Storage super-cycle bottoming out, Korean won poised for appreciation, "Storage chip duopoly" continues to play the mad bull market symphony in South Korea.

A senior trader at Wall Street financial giant State Street, with over 25 years of experience, stated that under the backdrop of the global fund flow into the South Korean stock market as a super bull market amidst the "storage super cycle" attracting global funds, and the continuous retreat of the pressure from foreign investors selling South Korean local stocks in the presence of the massive increase in global storage chip demand supporting the South Korean exchange rate due to the persistent current account surpluses, the South Korean won may see a significant rebound in the second half of the year. The head of the foreign exchange trading and sales department at the Seoul branch of State Street, Cho Ji-wang, stated in a media interview that if global risk sentiment/ preference continues to improve and foreign investors massively return to South Korean stock assets, the exchange rate between the won and the dollar may rise to over 1400 won per dollar by the end of the year. On Tuesday, the won exchange rate in the foreign exchange market was trading around 1517 won for 1 USD. If the won really follows the prediction of the senior trader at State Street, rising from around 1517 won/USD to over 1400 won/USD in the second half of the year, it will be even more favorable for the South Korean Kospi composite index, which has surged more than 100% this year. It also means the bull market trajectory of the two major storage chip giantsSK Hynix and Samsung Electronicswhich together account for over 50% of the South Korean benchmark index, is far from over. The continuous strong uptrend of these companies is a significant positive catalyst for the global storage sector and even the AI computing power industry chain. If the appreciation comes from the return of foreign capital, current account surpluses, the reduction of risk premiums on South Korean assets, and the potential increase in expectations of interest rate hikes by the Bank of Korea, it reflects not "damages to export competitiveness," but the continuous repair of credit sentiment towards South Korean assets. State Street's won appreciation logic is based on the gradual disappearance of foreign selling pressure, the continuation of current account surpluses, and the reconfiguration of South Korean assets, all of which will jointly promote a strong rebound of the won in the second half of the year. Morgan Stanley, the Wall Street financial giant, has significantly raised the target level of the South Korean stock market benchmark indexKospi composite indexwithin less than a month, which undoubtedly reflects its belief that the bull market trend under the AI infrastructure frenzy amidst the "storage super cycle" background is far from ending. This largest commercial bank on Wall Street has raised the target for the South Korean Kospi composite index benchmark to 9,000 points and the bullish target to 10,000 points, a historical milestone. In comparison, Morgan Stanley's previous benchmark and bull market targets at the end of April were 7,000 points and 8,500 points respectively. The opening of a US dollar-won spot exchange trading for 24 hours starting from July 6 will be a critical test for the South Korean foreign exchange market. This is one of the key measures taken by the country to improve market access and strengthen its efforts to gain a position in developed markets. Cho Ji-wang stated that the extended trading hours not only mean longer trading periods but also mark an important step towards the wons evolution into a more internationally traded currency. The ultimate test will be on liquidity. Cho Ji-wang said, "If there's ample liquidity round the clock and pricing is tighter and more rigorous than in the non-deliverable forward forex market, then this market can succeed in expanding." "If this happens, the entire South Korean market could become significantly larger than we remember from the past." Non-deliverable forward, or NDF market allows offshore investors to hold won exposures without converting the currency. Instead, profits and losses are settled in US dollars upon expiration. During the US summer time, trading hours will be extended from the current 9 am to 2 am the next day, to 6 am on Monday to 6 am on Saturday. Trial trading will begin on June 29. With the rebound of the won combined with the storage super cycle, the South Korean Kospi index and the two major storage chip giants are entering a moment of stock and foreign exchange resonance. As mentioned above, if the won appreciation comes from the return of foreign capital, current account surpluses, the reduction of risk premiums on South Korean assets, and the potential increase in expectations of interest rate hikes by the Bank of Korea, it reflects not "damages to export competitiveness," but the continuous improvement of credit sentiment towards South Korean assets, which is considered a continuous positive catalyst for the two major storage chip giants and the South Korean stock market. Following Samsung Electronics entering the trillion-dollar market value club, another South Korean storage giant SK Hynix's market value recently exceeded one trillion dollars. They have jointly driven the Kospi index to set new highs, supported by the almost endless demand for storage and the expectation of storage chip price increases driven by the AI infrastructure frenzy. For Samsung and SK Hynix, the short-term impact of the won appreciation may be relatively negative on profit conversion, but it is not enough to reverse the current bull market trend driven by the AI infrastructure frenzy amidst the "storage super cycle" backdrop. Demand for storage chips, especially HBM, server DRAM, DDR5, and enterprise-grade NAND storage components, is being pushed up by the near-endless demand