business, prompting speculation that the first signs of waning demand for computing capacity may be starting to appear on the horizon. Chipmaking stocks were down across the board on Wednesday, though they rebounded a bit in early trading Thursday. But technology analysts say that another development is likely more responsible for the midweek dip in chip stocks than any sign of excess computing capacity at Meta: an efficiency gain in the ability of artificial intelligence algorithms to answer particular questions, a process known as "inference." Frontier model OpenAI made a deal with chipmaker Cerebras in January to use the company's new AI chips in its inference stack, and OpenAI engineers are now seeing greatly reduced inference costs, according to a Monday report in The Information. Plentiful memory Cerebras' efficiency-boosting chips use a more plentiful type of memory - known as SRAM - than other AI chipmakers, and its success with OpenAI could ease the memory bottleneck and reduce the huge amount of pricing power in that section of the hardware supply chain. "The Cerebras chip uses SRAM, and SK Hynix, Samsung and Micron [produce] high-bandwidth memory DRAM, so if these guys really make a splash in the industry, then Micron gets hurt," Paul Meeks, head of technology research at Freedom Capital Markets, told CNBC Wednesday. Cerebras stock soared 19% on Monday, another 2% on Tuesday and was little changed on Wednesday. CBRS 5D mountain CBRS past five days. Efficiency gains at OpenAI could be the start of a trend, as hardware and software companies scramble to optimize their inference capabilities and find workarounds for the memory chip shortage. "We've got all these chip startups and small companies talking about new methods of doing inference with different chip architectures," analyst Bob O'Donnell at Technalysis Research said. "What we're seeing is significantly more diversity in the types of Silicon that's being used to do inference and the sense that all those options are viable because the demand is so high." Inference connection While the analysts CNBC spoke with didn't think Meta's cloud play was responsible for Wednesday's chip and memory rout, some do see a connection to what's happening in the inference field. Meta climbed 9% on Wednesday in anticipation of the hoped-for new revenue stream for the Instagram and WhatsApp parent, but some analysts are concerned that CEO Mark Zuckerberg's company could be missing out on efficiency gains in its own core advertising business. "We'd much prefer that Meta develop core AI products, leverage them over its base of around 4 billion users, and require massive compute for its own inference rather than selling access to its infrastructure," Doug Anmuth at JP Morgan wrote in a Thursday note to clients. Meta has been signaling its move into the cloud space for some time. At the company's shareholder meeting in May, Zuckerberg said the reason Meta hadn't started renting its servers is that it needed the computing power for itself. "It's possible that Meta is selling compute that's more suitable for inference, and its heavy spending on training will continue, but it still suggests to us that Meta's AI product traction beyond advertising remains limited," Anmuth wrote.

Wall Street flees software plays for triple-digit chipmaker boom

Software used to be the safe bet in artificial intelligence. Buy the companies writing the code, sit back, and let subscription revenue do the work. That trade has come undone. While the iShares Expanded Tech-Software Sector ETF sits about 20% below its record high, the chipmakers building the AI data centers are having one of the best stretches in stock market history. Micron, Intel, and Advanced Micro Devices all posted gains north of 100% in the second quarter. In addition to an expanding earnings base, investors are chasing order books and a supply squeeze that executives say could last years. Why money is leaving software for silicon The shift comes down to a simple idea on Wall Street. Everyone building AI needs the same scarce ingredients, and chipmakers are benefiting from a surplus of demand. Barclays analyst Anshul Gupta summed it up in a note published Tuesday, June 30, writing that the rotation out of AI hyperscalers and into AI enablers has pushed investor enthusiasm into semiconductors, fueling dramatic rallies, according to CNBC. Micron (MU), Intel (INTC), and AMD (AMD) gained a combined $2 trillion in market value during the quarter and now rank among the 10th, 11th, and 12th-most valuable technology companies in the country, CNBC reported. That's a remarkable jump for three companies that spent years trading in Nvidia's shadow. Even Nvidia, the biggest AI chip name by far, only gained 15% in the quarter, a