Artificial intelligence may have made many things easier now. You can write code in seconds, create presentations straight away or generate images. But all this comes at cost, and no we are not talking about the premium subscription prices. Rather, the increased demand for AI has led to a rise in prices for memory chips. While AI giants like OpenAI and Anthropic are looking to buy these chips to run their AI models in datacentres, consumer brands such as phonemakers are struggling to get these memory chips. This has led to an increase in the prices of various products, including smartphones and laptops. And now, a new report indicates that this chip shortage may not stop anytime soon.
As per a report from Nikkei Asia, chipmakers will not be able to increase capacity to fulfill demand at least till 2027. Globally, three suppliers enjoy the majority of memory chip market share – Samsung, SK Hynix, and Micron.
Memory chips are essential components in a wide range of products, from PCs and smartphones to automobiles and industrial equipment.
What is the memory chip situation?
The three companies collectively account for about 90 per cent of the global DRAM market and have shifted their focus towards high-bandwidth memory (HBM), crucial for AI semiconductors.
This shift has led to a delay in the production of general-purpose memory used in personal computers and smartphones, resulting in supply shortages since the fall of 2025, with a steady increase in prices of consumer goods.
As per the report, the three giants will only be able to meet 60 per cent of DRAM demand despite having plans to increase production. Do note that Micron has previously decided to shut down its consumer business entirely to focus on sales for AI datacentres.
The report states that the ongoing turmoil in the Middle East has further complicated matters when it comes to the chip shortage. The US’ strikes on Iran has driven up the cost of electricity and materials, thereby increasing uncertainty in production forecasts.
Increased production not enough
Samsung, the world’s largest memory chip maker, plans to bring a new fabrication plant at its Pyeongtaek campus online by 2026, but full-scale production will not commence until 2027 or later. This facility will also produce logic chips, that is processors used in devices like your smartphone, which will constrain its capacity to boost memory output.
Meanwhile, SK Hynix has started operating an HBM fab in Cheongju, marking the only increase in production among the top three memory suppliers for 2026. The company is also expediting the construction of another facility in Yongin, with completion anticipated in early 2027.
Micron plans to begin HBM production in Idaho and Singapore by 2027 and will start building a fab in Hiroshima, Japan, in May, with mass production slated for 2028.
Is there any hope?
Despite these efforts, SK Group Chairman Chey Tae-won has indicated that supply constraints in AI memory could continue into 2030, citing a wafer shortage and other production challenges.
Counterpoint Research estimates that resolving the memory shortage would require a 12 per cent annual increase in production through 2027, but current plans only foresee a 7.5 per cent expansion. The supply-demand balance is not expected to normalise until 2028.
Smartphone production may fall due to chip shortage
The rising demand for AI-related memory and the slow pace of new capacity coming online contribute to the uncertainty about when the market will stabilise. Although cutting-edge memory fabs are being developed, achieving high production yields will take time. Meanwhile, smartphone sales are projected to decline by 13 per cent in 2026, according to IDC, as memory costs rise.
That is, some consumers may not be able to afford smartphones due to the rising prices.
The share of memory costs in manufacturing low-priced smartphones is expected to increase to nearly 40 per cent by mid-2026, potentially leading to reduced production by smartphone makers.
