The memory market is in a state of flux, with investors and analysts alike grappling with the implications of China's rise in DRAM and fears of a slowing AI investment. The recent selloff in memory stocks, led by SK Hynix, has sparked a debate about whether this is a timing issue or a genuine collapse in demand. In my opinion, the answer lies in a nuanced understanding of the market dynamics and the unique pricing behavior of AI memory.
One thing that immediately stands out is the role of pricing in this narrative. The selloff in SK Hynix shares was triggered by a revision in earnings forecasts, which investors interpreted as a sign of fading AI memory demand. However, as Tessara research points out, this may be a case of misreading the market. The key issue here is not demand, but rather how quickly SK Hynix recognizes higher DRAM prices under long-term contracts. This is a critical distinction, as it highlights the unique pricing behavior of AI memory, which is negotiated over much longer periods than traditional commodity memory.
From my perspective, the physical data also tells a different story. Korean customs data shows that memory exports rose significantly in the second quarter, with commodity DRAM's implied price per kilogram increasing by 58%. This suggests that end-market demand remains strong, despite the market's bearish interpretation. However, it is important to note that this demand is concentrated in commodity memory, rather than the HBM chips that power AI infrastructure.
What makes this particularly fascinating is the structural transformation of the memory industry. As Yuri Khodjamirian, chief investment officer at Tema ETFs, points out, supply remains constrained well into 2028. This means that even as demand for AI memory grows, the industry may struggle to keep up, leading to potential pricing pressures. In my opinion, this raises a deeper question about the sustainability of the current market dynamics, and the potential for a broader AI memory slowdown.
The earnings report from SK Hynix this week could settle the debate. If the company confirms that AI demand remains robust and pricing is simply flowing through financial statements more slowly than investors expected, then the market may have spent the past month selling memory stocks for the wrong reason. However, if the earnings report reveals a more nuanced picture, with pricing pressures and demand concerns, then the market's fears about a broader AI memory slowdown could gain credibility. In any case, the report will provide valuable insights into the current state of the market and the future trajectory of the memory industry.
In conclusion, the memory market is a complex and dynamic space, with a unique set of pricing behaviors and market dynamics. As an investor or analyst, it is crucial to understand these nuances to make informed decisions. Personally, I believe that the earnings report from SK Hynix will provide a critical insight into the current state of the market, and the potential for a broader AI memory slowdown. However, the market's fears may be overblown, and the report could reveal a more nuanced picture of the industry's current health and future prospects.