from AI data centers in North America, long-term supply agreements, production capacity bottlenecks, and long-term customer lock-ins. In this environment, the exponential demand expansion, continuous price increases, and product price upgrades driving profits are often much greater than the conversion pressure of the exchange rate from 1517 to over 1400. If the stock price rise is mainly driven by the AI storage shortage, long-term contracts for HBM, and profit upgrades, the profit pressure brought by the won appreciation is just marginal noise. Whether it is Google's massive TPU AI computing cluster, or Nvidia's AI GPU computing power, they both rely on HBM storage systems that integrate AI chips comprehensively. The acceleration of building or expanding AI data centers by tech giants requires a massive purchase of server-grade DDR5 storage and enterprise-grade high-performance SSD/HDD, and Samsung Electronics, SK Hynix, and Micron are positioned in these three core storage sectors: HBM, server high-performance DRAM (including DDR5/LPDDR5X), and high-end data center-level SSD. They are the most direct beneficiaries in the "AI Memory + Storage Stack", enjoying the "super dividends" of the AI infrastructure frenzy. GPUs are responsible for generating intelligence, while HBM/DRAM provides high-speed transactions and enterprise-level NAND/eSSD manages hot data and caching, and HDD stores massive levels of cold/warm data for long-term retention. Therefore, Goldman Sachs believes that the AI computing arms race led by cloud computing giants is shaping storage chips from cyclical products to scarce strategic assets. The increase in the prices of DRAM/NAND by 2026 is not the end but could be just the beginning of the super cycle. Transitioning from cyclical commodities to strategic core assets, the shortage of storage chips may continue until 2028, reshaping stock price and profit logic through long-term supply agreements Under the eruption of AI computing demand, the fundamentals of the storage chip industry are undergoing unprecedented structural reshaping. The traditional cyclical storage market was dominated by consumer electronics and inventory cycles, with prices and demand highly volatile; whereas now, the persistent and near-endless demand for high-bandwidth memory (HBM), server DRAM, and enterprise-grade SSD due to AI training and inference has positioned storage as a strategic core asset of AI infrastructure, with a forecasted Total Addressable Market (TAM) of around $1.7 trillion by 2028, indicating a fundamental shift from a commodity nature to an essential attribute of AI computing power infrastructure. Morgan Stanley emphasizes that CPUs in the AI computing system are becoming important growth engines after GPUs, driving the transition of memory and NAND storage demand from marginal growth to explosive growth. In the coming years, the proportion of AI CPUs, ASICs, and GPUs requiring high-bandwidth, low-latency memory will significantly increase, with the advanced manufacturing processes and complex testing of HBM making supply-side expansion difficult, leading to a market situation of continuous supply shortages from 2026 to 2028. With the reshaping of supply and demand in the global storage market, long-term supply agreements (LTA) and structural distribution of demand are also reshaping price trends. The widespread adoption of long-term contracts not only enhances profit certainty but also provides cash flow protection for storage manufacturers through deferred revenue and prepayment mechanisms. This shift has moved pricing away from being solely influenced by spot and inventory cycles, to being tied to the long-term capital expenditures of large enterprise clients (particularly tech giants like Microsoft, Google, etc.). This mechanism significantly reduces the risk of cyclical fluctuations. This reshaping logic has even brought about a fundamental change in the valuation system. Goldman Sachs points out that in the environment of the AI infrastructure frenzy and the continuous popularity of long-term supply agreements (LTA), storage chip manufacturers are moving away from valuation based solely on price-to-book (P/B) ratios to a price-to-earnings (P/E) valuation framework based on profitability, prompting Wall Street to significantly raise the target prices for Samsung Electronics, SK Hynix, and Western Digital, among others, and their profit rules are shifting from "cyclical booms and busts" to "predictable and stable cash flow generation." As revealed by Jeremy Werner, Senior Vice President and General Manager of the Data Center business department at Micron Technology, in a recent interview, from the underlying AI data center data flow processing engineering logic, the driving force behind this trend is not just that "AI needs more computing chips," but that the AI Inference agents, like Claude Cowork, and OpenClaw, have transformed memory/storage from complementary components to system bottlenecks. AI training engineering relies more on large-scale parallel computing, while inference, especially those that involve long contexts, multi-round conversations, and Agentic AI workflows, require continuous storage of KV Cache, contextual states, and intermediate results. When there is insufficient memory/storage space, the model has to recompute historical states, leading to a decrease in GPU utilization and an increase in token generation costs. Therefore, HBM, DDR5, LPDDR, enterprise-grade SSDs, and HDDs/Data Lakes are forming an "AI Memory Chain" from near-end GPUs to remote storage, determining the throughput, latencies, concurrency capabilities, and unit economic viability of AI systems. This is why storage and data storage stocks like Micron, Samsung, SK Hynix, SanDisk, and Western Digital are experiencing a synchronous surge: the demand is not just concentrated on HBM but is spilling over the entire chain towards DRAM, NAND, SSD, and HDD.