The more powerful the AI model, the more sought-after the storage device becomes? OpenAI Astra reignites the "storage chip trade."

The release of OpenAI's latest flagship model, Astra, has reignited market enthusiasm for AI computing infrastructure, with memory chips being one of the first sectors to benefit. Last Friday, US semiconductor stocks rallied collectively, and this upward trend spread further to Asian markets on Monday, with leading memory chip companies such as Samsung Electronics, SK Hynix, and Kioxia Holdings showing significant gains.

ChatGPT-6 Astra was launched in two batches last Thursday and Friday. OpenAI President Greg Brockman called it "the world's smartest and most aligned model." As Astra is designed for more complex AI tasks, the market is further betting that the continuous iteration of cutting-edge models will drive up computing power demand and support the demand for core hardware such as HBM, DRAM, and GPUs.

This expectation was quickly reflected in stock prices. Micron Technology rose more than 6% on Friday, SanDisk rose nearly 12%, and SK Hynix Nasdaq Depositary Receipts rose more than 8%. On Monday, Asian markets continued the upward trend, with SK Hynix rising about 8%, Samsung Electronics rising more than 5%, and Kioxia rising about 10%.

The institution believes that as AI model capabilities continue to improve, the demand for computing power is still expanding rapidly, and bottlenecks in storage and network may become more prominent. "Storage chip trading" is expected to become an important theme in AI infrastructure trading again.

AI computing power expansion drives memory chips back into the spotlight.

The core of the rally driven by Astra is not just the release of a single model, but rather the market's renewed awareness that as AI model capabilities continue to improve, they will still correspond to higher demands for computing power and storage.

Charu Channa, chief investment strategist at Saxo Bank, pointed out that Astra's launch further supports the continued growth in AI spending. As cutting-edge models continue to increase their demand for computing power, bottlenecks in storage and networking are also widening, which will benefit memory chip manufacturers such as SK Hynix and Samsung Electronics.

Analysts at Goldman Sachs and Morgan Stanley predict that global AI capital expenditure will reach $1.3 trillion to $1.5 trillion in 2027, with more than half going towards memory infrastructure. If this trend continues, storage products such as HBM and DRAM are expected to continue to benefit from the expansion of AI computing power.

Meanwhile, Goldman Sachs' chief Asia equity strategist, Tim Moe, reiterated his year-end target of 12,000 points for the South Korean stock index, noting that earnings expectations for the South Korean market are still being revised upwards, with a 3.2% increase this week. He also emphasized that foreign investor ownership of South Korea's semiconductor sector remains significantly below historical averages.

Previous discussions about the semiconductor "supercycle" focused more on whether AI capital expenditures could be sustained and whether chip manufacturers could maintain high growth and high profit margins. Now, with the continuous increase in computing power demand driven by next-generation AI models, storage is once again becoming a key beneficiary of AI infrastructure transactions.

The more powerful the model, the higher the computing power required, and the greater the demand for storage. Astra's launch is reigniting this trading logic.

Risk Warning and DisclaimerInvesting involves risk; please exercise caution. This article does not constitute personal investment advice and does not take into account the specific investment objectives, financial situation, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article are suitable for their specific circumstances. Any investment decisions made based on this information are at your own risk.