Nvidia and other firms have raised AI chip prices as surging demand for artificial intelligence hardware pushes up memory costs in what is being described in the media as ‘RAMageddon’ or ‘RAMpocalypse’. Memory costs are now squeezing Indian IT firms, data-centre operators and electronics makers. A global shortage of DRAM or Dynamic Random-Access Memory and high-bandwidth memory (HBM) is deepening. The memory crisis that began in 2025 intensified in 2026. Memory prices rose sharply as chipmakers redirected capacity towards more profitable AI-related products.
AI turns up the heat: Rise of HBM chips
High-bandwidth memory (HBM) multi-stack chips are fast gaining ground over traditional RAM in AI systems. HBM is estimated to account for more than a fifth of DRAM wafer input among major suppliers. Producing the equivalent amount of HBM memory can require roughly three to four times the wafer capacity used for conventional DDR5 chips.
Hyperscalers including Microsoft, Google, Amazon and Meta are securing large volumes of HBM and server DRAM for AI training and inference. Such long-term supply agreements and open-ended orders have led to a squeeze in memory availability.
Memory makers like Samsung, SK Hynix and Micron have shifted capacity towards HBM and high-capacity server memory. These three companies together control about 95 per cent of the DRAM market. Their shift in priorities has left less capacity for PCs, smartphones, gaming consoles and conventional enterprise hardware.
Trending Stories
Prices are surging for memory
Conventional DRAM contract prices jumped by about 90-98 per cent quarter-on-quarter in the first quarter of 2026. Overall DRAM and NAND prices have risen by 200-400 per cent in some segments since the shortage intensified.
For instance, retail DDR5 memory kits that used to sell for about $80-250 in mid-2025 are now worth $300-650 or more. Server DRAM lead times have also exceeded 40 weeks.
The supply squeeze is unlikely to disappear quickly
Chip fabrication plants can cost up to $20 billion and take years to reach full production. This is the reson why US firms are locking in supplies, signing multi-year contracts and passing on higher costs to consumers by raising prices.
Besides Nvidia raising prices for AI chips, firms like Apple, Microsoft, Dell, Lenovo, HP and HPE have increased prices on their phones, computers, gaming consoles and servers by up to 15 per cent. Others that have recently raised prices include Asus and Cisco.
Samsung, SK Hynix and Micron have also announced substantial increases for DRAM and HBM as they prioritise higher-margin products.
Indian IT sector feels the squeeze of memory crisis
Indian technology companies are facing the impact in the form of costlier hardware, data centres and infrastructure. They are trying to adjust to the situation with various strategies. For instance, Tech Mahindra avoided guaranteeing fixed memory and chip prices to customers. HCL Technologies is investing $350 million in data-centre capacity.
The consumer market is feeling the pressure too. Smartphone shipments in India fell by about 10 per cent year-on-year in the second quarter of 2026. The sub-Rs 20,000 segment is particularly hit, due to the squeeze in hardware availability.
Also read: Anthropic has stacked $71 billion in chip-lease debt in 60 days and kept it off its books
India is also trying to reduce its dependence on imported chips and chip components. The India Semiconductor Mission is backing further investment in semiconductors and memory-related infrastructure. Micron has begun commercial production at its assembly-and-test facility in Gujarat. Projected investment in India's data-centre ecosystem has crossed $200 billion.
For Indian IT firms, the challenge is not limited to the price of a memory chip. The rising cost is hitting servers, data centres, smartphones and AI infrastructure as demand for computing capacity soars.

&imwidth=800&imheight=600&format=webp&quality=medium)
&im=FitAndFill=(700,400))
)
)
)
)
&im=FitAndFill=(700,400))
)
)
)
)
)
)
)
)
)
)
)
&im=FitAndFill=(700,400))
)
)
)
)
)
&im=FitAndFill=(700,400))
)
)
)
)
)
&im=FitAndFill=(700,400))
)
)
)