modest number by comparison. Its customers had a mixed few months, too, with Meta stock slipping almost 2% while Alphabet climbed 24%, CNBC noted. Micron, Intel, AM: the numbers behind the rally * Micron stock more than tripled in Q2, adding $920 billion in market cap. * The company's revenue more than quadrupled as memory prices spiked, and gross margin jumped to 84.9% from just 39% a year earlier. * Intel stock surged 216% in the June quarter, adding roughly $480 billion in value as the company benefited from renewed CPU demand alongside its U.S. factory buildout. * AMD added about $615 billion after nearly tripling, helped by soaring demand for its server processors. * Networking chipmaker Marvell climbed about 200%, and Arm, which licenses chip designs, rose 134%. * The broader VanEck Semiconductor ETF gained 71% in the quarter, its best three months since the fund launched in 2000. Source: CNBC Micron CFO Mark Murphy told analysts on the company's June 24 call that free cash flow is expected to top $30 billion next quarter, with essentially all of it returned to shareholders through buybacks and dividends. "We're really pleased with the financial trajectory of the business," Murphy said. "The combination of memory being so important to so many markets, AI data center, the edge, enabling this or helping enable this technology revolution we have underway." Chief Business Officer Sumit Sadana said demand for high-bandwidth memory chips remains well above what Micron can supply through 2027 and even 2028, with the HBM market expected to top $100 billion in 2027. Over at Intel, CFO David Zinsner described a cultural overhaul under CEO Lip-Bu Tan that cut management layers and refocused the company on execution, telling a Bank of America conference the CPU market opportunity could reach $200 billion. AMD's Jean Hu pointed to similar strength, noting that CPU revenue grew more than 50% last quarter and is guided to grow more than 70% this quarter as agentic AI workloads drive demand for higher-core-count chips. For now, the message from chip executives hasn't changed. Supply is tight, customers are locking in multi-year agreements, and nobody on those earnings calls sounded like they expect the shortage to ease anytime soon. Cheng Xin/Getty Images Is there more upside left for chip stocks? Out of the 30 analysts covering Micron stock, 29 recommend "buy" and one recommends "hold." The average Micron stock price target is $1,564, indicating 52% upside from current levels. Out of the 35 analysts covering AMD stock, 28 recommend "buy" and seven recommend "hold." The average AMD stock price target is $510, indicating a 6% downside from current levels. Out of the 39 analysts covering Intel stock, 11 recommend "buy," 26 recommend "hold," and two recommend "sell." The average INTC stock price target is $97, indicating a 24% downside from current levels. The Arena Media Brands, LLC THESTREET is a registered trademark of TheStreet, Inc. This story was originally published July 2, 2026 at 2:47 PM.

Citi sends warning on semiconductor and hyperscaler stocks

Semiconductor stocks have been the most crowded corner of the AI trade in 2026. Memory company margins are at records, and chip ETFs have rallied sharply. The consensus view on Wall Street has been straightforward: As long as the biggest cloud platforms keep spending on AI infrastructure, demand for chips has nowhere to go but up. On June 30, Scott Chronert noticed something that puts a question mark on that assumption. Chronert is Citi's head of U.S. equity strategy and has been among the more bullish voices on equities this year, carrying an 8,100 year-end S&P 500 target above most of his peers. The note he published on June 30 is not a bearish call on the market. It is a specific observation about a tension building inside the AI trade that semiconductor investors may not have fully priced in yet. Citi said hyperscalers need something to show for their investment Chronert's characterization of the setup was direct, Seeking Alpha reported. "The tech trade is approaching a pivotal moment as rising semiconductor memory prices are set to clash with hyperscalers' return on investment expectations." The note is specifically about memory. Memory chip prices have quadrupled over the past year as AI data center buildouts from Microsoft, Google, Meta, and Amazon consumed supply ahead of consumer electronics. That surge lifted Micron Technology's gross margin from 39% to 84.9% in a single year and made memory one of the strongest-performing subsectors of the AI trade. SMH, SOXX, and XSD, the major semiconductor ETFs, have all rallied significantly on the back of that demand story. Chronert is raising a specific question about the other side of that dynamic. The hyperscalers