Micron CEO: AI demands 50% more memory capacity; company invests billions in US expansion
Translated from Chinese, summarized and contextualized by DistantNews.
At a glance
- Micron CEO Sanjay Mehrotra stated that AI systems critically need higher performance, lower power consumption, and larger capacity memory, declaring "there is no AI without memory."
- The company faces a 50% production capacity gap, with demand from data centers exceeding current supply by 150% as tech giants scramble for AI computing power.
- Micron is investing $250 billion in U.S. manufacturing and R&D, including two massive fabrication plants near its Boise headquarters, with the first expected to start production by mid-2027.
The global semiconductor market is undergoing a historic transformation, with artificial intelligence (AI) completely reshaping the memory industry's boom-and-bust cycles. Micron Technology CEO Sanjay Mehrotra declared that AI systems are in desperate need of higher performance, lower power consumption, and larger capacity memory, emphasizing, "there is no AI without memory!" This critical technology has become an indispensable strategic infrastructure for the AI era, redefining the value of memory chips in the entire tech supply chain.
Mehrotra revealed in an interview that Micron's production capacity is already sold out, with terminal market customers acquiring almost all manufactured products. Demand, particularly from data centers, exceeds Micron's current supply commitments by approximately 50%. This means market demand is 150% of actual supply, highlighting the intense hunger among major tech giants for memory hardware as they vie for AI computing dominance.
In response to this unprecedented market demand, Micron is making a significant bet on the future with an astonishing expansion plan. This initiative is part of Micron's massive $250 billion investment in U.S. manufacturing and research and development. A particularly noteworthy project is the large-scale fabrication plant being built near its headquarters in Boise, Idaho. This facility will eventually house two giant wafer fabrication plants, each covering an area equivalent to 10 football fields. The amount of steel required to construct just one of these plants would be enough to circle the Earth twice. The first plant is projected to begin production by mid-2027, demonstrating strong management confidence in the long-term, stable demand driven by AI.
Beyond production capacity and investment scale, this revolution is profoundly altering the memory industry's business models and ecosystem. Previously, memory was treated as a standardized commodity, with customers often comparing prices and choosing the lowest bidder, leaving manufacturers vulnerable to sharp fluctuations in the spot market. Now, customers are shifting towards "co-design" with Micron, participating in product development from the earliest stages to deeply optimize memory with processors and the overall system. Memory is no longer a simple plug-and-play component but a core element determining the performance of the entire AI system.
To ensure these partnerships and stabilize future expectations, Micron is securing orders through long-term contracts. Following the announcement of five-year strategic agreements with 16 major clients during a previous earnings call, Micron has since reached similar collaborations with more customers. These long-term contracts allow clients to lock in production capacity in advance, providing Micron with clear and guaranteed demand visibility, significantly reducing the risks associated with traditional cyclical markets. Mehrotra also emphasized that this strong memory wave is not limited to data centers. With the future proliferation of autonomous vehicles, robots, and various AI consumer devices, immense memory demand is expected to erupt across a wide range of end-use sectors, ushering in a new golden age for the memory industry.
Originally published by Liberty Times in Chinese. Translated, summarized, and contextualized by our editorial team with added local perspective. Read our editorial standards.