"The AI race is not just a computing power race, but also a storage race," concluded Sanjay Mehrotra, CEO of Micron Technology.
In the podcast "A Bit Personal" on June 5th, Sanjay gave a rare, in-depth interview recorded at home. Beyond the usual industry insights, this personal conversation led him to voluntarily discuss his upbringing, family influences, and career decisions.
AI is still in its very early stages, which is one of Sanjay's core judgments.
In his view, as large models, Agent AI, and inference applications evolve, AI needs not only more powerful computing but also stronger "memory capabilities."
Longer context windows, larger model sizes, and ever-increasing token consumption are all driving a continuous climb in storage demand.
The essence of AI is data, and data cannot exist without storage. Therefore, storage will become one of the most critical infrastructures in advancing AI capabilities.
At the same time, the supply side is not fully prepared. Sanjay pointed out that the storage industry faces not a short-term supply-demand mismatch but a structural supply constraint. Advanced storage products consume more wafers, and building new fabs typically takes three to four years, followed by a long capacity ramp-up period.
More importantly, as technology nodes advance, the increase in storage capacity output per wafer is diminishing. He predicts the industry's tight supply condition is likely to persist beyond 2026.
Explaining why storage technology has long been undervalued, Sanjay said bluntly: "People often misunderstand memory; they don't know how hard it is to make memory." The technical challenges are immense, spanning physics, chemistry, and materials science, to ensuring every single one of trillions of bits behaves correctly in mass production. He believes the AI race is equally a storage race, a point long overlooked by the market.
From a longer-term perspective, Sanjay believes the fundamental logic for success for both companies and individuals remains unchanged. Whether driving a $200 billion investment plan or leading Micron through storage industry cycles, the keywords he repeatedly emphasizes are resilience, discipline, and long-termism. Investments must be based on data and fundamentals. Leaders must see industry trends clearly and also deeply understand technical details.
Just as he learned from his father, success requires both the resilience to persist and the ability to seize opportunities at critical moments.
Core points from the interview with Sanjay Mehrotra, CEO of Micron Technology:
Storage is an underestimated bottleneck for AI; its manufacturing difficulty and strategic value far exceed market recognition. AI is extending from a "computing power race" to a "storage race." Expanding model sizes, longer context windows, and surging token consumption mean AI relies not only on stronger computing power but also on stronger "memory capabilities." Without sufficient storage capacity and bandwidth, even the strongest computing power cannot be unleashed.
Structural constraints on the supply side mean the storage shortage is not a short-term fluctuation but a long-term condition. Advanced storage products consume more wafers, while building new fabs takes three to four years, with equally long capacity ramps. Meanwhile, advancing technology nodes reduce the output gain per wafer. With supply-demand mismatch, tight supply will last at least beyond 2026.
People always underestimate the difficulty of manufacturing memory, but this is precisely the industry's deepest moat. The engineering complexity is extremely high—from physics, chemistry, and materials science to ensuring trillions of bits are error-free in mass production from design to manufacturing. Making memory chips is as difficult as any semiconductor field, and in many aspects, even harder.
Success comes from resilience, discipline, and long-termism, not short-term trend spotting. Whether driving a $200 billion investment or navigating the cyclical volatility of the storage industry, leaders need to see industry trends clearly and dive deep into technical details. Just as his father did not give up after his visa was rejected three times, success requires both the resilience to persist to the end and the ability to seize opportunities at key moments.
Storage is Becoming the Backbone of AI
Speaking about the storage industry's current historical position, Sanjay said bluntly: "I've been in this industry for over 45 years. This is the most exciting moment I've experienced for the entire industry."
He further explained the strategic importance of storage for AI:
"Without semiconductors, there is no AI. And storage is the backbone of AI, the key foundation supporting AI's continuous evolution."
In his view, storage's role is no longer just a component in devices but directly carries "intelligence" itself: "Today, storage isn't just making devices run; it is supporting the 'intelligence' within AI itself, helping artificial intelligence become smarter."
As model sizes expand, inference demand explodes, and agent AI rapidly emerges, the logic behind storage demand growth is clear to Sanjay: "As models get larger, as inference demand continues to grow, as AI moves from training to inference, from data centers to the edge, demand for storage will only increase—it needs greater capacity, higher performance, and lower power consumption."
He also specifically mentioned token economics' reliance on storage: "When you look at token economics, it also heavily depends on storage. As token usage grows, context windows become longer, KV cache demand increases, and models themselves become larger. AI needs not just computational ability, but also the ability to 'remember.'"
Tight Supply to Persist Beyond 2026
Regarding the market's most concerned supply-demand issue, Sanjay gave a clear judgment: The industry's tight supply will persist beyond 2026 and continue for a considerable period.
He explained the structural constraints on the supply side: "It takes a long time to build a fab. From groundbreaking to the first wafer out, it typically takes three to four years. After that, there's continued ramp-up to gradually increase the output."
More critically, rising technical difficulty is compressing output efficiency per wafer: "The productivity improvement from each new generation of technology, the bit increase each wafer can bring, is becoming less."
Sanjay revealed that Micron anticipated this trend as early as around 2021.
At that time, High Bandwidth Memory (HBM) accounted for less than 1% of the entire storage industry, but they already saw that future generations of HBM would consume massive amounts of silicon and significantly impact the supply landscape: "So as early as 2021, we said the industry needed to build new fabs from scratch. It's just that no one truly predicted AI would explode at such a speed."
Regarding market concerns about "overcapacity returning once supply catches up," Sanjay did not directly dismiss them, but he emphasized that AI is still in its early stages, and long-term structural demand growth is his basis for confidence: "From the demand side, all of this is still at a very, very early stage. We believe AI still has a long way to go."
The Underlying Logic of the $200 Billion Investment: Discipline
Micron's announcement of a $200 billion investment in building a storage manufacturing system in the U.S. is one of the most notable capital decisions in the semiconductor industry in recent years. Regarding the underlying logic of this decision, Sanjay repeatedly emphasized the word "discipline":
"Investments are never made blindly; they must be disciplined and based on data. You need to understand the technology, understand the applications, understand where those applications are going. You also need to work closely with customers to understand where they are headed in the future and what role Micron will play in that."
He further explained the discipline in execution: "Today, we are investing in building several new fabs from scratch. The first step is constructing the building and infrastructure. Once those buildings are complete, we will still maintain discipline when installing equipment and forming actual capacity—continuously evaluating demand forecasts, assessing how much growth technological progress can bring, evaluating how product demand will change."
When asked if he ever had self-doubt, Sanjay's reply was straightforward:
"We have no self-doubt. We absolutely believe in the opportunity for storage; today, that is very clear. Of course, in our business, it's always important to maintain adaptability and agility."
