Micron Technology (MU) The Six Five Summit: AI Unleashed 2026 summary

Strategic importance of memory in AI

• Memory has become a critical enabler for AI, with demand outpacing supply across all market segments due to escalating performance requirements and the need for differentiated system designs.

• Customers are embedding memory planning into multi-year roadmaps, seeking competitive advantage through co-design and custom capabilities, as seen in collaborations like LPDRAM for data centers.

• The shift from commoditized, off-the-shelf memory to strategic, differentiated solutions is driving new business models, including multi-year strategic customer agreements with committed supply and demand.

• AI's proliferation from data centers to edge devices, autonomous vehicles, and robotics is fundamentally changing memory requirements, emphasizing low power, high performance, and on-device processing.

• Robotics and next-generation devices are expected to drive massive growth in DRAM and NAND demand, with each humanoid device requiring significant memory content.

Industry transformation and investment

• The memory industry is undergoing a transformation, with greenfield capacity expansion required to meet AI-driven demand, as technology transitions alone are insufficient.

• Large-scale investments are underway globally, including $250 billion in the U.S., with new fabs in Idaho, New York, Taiwan, Japan, and Singapore, aiming for meaningful supply ramping by 2028.

• Building new fabs is a multi-year, capital-intensive process, constrained by construction talent, regulatory approvals, and the need for skilled technicians.

• Investments in workforce development and community partnerships are essential to support the scale and complexity of future memory manufacturing.

• The industry expects continued expansion and clustering of fabs over the next decade to achieve cost efficiencies and meet demand.

Technological complexity and innovation

• Memory manufacturing is highly complex, involving thousands of process steps, EUV technology, and long cycle times, with products like HBM and high-layer NAND representing engineering marvels.

• Co-design with partners is becoming the norm, with deep, ASIC-like relationships enabling differentiated products and long-term R&D collaboration.

• Innovations span from ultra-low-power devices to gigawatt-scale data centers, requiring breakthroughs in packaging, thermal management, and process scaling.

• The pace of innovation and investment is crucial to sustaining the cycle of technological advancement and meeting the evolving needs of AI and emerging markets.

• Robotics and physical AI are anticipated as the next major growth drivers, compounding demand for advanced memory solutions.