Memory prices continue to rise, and the supply of these devices to the consumer market remain tight. But astonishingly, an SK Group Chairman was recently quoted as saying these prices are abnormally high, suggesting that the market may soon face adjustments.
SK Group Chairman Says Memory Prices “Abnormally High”
Recently, SK Group chairman Chey Tae-won was interviewed by Bloomberg, where he was asked about the current situation surrounding memory semiconductors, and according to Chey, memory semiconductor prices are currently “abnormally high”.
The rising cost of memory is already creating challenges for many manufacturers of electronic devices, with PC makers and smartphone developers both experiencing pressure from increasing semiconductor costs. As a result, many companies may have no choice but to pass these additional costs onto consumers through higher product prices.
However, Chey also mentioned that SK Group is considering constructing a memory semiconductor plant in the United States, which would help increase supply and provide greater stability within the memory market.
At the same time, Chey warned that high memory prices could encourage new competitors to enter the industry. This is something that has happened before, particularly during the DRAM price competition between South Korea, Japan, and the United States in the early 2000s.
In recent years, several companies have already begun sourcing memory from Chinese manufacturers. Companies such as Lenovo and Corsair have explored alternative suppliers, while Apple is reportedly looking towards Chinese memory manufacturers as existing suppliers such as SK hynix struggle to meet demand.
One Chinese manufacturer gaining attention is ChangXin Memory Technologies (CXMT), which has started shipping memory products to international markets. Another is Yangtze Memory Technologies (YMTC), which is looking to expand further into mobile memory solutions.
Furthermore, Chey also noted that companies such as Elon Musk’s SpaceX could become future competitors to SK hynix, following Musk’s announcement that he intends to establish semiconductor manufacturing capabilities in Texas, although such a project has yet to materialise.
Are We Headed for a Disaster?
While memory prices naturally fluctuate over time, the current situation is very different from previous market cycles. The sudden rise in demand for AI hardware and services has become the primary driver behind the increase, with data centres around the world placing enormous orders for memory components.
This sudden demand has effectively placed consumers at the mercy of the market. While this type of boom is common in industrial equipment and heavy machinery, it is far less normal for consumer electronics, where millions of people rely on affordable memory for everyday devices.
Normally, when a market experiences a significant increase in demand, manufacturers respond by constructing new production facilities and increasing capacity. Over time, this establishes a new baseline for supply and demand.
However, the current memory cycle is different.
Instead of demand being spread evenly across different types of memory, the majority of investment is focused on high-bandwidth memory (HBM) used in AI accelerators. Manufacturers are prioritising these products because AI companies are willing to pay significantly higher prices for them.
The problem is that if the AI market turns out to be a bubble, the consequences could be severe. When such a bubble bursts, the excess memory supply would not simply move into consumer products. The issue is that HBM memory is highly specialised and cannot easily be converted into conventional consumer memory such as DDR RAM.
This means manufacturers could suddenly find themselves with enormous amounts of expensive HBM production capacity and very few customers willing to purchase it. The result could be massive financial losses and potentially a significant reduction in memory manufacturing capacity.
Ironically, this could make consumer memory even harder to produce. If memory manufacturers suffer large enough losses from the collapse of AI demand, they may be forced to scale back operations, making affordable RAM production even more difficult.
In the worst-case scenario, major memory suppliers could face financial difficulties, while the AI companies that drove the demand could walk away having benefited from inflated prices without facing the same consequences.
This risk becomes even more concerning when considering companies such as OpenAI, which have publicly discussed purchasing enormous quantities of computing hardware while also facing challenges around long-term profitability and meeting financial obligations.
So, are we heading towards a disaster? Will memory prices continue to rise? Will enough memory be manufactured for consumers? And will memory companies survive if the AI market experiences a major correction?
Only time will tell.