According to a report from South Korean media MK on August 11, Japanese memory chip maker Kioxia disclosed a change in its ownership structure. Toshiba, which had been the largest shareholder, reduced its stake from 14.48% to 14.12%. At the same time, BCPE Pangaea Cayman2, a special purpose vehicle established by Bain Capital, rose to become the top shareholder with a 14.19% position. That SPV is an investment platform funded by SK Hynix through convertible bonds, as also noted in Kioxia's latest annual report. If SK Hynix converted those bonds into shares, it would hold a 14.19% stake in the company, making SK Hynix the de facto largest shareholder. Nevertheless, the report points out that SK Hynix faces practical obstacles in taking direct operational control in the short term. These include merger review procedures in various countries and a cap on voting rights at 15% until 2028. The ownership change is being interpreted as an important signal of shifts in the global NAND flash memory competitive environment. The disclosure was confirmed in Kioxia's annual report and first reported by MK.

SK Hynix has emerged as the largest shareholder in Japan’s Kioxia through a layered investment structure after Toshiba cut its stake, according to Kioxia’s Aug. 11 disclosure. Toshiba sold shares in seven tranches between July 15 and Aug. 3, reducing its holding from 15.10% to about 14.12%. That left BCPE Pangea Cayman2, a special purpose company set up by Bain Capital, as Kioxia’s biggest single shareholder with a 14.19% stake. The structure matters because SK Hynix holds convertible bonds tied to nearly all of SPC2’s voting rights. If those bonds are converted, the South Korean memory chip maker could gain indirect voting control over Kioxia shares held by SPC2. Kioxia said in its latest annual report that the arrangement presents a potential conflict of interest, given that SK Hynix is a direct competitor in the global NAND Flash market. The investment link goes back to Bain Capital’s 2018 leveraged buyout of Kioxia, when SK Hynix invested about 4 trillion won through two special purpose companies. One vehicle, SPC1, was unwound after Bain exited in June, but SPC2 and its roughly 1.3 trillion won convertible bond position remain in place. Any move by SK Hynix to convert the bonds would still face contractual limits through 2028, antitrust approvals in multiple jurisdictions, and likely political resistance in Japan.

BlockBeats reported on Aug. 11 that SK Hynix has indirectly become the largest shareholder of Japan's NAND flash maker Kioxia through SPC2, a special purpose company it invested in. Kioxia disclosed that its largest shareholder changed from Toshiba to BCPE Pangea Cayman2 (SPC2), which holds 14.19% of the company. Toshiba's stake dropped to 14.06% after selling shares seven times from July 15 to Aug. 3. SPC2 represents about 1.3 trillion won of SK Hynix's 2018 investment made via convertible bonds. Voting rights would still require antitrust approvals, and Japan may decline to approve, so direct influence over Kioxia remains unlikely. Kioxia ranks third globally in NAND market share.

IOSG argues that the current storage rally is being driven by artificial intelligence, but not in the way traditional IT buyers used to think about storage. In its view, the market is no longer rewarding raw capacity first. It is rewarding the ability to keep GPUs fed, move checkpoints quickly, support retrieval-augmented generation with very low latency, and raise overall compute utilization across tightly coupled infrastructure stacks. That shift, the article says, is why components such as HBM, DRAM, CXL, enterprise SSDs, SSD controllers, NVMe pathways, and performance storage software have become central to the AI investment narrative. The piece draws a sharp distinction between AI storage and decentralized storage. AI storage is framed as an efficiency system built for hot data and commercial output. Decentralized storage, by contrast, is described as a trust system for cold data, focused on permanence, censorship resistance, auditability, and public memory. IOSG uses Filecoin and Arweave as the main examples, outlining how the two networks diverge in architecture and product direction, while also listing persistent problems across the sector, including weak enterprise service layers, retrieval limits, supply-demand incentive mismatches, privacy and compliance tensions, and token economics that can amplify market cycles rather than solve product-market fit.

