YB new stock pitches (Mon, Jul 20)

China Merchants China Direct Investments (0133.HK), a Hong Kong-listed closed-end fund, offers the only meaningful listed backdoor exposure to Moonshot AI ahead of the anticipated launch of its Kimi K3 open-weight model, which could replicate the global developer-adoption surge seen with GLM-5.2 (released in June with public weights and a 1M-token context window, gaining traction on OpenRouter as a coding agent). Moonshot teased K3 via a 16 July trailer but has not released a model card, benchmarks, pricing, weights, or licensing; reported (unconfirmed) specs include a ~2.5 trillion-parameter mixture-of-experts architecture and ~1M-token context, focused on long-horizon agent and coding tasks competing against GLM-5.2, Opus, and Fable. Early tester feedback is positive but mixed—some Arena comparisons show K3 leading with high-quality frontend output (though one generation took 35 minutes, raising latency concerns), while others favor Fable 5 for speed and cleaner controls. Moonshot's momentum has accelerated, with reported ARR rising from over US$100M in March to above US$200M in April to over US$300M by mid-June (70%+ from API), and TickerTrends Developer Activity up 223% over three months before plateauing ahead of K3. The fund initially invested US$10M for a 1.29% Moonshot stake in the December 2023 Series A (since diluted to ~0.60%), and on 15 July raised the carrying value from US$17M to US$55M, now 5.68% of the fund's US$967M total assets (up from 1.67% in March), implying a ~US$9.1B Moonshot valuation—below the May round led by Meituan's investment arm (~US$2B raised at a >US$20B post-money valuation) and a reported (unclosed) new round near US$30–31.5B. At the May price the position would be worth ~US$121M (~11.7% of assets); at US$30B, ~US$169M (~15.6%). The Moonshot markup added ~HK$1.96/share to NAV last quarter but was more than offset by declines elsewhere, leaving reported NAV at US$4.954 (HK$38.83) per share as of 30 June. Key post-launch factors are whether K3 is released open-weight and under what license, consistent performance across coding/agent/long-context tasks, economics (pricing and inference speed), and sustained developer/consumer usage. Risks include K3 underperforming early hype, inaccurate reported specs, restrictive licensing, strong performance not translating into usage or API revenue, the financing round closing lower or not at all, and portfolio losses elsewhere offsetting any Moonshot markup.

Read the full article here. Read time: 5 min

Gravity Co. (GRVY), a South Korean game company built on the Ragnarok Online IP (spanning PC, console, and mobile titles across key Southeast Asian markets like Taiwan, Thailand, the Philippines, and Indonesia), has long been a notorious value trap, trading at just 0.5x EV/EBIT run rate (based on 1Q26 EBIT of ~$20m) and 0.7x LTM EBIT ($56m), with net cash of ~$401m covering 91% of its $439m market cap after decades of never paying dividends or buybacks. The thesis hinges on a June 18 6-K announcing an interim dividend (potentially paid as early as July), signaling a pivot away from 20+ years of cash hoarding, driven by ~59% owner GungHo Online Entertainment—whose business is now over 90% Puzzle & Dragons (making Ragnarok negligible)—facing activist pressure from Strategic Capital that forced GungHo to raise its dividend payout floor from 30% to 50%+, add a 4% Dividend-on-Equity target, initiate buybacks, and overhaul executive comp (top exec pay cut ~63% from ¥288.4m to ¥107.7m, now tied to profit and TSR vs. TOPIX), giving GungHo incentive to tap GRVY's overcapitalized balance sheet. GRVY operates a dual-revenue model (in-house and third-party-developed titles, with 81.3% of online/mobile revenue from third-party games but GRVY billing players directly and taking revenue share), though its four largest mobile titles (~59% of FY25 revenue) are externally co-developed, creating pipeline dependency amid declining Ragnarok Online user metrics (down 15%+ in most markets since 2023), offset by a strong 1Q26 (revenue +18% y/y, +43% q/q; EBIT margin +870bps q/q) and a robust 2026-2027 pipeline (24% more launches than 2024, headlined by Ragnarok Online 3, whose China launch slipped to 1Q27). The author estimates fair value of $90-200/share (40%+ upside at the floor), noting peers Webzen and Netmarble trade at double-digit EBIT multiples, that a full cash distribution implies a 93% yield, and that cumulative FCFF through 2028 (~$151m) would push EV negative absent a rerate; catalysts include a material dividend, a new derivative hit, or GRVY adopting a GungHo-style comp scheme, while risks include dividend disappointment/one-off payout, GungHo control (capital allocation depends entirely on GungHo's incentives), pipeline dependency, and a persistent discount due to thin ADR liquidity, PFIC tax drag for U.S. holders, and governance overhang.

Read the full article here. Read time: 11 min

ZIM Integrated Shipping trades at a negative ~$200M enterprise value (~$3B market cap, ~$2.6B cash, ~$600M cash from operations to date) after its ~$4.2B buyout by Hapag-Lloyd ($HLAG) collapsed due to government opposition, creating an opportunity where the market values one of the world's youngest, most efficient container fleets, a leading Trans-Pacific carrier, and billions in earnings potential at negative EV. ZIM operates ~710,000 TEU plus 14 profitable car carriers (earning ~$70-80M/quarter), controls ~35% of the world's modern LNG-powered container ships despite representing only 1.8% of global capacity, and has one of the highest sensitivities to spot freight rates—currently earning an estimated $0.7-0.8/share weekly with peak season beginning unusually early. The thesis rests on tight effective (not theoretical) capacity from Suez/Red Sea disruption (removing ~8-10% of global capacity via longer Cape of Good Hope routes), port congestion, and stronger-than-expected global trade growth (~7% in 2024, ~4.8% in 2025, ~5% expected 2026), all confirmed by the HARPEX charter index making new multi-year highs despite fleet growth. ZIM ordered 48 newbuilds (~80% LNG-powered, delivered June 2026–December 2028) worth ~$4.5B, now estimated ~$1.5B more valuable than at signing and capable of roughly double contracted rates on short-term charters or a 30-40% premium long-term; fleet capacity is expected to reach ~920,000 TEU by end-2028, though management is expected to optimize down to ~600,000 TEU, with ship sales potentially generating ~$25-35/share in additional cash. Hidden assets include ~1.08M containers (~55% owned) carried at ~$800M but with owned-fleet economic value of ~$3.5-4.0B. Under HLAG's own merger math, ZIM would have contributed ~35-38% of combined EBIT, implying a valuation exceeding $10B versus HLAG's ~$26B market cap—yet the market values ZIM at only ~one-tenth of HLAG. EPS estimates (freight-rate dependent) are Q2 $1-2, Q3 $6-10, Q4 $0-4, with normalized cycle earnings of ~$10/share (strong years >$22, weak years ~-$2 loss). Dividends are expected at ~$0.30/share (Aug 2026), ~$2-3/share (Nov 2026), ~$2-3/share (Mar 2027), plus a potential ~$10-14/share special dividend in 2027 or 2028. The author expects modest ~15% share appreciation plus dividends to produce ~100% total return over 24 months (with ~70% of today's price returned via dividends), the returns coming mostly from cash distributions rather than capital appreciation. Key risk is the highly cyclical nature of container shipping and dependence on freight rates remaining near current levels. Portfolio allocation (updated): 20%.

Read the full article here. Read time: 5 min