Start by Topic

11 research hubs

Korea Foreign-Investor Playbook: KOSPI 168, KOSDAQ 355 and the MSCI DM Optionality

A Research OS breakdown of 2026 YTD Korean foreign-investor activity shows that KOSPI foreign attention is concentrated in 168 names and KOSDAQ in 355 names. The practical MSCI DM-upgrade question is not whether foreigners buy Korea broadly, but which Korean stocks are already in the foreign-investor playbook.

Alibaba AI Pay's 300M Transactions: Korea's Real Risk Is Seller OS and Payment Permission, Not a 272T Won E-Commerce Collapse

A follow-up to the NAVER rerating case: Alipay AI Pay's 300 million transactions and Alibaba.com's Accio Work Korea launch matter less as a direct attack on Korean consumer e-commerce, and more as a threat to seller workflows and agentic payment trust layers.

US Non-Semiconductor Re-Rating Check: AI Power, Software and Stablecoin Rails Through a Korean Equity Lens

A two-month review of the strongest US non-semiconductor re-rating themes, from quantum and fuel cells to AI software, stablecoin rails and space-defense, then translated into Korean equities such as NHN KCP, Douzone Bizon, Hanwha Systems, Semyung Electric, LG CNS and NAVER.

In an Earnings Bubble, Estimates Get Cut Last: AI Infrastructure Crowding and the Value of a Deep-Dive

BCA Research argues that the AI bubble is not a valuation bubble but an earnings bubble. In an earnings bubble, the share price falls well before earnings estimates are cut. So precisely now, with capital crowding into AI infrastructure, it matters more to read system structure and leading indicators directly through deep-dive research rather than wait on consensus estimates. This note lays out that methodology plainly, through Thesis OS and our blog's actual work.

Goldman's Token Demand vs. J.P. Morgan's Memory ASP Peak-Out: Do the Two Calls Really Clash?

Goldman Sachs sees AI agents lifting token usage 24x by 2030 while cost per token falls 60-70% per year. J.P. Morgan sees the year-over-year gains in DRAM and NAND ASP rolling over from 2027. Split the two seemingly opposed calls into P, Q and C, and one coherent path emerges: alpha rotating out of 2026 memory beta and into a post-2027 token-cost-reduction stack.