While Nvidia and OpenAI dominate AI headlines, a quiet revolution is happening in the memory chip sector. Companies like SanDisk, Micron, and Western Digital – once considered the unglamorous backbone of tech infrastructure – are seeing their stock prices skyrocket as AI’s insatiable appetite for data storage creates unprecedented demand. SanDisk shares have nearly doubled since January and are up almost 1,100% since August last year, with Micron and Western Digital tripling over the same period.
The Hidden Choke Point
“The narrative in the AI rally has shifted to memory being the choke point in the sustained AI capex build-out,” said Arun Sai, a multi-asset strategist at Pictet Asset Management. Nvidia CEO Jensen Huang recently declared that “holding the working memory of the world’s AIs” could soon become “the largest storage market in the world.” As AI models grow more sophisticated, the amount of data they consume and produce has skyrocketed, boosting sales of both flash-based storage and traditional magnetic hard drives.
Supply Squeeze and Investor Frenzy
The supply-demand imbalance is creating what Richard Clode, a tech portfolio manager at Janus Henderson, describes as pricing that’s “going berserk.” Manufacturers aren’t significantly increasing production capacity due to the historic cyclicality of the memory market and the high cost of building new factories. Ben Bajarin of Creative Strategies predicts shortages will continue until “2028 at least.” This scarcity has created a gold rush for hedge funds, with DE Shaw potentially making $3.9 billion in profits since September from timely bets on memory stocks.
Bubble Warnings from Industry Leaders
Amid this frenzy, Google DeepMind CEO Demis Hassabis has issued a sobering warning. “There are parts of the industry that do look bubble-like,” he said, specifically citing “multibillion-dollar seed rounds in new start-ups that don’t have a product, or technology, or anything yet.” His comments at Davos highlight growing concerns that AI investment patterns may be unsustainable, even as companies like Google maintain strong positions due to their scale and existing technology.
The Infrastructure Challenge
The AI boom is facing another critical constraint: energy infrastructure. Data center operators are planning coordinated lobbying campaigns to counter growing public backlash against AI infrastructure projects. Over two dozen projects were blocked or delayed in January 2025 alone due to concerns over rising energy costs, water consumption, and air pollution. “Nimbyism is coming to our space real fast,” said Andrew Power, CEO of Digital Realty. “There’s a tremendous amount of misperception that is slowing development.”
Job Market Realities
While memory stocks soar and infrastructure debates rage, the human impact of AI is becoming increasingly clear. Research from Stanford University suggests workers ages 22 to 25 are seeing the steepest declines in employment, especially in fields most exposed to AI-enabled automation. Software engineer jobs for this age group declined nearly 20% in 2025 compared to their peak in 2022. Business leaders emphasize that professionals need to develop blended capabilities, critical thinking, and translation skills between business and technology to remain competitive.
A More Discerning Market
“The AI trade is no longer just about holding a basket of exposed names,” noted Sai at Pictet. “The market has turned more discerning between winners and losers.” This shift comes as the long-running rally in megacap tech stocks has petered out following concerns over high valuations and massive spending. Nvidia remains 11% below its October peak, and among hyperscalers, only Alphabet has reached new highs since November.
Looking Ahead
The memory boom represents both opportunity and risk. While companies like HEN Technologies show how AI can drive innovation in unexpected sectors – using sensor data from firefighting equipment to create valuable physics data for AI training – the broader industry faces questions about sustainability. As Hassabis estimates artificial general intelligence is 4-8 years away with a 50% chance by around 2030, the infrastructure, investment, and talent challenges will only intensify. The memory sector’s current surge may be just the beginning of a much larger transformation – or a warning sign of overheating in an industry racing toward an uncertain future.