The Bank of Korea has resumed purchasing physical gold after a 13-year hiatus, aiming to diversify its foreign exchange reserves.

Per Lookonchain’s monitoring, the CXMT short whale address 0xf292 has paid more than $2.04 million in funding fees. As CXMT’s price continues to rise, the whale’s short position currently carries an unrealized loss of roughly $4.3 million, with a liquidation price of $15.8. Market data shows that this address has been consistently shorting CXMT, yet has been incurring funding costs amid the price rally, and is now facing substantial position pressure.

The $CXMT mega bear 0xf292 has now paid more than $2.04M in funding fees. As $CXMT keeps rising, the whale is now down $4.3M on the short position. Liquidation price: $15.8

According to Bitget market data, Kioxia rose 2.97% intraday, while the Nikkei 225 index fell 0.53%. In related news, Kioxia officially announced its first GP series solid-state drive (SSD) — the KIOXIA GP1 — which supports direct GPU access to high-speed flash memory. The product delivers extreme random read performance of 100 million IOPS. Kioxia stated that the KIOXIA GP1 series is designed to support the emerging AI storage architecture that integrates high-speed flash memory media into graphics memory systems. This approach enables AI systems to access larger datasets at a far lower cost than adding HBM, while boosting GPU utilization.

South Korean cryptocurrency exchange Bithumb will list META2 and USDG KRW trading pairs.

Citrini analyst Jukan published a report stating that NVIDIA is discussing adjusting the HBM configuration of its next-generation Rubin Ultra GPU, possibly shifting the main version from HBM4E 12Hi 384GB to HBM4 8Hi 192GB, with an HBM4E 8Hi version to be launched later. The analysis notes that if NVIDIA’s CoWoS capacity allocation reaches 1.2 million units in 2027, its accelerator output is expected to hit around 9.9 million units, of which Rubin Ultra production will be roughly 1.6 million units. If the lower-spec HBM scheme is adopted, NVIDIA’s 2027 HBM demand is projected to drop from 24.1 billion Gb to 21.6 billion Gb, a decrease of about 10%, with overall HBM demand potentially falling by around 4%. Jukan argues this "spec downgrade" is not due to weak demand, but rather NVIDIA’s strategy to boost accelerator output amid limited HBM supply. As DRAM manufacturers’ HBM expansion pace cannot match TSMC’s advanced process and packaging capacity growth, HBM supply has become a bottleneck for AI chip production. Additionally, DRAM makers also struggle to significantly adjust their capacity allocations. Currently, smartphone and PC vendors in the consumer electronics sector are grappling with storage shortages, and further cuts to traditional DRAM supply could intensify pressure on the end market. The analysis concludes that by adjusting HBM configurations, NVIDIA is poised to unlock more AI accelerator capacity with limited storage resources while easing the HBM supply bottleneck.

Palantir (PLTR)’s latest financial report shows that the company’s second-quarter revenue rose 93% year-over-year to $1.94 billion, net profit hit $1.06 billion, and it significantly raised its full-year performance guidance. Boosted by the results, the company’s stock price climbed more than 12% in after-hours trading on Monday. Palantir forecasts full-year 2026 revenue of at least $8.15 billion, up from its prior expectation of less than $7.7 billion. U.S. commercial business revenue is projected to reach $3.4 billion, growing at least 134% year-over-year, making it the key growth driver. Markets had previously worried that the rapid expansion of AI model firms like OpenAI and Anthropic could erode Palantir’s competitiveness, but this earnings report has reignited investors’ focus on its "AI Sovereignty" business model. Palantir CEO Alex Karp noted that enterprises and governments are seeking control over their own data, AI systems, and decision-making processes, a key source of the company’s growth. The so-called "AI Sovereignty" refers to enterprises preventing their core data from being used for training by external AI model providers, while ensuring AI systems align with their own business needs. Palantir argues that future enterprises need not just base models, but AI platforms that can manage data, integrate into business workflows, and protect competitive advantages. In recent years, Palantir has been expanding its enterprise AI business and criticized the business models of some AI labs, claiming that enterprises pay large fees to model suppliers yet may lose control over their core data. However, Palantir still faces pressure in overseas markets. Due to the company’s long-standing work with the U.S. government and defense sector, some European nations are reducing their reliance on U.S. tech firms. As of the latest quarter, international client revenue makes up about 19% of Palantir’s total revenue, down from 26% in 2025.