TechFlowPost published a market analysis by Chaoxiang Research that highlighted a recent strategy report from Guotou Securities, titled Investment Methodology for the Technology Industry. The report argues that tech stocks are difficult to value with the same discounted cash flow approach often used for mature companies because technological progress can shift abruptly after a catalyst event. It says returns in the sector come from trading distinct phases rather than simply holding through the full cycle. The core framework is an N-shaped pattern with four points: A, B, C and D. The A-to-B move is the first rally, driven by narrative and total addressable market expectations. The B-to-C phase is the pullback, where many tech names fail to recover. The C-to-D phase is the second rally, led by earnings growth and fast penetration. The report calls point C the most important entry zone for institutional investors and says it can be identified through three conditions appearing together: large-company capital spending, a breakout hit product and order flow reaching the supply chain. For exits, the report uses an M-top framework, watching macro slowdown, price wars on the supply side and cuts in capital spending on the demand side. It also says U.S. stock mapping has been an important historical method, with supply-chain mapping seen as especially effective.

Kioxia said it will launch a share buyback of up to ¥800 billion and carry out a 3-for-1 stock split after reporting record quarterly revenue and operating profit. The Japanese flash memory maker said demand for SSDs used in AI data centers helped drive the latest results, with quarterly revenue reaching ¥1.767 trillion and operating profit climbing to ¥1.27 trillion. The company plans to repurchase as many as 30 million shares, equal to about 5.5% of shares outstanding, between Aug. 3 and Oct. 30, 2026. It also set a total shareholder return target of around 50% and said recurring dividends are planned from fiscal 2028. Kioxia said the stock split will use Sept. 30, 2026 as the record date and take effect on Oct. 1, with each existing share split into three. The move is intended to improve liquidity and lower the entry threshold for retail investors. ABMedia said Japan’s standard trading lot is 100 shares, and at Kioxia’s current share price of ¥50,250, the minimum purchase amount is about ¥5.025 million before fees and taxes. Despite the record quarter, the company’s figures came in slightly below market expectations, and its cautious outlook has raised questions about the pace of future AI infrastructure demand. Kioxia shares rose 10% on the news, after having fallen more than 60% from their June peak following Bain Capital’s exit and a reduction in Toshiba’s stake to 15.1%.

Fujian Jinhua Integrated Circuit Co. Ltd., once seen as one of China’s three major memory projects alongside Yangtze Memory Technologies and ChangXin Memory Technologies, is back in focus after years of disruption. The company was hit by a U.S. export blacklist and criminal charges just as its DRAM production line was coming together, bringing progress to a halt and pushing it out of the public eye for years. That legal overhang began to lift at the end of 2023, when Micron Technology reached a global settlement with Fujian Jinhua and both sides withdrew lawsuits worldwide. On Feb. 27, 2024, a federal court in San Francisco ruled that prosecutors had failed to prove Jinhua stole Micron trade secrets. The decision cleared the company of the charges that had defined much of its recent history. The article traces Jinhua’s rise from a state-backed strategic project launched in Jinjiang in 2016, its early technology partnership with United Microelectronics Corp. (UMC), the central role of executive Chen Zhengkun, and the impact of years of sanctions on its production roadmap. It also places the company inside the economics of the DRAM industry, where scale, capital intensity, yield management and intellectual property disputes have long shaped the global pecking order. Jinhua is still far behind larger Chinese peers. Its 12-inch fab is producing about 40,000 wafers a month, with expansion to 60,000 planned for 2026, and the company remains on the U.S. entity list. But with local state backing, a niche DRAM focus and more than 1,007 related patents, it has not dropped out of the race.

Bain Capital has fully exited Japanese memory maker Kioxia, closing out a deal that began with Toshiba’s distressed sale of its storage business and ended during one of the hottest runs the memory sector has seen in years. According to the article, Bain had sold down its position by early July after starting a large-scale reduction in November 2025. Nikkei estimated the firm’s gain from the share sale at about JPY 2.5 trillion, or roughly $17 billion. The report traces the investment back to 2018, when a Bain-led consortium that included SK hynix, Apple, Dell and Seagate acquired about 55% of Toshiba Memory for $18 billion. Toshiba Memory was later renamed Kioxia in 2019. At the time, the deal drew skepticism because the memory market was already sliding as smartphone demand softened and NAND prices weakened. That picture changed after AI-driven demand reshaped the storage market. The article says HBM demand surged from the second half of 2025, pushing major suppliers to shift capacity toward higher-end products and tightening consumer storage supply. It also reviews Kioxia’s repeated failed IPO attempts, Bain’s long holding period, and compares that patience with the path taken by investors in China’s CXMT.