paying those higher memory prices need to show investors the cost is generating returns. If they cannot make that case convincingly, their willingness to keep spending at the same pace becomes harder to assume. Why hyperscaler ROI pressure is the next test for chip stocks The scale of what hyperscalers are spending is what makes this a market-level question. As TheStreet reported, Goldman Sachs estimates the largest cloud platforms will spend approximately $754 billion on capital expenditure this year, an 83% increase from 2025, with that figure expected to exceed $900 billion in 2027. All of that spending flows directly into semiconductor demand. FactSet estimates semiconductor and semiconductor equipment earnings will grow 121% in Q2, making chips the fastest-growing sector in the S&P 500 by a substantial margin. That number tells you how much the market has already priced into these stocks. The tension has already started showing up in market behavior. As TheStreet reported, Mag7 stocks including Microsoft, Meta, and Amazon have sold off in recent weeks, while memory stocks including Micron kept climbing. One Roundhill ETF strategist described it as a "tale of two trades," where the hyperscalers' free cash flow has taken a hit from capex spending while memory company margins have benefited. That split is the market pricing in the same tension Chronert's note is describing. Which semiconductor stocks and ETFs carry the most exposure The stocks most directly in Chronert's frame are the ones most dependent on hyperscaler purchase orders. Micron Technology (MU) is the clearest example. Its revenue and margins have surged almost entirely on the back of AI data center demand, with gross margin climbing from 39% to 84.9% in a single year. That performance is also why Micron sits as the third-best performer in the S&P 500 year-to-date, with gains approaching 730%, according to Slickcharts data. High-bandwidth memory suppliers more broadly, including SK Hynix and Samsung, face the same dynamic. Each of them has seen demand and pricing driven almost entirely by the same handful of hyperscaler customers. If those customers moderate their procurement pace, the pricing environment these companies are benefiting from changes quickly. On the ETF level, the three Chronert specifically named are SMH (VanEck Semiconductor ETF), SOXX (iShares Semiconductor ETF), and XSD (SPDR S&P Semiconductor ETF). All three have rallied sharply on AI infrastructure demand in 2026. All three would feel the impact if hyperscalers begin signaling more discipline around memory procurement costs in their Q2 earnings calls. Nvidia sits in a different position. Its products are compute accelerators rather than memory, and its customer base includes a broader mix of enterprises and research institutions alongside hyperscalers. But even Nvidia is not immune to a broader AI spending recalibration, and any slowdown in data center buildout activity would affect demand across the AI hardware supply chain, not just memory. What this means for semiconductor investors heading into Q2 earnings Chronert is not making a bearish call on the market or on tech broadly. Citi's 8,100 year-end target for the S&P 500 remains in place. The note is a specific caution about a specific mechanism: the point where rising memory costs for the biggest AI buyers could start moderating the demand signal that semiconductor stocks are built around. In April 2026, Chronert called the earnings setup for tech a "reverse perfect storm," a term he used for conditions favorable enough to surprise to the upside. The June 30 note suggests at least one of those tailwinds is now working in a different direction, at least for memory and the hyperscalers buying it. The semiconductor names most exposed are those with the highest revenue concentration in hyperscaler customers. If the major cloud platforms start guiding more cautiously on capex or signal tighter scrutiny on memory procurement in their Q2 earnings calls, the impact on chip order expectations could move quickly. Google, Microsoft, Meta, and Amazon all report in mid-July. Investors in semiconductor stocks have those calls as the next concrete test. Any language suggesting spending discipline or slower data center expansion would put Chronert's concern in sharper focus. Any language reaffirming aggressive capex plans would push it back. Until then, the semiconductor trade sits in the position Chronert described: a market that has priced in strong hyperscaler demand, watching to see whether the ROI case holds up under shareholder pressure at the biggest buyers in the world. The Arena Media Brands, LLC THESTREET is a registered trademark of TheStreet, Inc. This story was originally published July 2, 2026 at 7:49 AM.