Memory and storage will play an increasingly critical role in scaling artificial intelligence, as the performance of advanced AI systems depends heavily on how quickly data moves between processors and memory, according to Anand Ramamoorthy, Managing Director of Micron India.
Speaking at the inaugural plenary session of BTS 2026, Ramamoorthy noted that the growing complexity of AI models, larger context windows, and rising user volumes will drive sustained demand for greater memory capacity and bandwidth.
“The world’s most powerful AI systems today are not necessarily constrained by compute alone but by how fast the memory can feed that compute,” Ramamoorthy said.
He explained that data movement bottlenecks between computing processors and memory create latency, limiting the performance gains promised by advancements in computing hardware. He referred to this industry challenge as the ‘memory wall,’ where high computing power is underutilized due to slow data transaction speeds between the compute unit and memory.
Ramamoorthy stated that demand for memory will increase across all stages of AI development, ranging from model training and inference to agentic and physical AI applications. Highlighting high-bandwidth memory (HBM) as a foundational innovation for the ecosystem, he explained that the technology accelerates data transactions through the stacking of multiple DRAM dies placed alongside GPUs or AI CPUs. He described HBM as the ‘crown jewel of AI memory,’ adding that energy efficiency will grow equally important as workloads expand and consume more power.
On the domestic front, Ramamoorthy highlighted India's expanding contribution to Micron's global operations, particularly in memory engineering and semiconductor manufacturing. Micron established its presence in Bengaluru approximately eight years ago and has since grown its local workforce to over 8,000 employees engaged in design and engineering.
“Every data center you see in the world has a lot of Bengaluru engineering as a part of it,” he said, pointing out that the company has also initiated semiconductor back-end manufacturing operations in the country as the first multinational to do so.
Ramamoorthy attributed these operational expansions to India's available engineering talent, the local startup ecosystem, and supportive government policies, reinforcing Bengaluru's position as a key hub for the company's global technology development.
"The semiconductor and memory sector is foundational to the rapid commercialization of artificial intelligence. Micron India's focus on high-bandwidth memory addresses the core bottleneck of data latency that hardware developers face globally. Furthermore, the expansion of advanced engineering and manufacturing operations in India highlights the country's rising capability to support deep-tech supply chains. For businesses and startups building AI infrastructure, understanding these hardware limitations and talent ecosystems will be critical for long-term scalability and operational efficiency." — Dr. Shishir Gupta, Founder & CEO, StartupLanes
Recent StartupLanes Articles
Browse through our 30 latest publications on venture capital, startups, and angel investing.