Data storage companies are experiencing a dramatic rally on Wall Street following comments from Nvidia CEO Jensen Huang at CES 2025, who highlighted data storage as the next critical bottleneck in the AI revolution. SanDisk, the flash-memory company recently spun out of Western Digital, led the surge with shares jumping approximately 28% on Tuesday, marking its strongest intraday performance since February and making it the top S&P 500 performer for the day.
The rally extended across the storage sector, with Western Digital shares climbing 17% and Seagate Technology rising 14% on the same day. SanDisk has already gained more than 40% in just the first few days of January, underscoring investor enthusiasm for the data storage opportunity in AI infrastructure.
Huang described the memory storage market as “completely unserved” during his CES appearance, emphasizing that it will likely become “the largest storage market in the world, basically holding the working memory of the world’s AIs.” He pointed to the exponential growth in AI storage requirements, noting that “the amount of context memory, the amount of token memory that we process, KB cache we process, is now just way too high.”
The storage boom is being driven by unprecedented demand from AI workloads, with global stored data expected to exceed 200 zettabytes in 2025—equivalent to 200 billion terabytes. This surge is creating significant supply constraints and driving dramatic price increases across the memory market.
Memory prices are skyrocketing as a result. According to Korea Economic Daily, Samsung Electronics and SK Hynix are pursuing server DRAM price increases of 60% to 70% in the first quarter compared to the previous quarter. Market analysis firm Counterpoint projects that memory prices could rise by 40% through the second quarter of 2025 as AI-driven demand continues to outpace supply.
An analyst from International Data Corporation characterized the situation as “an unprecedented inflection point, with demand materially outpacing supply,” noting that “for an industry that has long been characterized by boom-and-bust cycles, this time is different.” The rapid build-out of AI infrastructure is fundamentally straining the memory ecosystem, signaling “the end of an era of cheap, abundant memory and storage, at least in the medium term.”
Key Quotes
This market will likely be the largest storage market in the world, basically holding the working memory of the world’s AIs.
Nvidia CEO Jensen Huang made this statement at CES 2025, fundamentally reframing data storage from a commodity business to a critical AI infrastructure component that could become the world’s largest storage market.
The amount of context memory, the amount of token memory that we process, KB cache we process, is now just way too high.
Huang explained the technical drivers behind the storage crunch, highlighting how AI models’ growing context windows and token processing requirements are creating unprecedented memory demands that current infrastructure cannot adequately serve.
The memory market is at an unprecedented inflexion point, with demand materially outpacing supply. For an industry that has long been characterized by boom-and-bust cycles, this time is different.
An analyst from International Data Corporation emphasized that AI-driven demand represents a structural shift rather than a cyclical upturn, suggesting sustained growth and pricing power for storage companies as AI infrastructure build-out accelerates.
For consumers and enterprises alike, this signals the end of an era of cheap, abundant memory and storage, at least in the medium term.
The IDC analyst warned that the supply-demand imbalance will fundamentally change the economics of computing, with higher storage costs affecting everything from enterprise AI deployments to consumer device pricing in the coming years.
Our Take
The storage sector’s sudden prominence reveals how AI infrastructure requirements are evolving faster than the industry anticipated. While much attention has focused on GPU shortages and power constraints, the memory bottleneck may prove even more challenging to resolve quickly, as expanding storage manufacturing capacity requires significant capital investment and time. What’s particularly striking is Huang’s characterization of the market as “completely unserved”—a remarkable statement given the maturity of the storage industry. This suggests AI workloads are fundamentally different from traditional computing, requiring new architectures and approaches to memory management. The 60-70% price increases also indicate that storage manufacturers now have unprecedented pricing power, potentially making these previously cyclical businesses more attractive long-term investments. However, these cost increases could slow AI adoption among smaller companies, potentially concentrating AI capabilities among well-funded tech giants who can absorb higher infrastructure costs.
Why This Matters
This development marks a critical shift in the AI infrastructure narrative, moving beyond chips and data centers to highlight data storage as the next major constraint and investment opportunity. For years, the AI boom has focused primarily on GPU manufacturers like Nvidia, but Huang’s comments signal that storage capacity is becoming the new bottleneck as AI models require exponentially more memory for context windows and token processing.
The dramatic price increases—potentially 60-70% for server DRAM—will have significant implications for businesses deploying AI systems, increasing the total cost of AI infrastructure and potentially slowing adoption among cost-sensitive organizations. For investors, this represents a fundamental revaluation of the storage sector, which has historically been cyclical and commodity-like, now positioned as essential AI infrastructure.
The supply-demand imbalance suggests sustained pricing power for storage manufacturers, potentially breaking the traditional boom-bust cycle that has characterized the industry. This could reshape competitive dynamics in tech, favoring companies with strong storage capabilities and creating new opportunities for innovation in memory technologies. For enterprises and consumers alike, the era of cheap storage is ending, which may influence everything from cloud computing costs to device pricing.
Related Stories
- Nvidia CEO Jensen Huang Reveals Public Speaking Struggles Despite AI Success
- Meta and Nvidia Billionaires’ Wealth Soars $152B in AI Boom
- Nvidia faces US probe over potential export rules violations to China
- Groq Investor Warns of Data Center Crisis Threatening AI Industry
- Big Tech’s 2025 AI Plans: Meta, Apple, Tesla, Google Unveil Roadmap