"China's DRAM chipmaker surges as AI memory demand reshapes semiconductor geopolitics"
DRAM (Dynamic Random Access Memory)
volatile memory chip used in all computing devices for active data processing.
HBM (High-Bandwidth Memory)
vertically stacked DRAM placed directly alongside AI processors for ultra-fast data access.
The global semiconductor industry has undergone a structural transformation driven by the rapid scaling of artificial intelligence workloads. For decades, memory chips — particularly Dynamic Random Access Memory (DRAM) — were treated as low-margin, commoditised components. The competitive landscape was dominated by a tight oligopoly: South Korea's Samsung and SK Hynix, alongside America's Micron Technology, collectively control roughly 90% of global DRAM supply. This concentration, combined with the capital-intensive nature of chip fabrication, created formidable barriers to entry for new players.
China's semiconductor ambitions have been a defining feature of its industrial policy for over a decade. Recognising that dependence on foreign chips represented a strategic vulnerability, Beijing channelled an estimated $150 billion in state-led funding into domestic semiconductor development. Yet progress was uneven: China's leading foundry, SMIC (Semiconductor Manufacturing International Corporation), remains constrained at the 7-nanometer node because Western export controls — particularly the denial of ASML's Extreme Ultraviolet (EUV) lithography machines — prevent it from advancing to finer process nodes. Memory manufacturing, however, presented a somewhat lower barrier to entry than cutting-edge logic chip fabrication, making it a more tractable target for domestic champions.
China's domestic semiconductor self-sufficiency reportedly rose from approximately 16% to 28% between 2024 and late 2025, reflecting the scale of state investment. The government's stated ambition is to reach 80% self-sufficiency by 2030 — a target that would require breakthroughs across the entire semiconductor value chain.
The semiconductor ecosystem can be understood through five distinct segments:
Memory chips occupy a unique position in the AI era. An advanced AI processor is functionally limited by how quickly it can access and process data — a bottleneck addressed by High-Bandwidth Memory (HBM), a specialised architecture where memory chips are stacked vertically and placed directly alongside the processor. …
Which one of the following pairs of semiconductor plants in India and their locations is not correctly matched?
Consider the following statements:
I.It is expected that Majorana 1 chip will enable quantum computing.
II.Majorana 1 chip has been introduced by Amazon Web Services (AWS).
III.Deep learning is a subset of machine learning.
Which of the statements given above are correct?
Consider the following statements regarding the global semiconductor industry:
1.High-Bandwidth Memory (HBM) is a type of DRAM stacked vertically to work alongside AI processors.
2.ASML's Extreme Ultraviolet (EUV) lithography machines are currently exported to China under a special licence regime.
3.Samsung, SK Hynix, and Micron together control approximately 90% of global DRAM supply.
Which of the statements given above is/are correct?
Which of the following correctly describes the role of a 'fabless' semiconductor company?
With reference to China's semiconductor self-sufficiency drive, consider the following:
1.China's leading foundry SMIC has successfully achieved 5-nanometer chip production.
2.China has invested an estimated $150 billion in state-led semiconductor funding.
3.China's stated target is to achieve 80% domestic semiconductor self-sufficiency by 2030.
Which of the statements given above is/are correct?
The term 'RAMageddon', as used in the context of global technology markets, refers to:
Want the full note, Prelims Facts & MCQ answers?
Create a free account — takes less than a minute.
Create free account