Samsung Q2 Operating Profit Seen Near 85 Trillion Won, Eyeing Nvidia

*Editor's Note: 'AI PRISM' (Personalized Report & Insight Summarizing Media) is an "artificial intelligence (AI)-based customized news recommendation and summary service" developed with support from the Korea Press Foundation. It selects and provides six customized news items by reader type.

1. Samsung Electronics Operating Profit to Exceed 100 Trillion Won? Global No. 1 in Sight

- Key Summary: Samsung Electronics will announce its second-quarter preliminary earnings on the 7th, with operating profit projected at around 85 trillion won. This represents a 49% increase from the first quarter, which recorded the highest earnings ever, and would surpass the $53.5 billion of Nvidia, currently ranked first in quarterly operating profit among global tech companies. Samsung Electronics' DRAM production stands at 650,000 to 700,000 wafers per month, more than double that of third-ranked Micron, and its high bandwidth memory (HBM) supply and next-generation technology competitiveness are credited with driving this profit structure. Analysts also suggest that, excluding some 8 trillion to 17 trillion won in performance bonus provisions, actual quarterly operating profit would already have surpassed 100 trillion won.

2. Shinhan Financial Chairman Jin Ok-dong: "Without Changing Now, There Is No Future… Must Strengthen AI Capabilities"

- Key Summary: Shinhan Financial Group held its "Second Half 2026 Management Forum" on the 3rd and 4th at Shinhan Bank's Blue Campus in Yongin, Gyeonggi Province, with about 300 executives attending. Under the slogan "Vibrant Shinhan, Overwhelming Immersion," the forum focused on strengthening AI transformation (AX) execution. On the first day, AI was used throughout the entire forum, including an AI agent taking on a "red team" role by analyzing debates in real time and presenting counterarguments and supplementary opinions. Chairman Jin Ok-dong stressed, "As generative AI becomes universal, the information gap between ranks will narrow, so leaders who can cross various boundaries and understand context will become necessary." On the second day, case presentations of AI agents by subsidiaries were made, along with the sharing of second-half plans to leap into an AI-native company.

3. Conditions for Success of the Three Mega-Projects

- Key Summary: The three mega-projects of semiconductors, physical AI, and data centers drew the attention of the global industrial community to the extent of appearing on the front page of the Financial Times. The core of these projects lies in ensuring that the three pillars do not stop at fragmentary development but build the entire AI value chain into a single organic ecosystem, and they also include a nationwide expansion model utilizing regional infrastructure such as Yeongnam, Honam, and Chungcheong. It is pointed out that for success, three tasks must be addressed first: institutional support such as shortening permit periods, securing labor flexibility in the research and development domain, and stabilizing the supply chain for materials, parts, and equipment. The government's policy of placing a dedicated official directly under the presidential office to personally oversee permits and infrastructure construction appears to be viewed positively.

4. Naver AI Search Cuts Hallucination by 30 Percentage Points and Reduces Costs to One-Third

- Key Summary: Naver newly applied three core technologies—a product-native large language model (LLM), harness engineering, and multimodal technology—to its conversational AI search "AI Tab." Instead of competing with general-purpose ultra-large models, it developed a lightweight model specialized for actual service environments such as search, shopping, and reservations, reducing hallucination—a chronic problem of generative AI—by 30 percentage points compared to its existing in-house model HyperCLOVA X, and cutting equipment operating costs to as low as one-third of the previous level through a method combining small language models (SLMs) specialized by role. Response speed was also improved more than twofold, and image search capabilities were strengthened by building a multimodal dataset of 35 million. This reflects a plan to respond to the offensive of global big tech such as Google in the domestic search market by simultaneously raising cost efficiency and service accuracy.

5. U.S. and EU Strengthen Robot Security, but Korea Has Neither Manpower Nor Measures

- Key Summary: The government and companies have unveiled plans to invest more than 1,000 trillion won in AI data center construction by 2035 and to commercialize humanoid robots in 10 major industries by 2028, but strategies to secure robotics security personnel to operate them safely remain in the early stages. Global research firm Future Market Insights forecast that the global robotics security market will grow about threefold, from $4.7 billion (about 6.3 trillion won) last year to $14.3 billion (about 19.3 trillion won) by 2035. The European Union will make the application of cybersecurity mandatory from the design and development stages of digital products through the enforcement of the Cyber Resilience Act (CRA) beginning this September, and the United States is also pushing the "Cyber Trust Mark" system, a security certification label for consumer Internet of Things (IoT) products. The industry assesses that finding convergence talent equipped with expertise in information security, robotics, and AI is like "picking stars from the sky," and there are growing calls for Korea to quickly establish a robotics security specialization track and personnel training system.

6. Semiconductor Industry to Need Water for 800,000 Swimming Pools by 2035… "Must Draw Up Dam and Weir Plans That Don't Swing with Administrations"

- Key Summary: With enormous amounts of water needed for semiconductor processes and AI data center cooling, global market analysis firms forecast that the semiconductor industry's water demand in 2035 will reach 2 billion tons, enough to fill 800,000 Olympic-sized swimming pools. Water usage at the domestic and overseas facilities of Samsung Electronics and SK hynix (000660.KS) has already reached an all-time high of 314 million tons, a figure that surged 40% in six years from 2019. The current National Water Management Master Plan presents measures only through 2030, so new large-scale demand such as semiconductor fabs in the Honam region is not reflected, and climate volatility—such as growing annual variation in summer rainfall in the Gwangju region—is increasing uncertainty in water supply. The semiconductor industry is urging the establishment of a predictable and coherent national water management policy, saying, "Under the current structure, in which dam and weir construction plans dance according to the administration, it is difficult to draw up stable long-term investment plans."