To The Moon

My takeaway is that the Treasury buyback is more of a signal than a solution. Increasing long-end purchases may temporarily cap yields, but it cannot fix the deeper problem: huge fiscal deficits, heavy Treasury issuance and growing competition for capital from AI infrastructure spending. The market’s reaction already suggests investors are skeptical—the 30-year yield quickly recovered after the initial drop. For equities, I wouldn’t chase valuation expansion. If earnings continue to deliver, especially in AI, higher profits can justify today’s multiples. But July’s forced unwind is also a reminder that leverage and positioning can overwhelm fundamentals in the short term. Personally, I’d rather pair AI growth exposure with gold or other defensive assets than build a 100% AI portfolio. Gol

Reflect on August, Plan for September: What the Market Taught Me August was a good reminder that investing is not simply about being right. It is about being right for the right reason, at the right price, with the right position size. Looking back at my August trading, the biggest lesson was not a particular stock or a particular return. It was learning to distinguish between a good company, a good story, and a good trade. They are three completely different things. ① August Recap — What Did I Get Right? The trade I am most satisfied with this month was staying focused on the areas where earnings and fundamentals were actually improving, rather than blindly chasing whatever stock was moving the most. AI infrastructure remained one of the strongest structural themes. NVIDIA's latest result

From Cheap to Compounding: My Biggest Investing Lesson I think value investing is not about choosing between Graham and Buffett, but knowing when to use each mindset. Cheap valuations provide a margin of safety, while industry insight helps identify businesses whose earnings power is still underestimated. For me, the real edge is understanding an industry before the market fully prices in its growth—watching consumer behavior, supply chains, competitive moats and the stage of the cycle. But insight means little without survival. Position sizing, cash reserves and disciplined rebalancing protect capital when our thesis is wrong. I would rather miss an opportunity than lose the ability to participate in the next one. Survive first, compound second.

Jeremy Tan’s biggest lesson isn’t about finding the next NVIDIA—it’s about surviving long enough to capture the next opportunity. The –8% stop-loss rule, disciplined position sizing, and staying within your circle of competence are simple ideas, but extremely difficult to follow when emotions take over. I especially like the “core-satellite” approach: build a stable foundation with dividend stocks, bonds or ETFs, then use a smaller portion for high-conviction growth opportunities. The smartest investor isn’t the one who predicts every winner. It’s the one who protects capital, controls risk, and stays in the game long enough for compounding to work. @TigerClub [胜利]

The biggest edge for retail investors may not be faster data, but better observation. Fund managers can analyze thousands of companies, but they cannot personally experience every industry. Pop Mart is a perfect example: people who noticed customers collecting, sharing and lining up for Labubu could recognize the brand’s emotional moat before it became obvious in financial statements. For me, the key lesson is to combine industry intuition with Stage Analysis. Your daily observations generate the idea; fundamentals validate it; charts help identify when the market is finally recognizing it. The next Pop Mart may already be hiding in plain sight — perhaps in an industry you understand better than Wall Street does. @TigerClub [鼓掌]

Two strong earnings reports, yet two completely different stock reactions. That’s not contradictory—it’s the market pricing expectations. NVIDIA delivered $96.2B in revenue, up 106% YoY, with data-center revenue surging 117%. The numbers were already enormous, but Jensen Huang’s comments on accelerating AI demand and the longer-term growth outlook gave investors a reason to raise their expectations again. Marvell was different. Revenue reached a record $2.74B, up 37%, data center grew 46%, and FY2027/FY2028 targets were raised. Fundamentally, little went wrong. The problem was that investors had already priced in a much bigger Google-driven upside. The lesson is simple: stocks don’t trade on results alone. They trade on the gap between reality and expectations. Sometimes great earnings ra

My pick: D. ServiceNow. Salesforce has delivered the most eye-catching AI growth, but I believe ServiceNow has the strongest long-term AI monetization story. The reason is simple: ServiceNow doesn’t just provide AI assistants—it owns critical enterprise workflows across IT, HR, customer service and operations. That gives its AI agents a natural path from answering questions to actually executing tasks and automating processes. As companies shift from paying for software seats toward paying for AI agents, workflows and outcomes, ServiceNow could capture a larger share of enterprise spending. Salesforce’s Agentforce growth is impressive, while Microsoft has unmatched distribution and infrastructure. CrowdStrike remains a cybersecurity AI leader. But for pure AI-driven expansion of software

Singapore Market: Where the Real Momentum Lies Singapore trading remains dominated by blue-chip names such as DBS, OCBC, UOB and Singtel, reflecting continued demand for defensive, high-quality assets. However, the real standout is Yangzijiang Shipbuilding (BS6.SI), the only stock on the list hitting a 52-week high. Its strong orderbook and improving earnings provide a solid fundamental backdrop for the rally. At the same time, unusual volume in Micro-Mechanics, CATL, Baidu and Keppel Infrastructure suggests investors are rotating into selected growth opportunities. The surge in leveraged warrants on Nvidia, Micron and SMIC also shows strong appetite for AI and semiconductor exposure. My takeaway: momentum is strongest where fundamentals, industry trends and capital flows converge.

Nvidia’s guidance changes the AI conversation from “future potential” to visible, accelerating demand. A $108B Q3 revenue outlook, alongside surging data-center sales, suggests hyperscaler capex still has plenty of fuel. I’m therefore more constructive on AI hardware into September—but I wouldn’t chase blindly. Rising memory costs and Nvidia’s projected margin compression show that even winners are starting to face capacity constraints and higher input costs. For valuations, I’d still put AI capex ahead of Fed policy in the near term. A hawkish Jackson Hole could trigger volatility, but if AI spending continues compounding, strong earnings can ultimately overpower higher-rate pressure. As for Meta and Snap, I think regulatory risk is becoming a structural theme rather than a one-off. Teen

I’m leaning toward a genuine memory supercycle, but I wouldn’t chase the rally blindly. The biggest difference this time is AI: HBM, DRAM and NAND demand is increasingly tied to data-center expansion rather than the traditional PC and smartphone cycle. At the same time, supply cannot respond quickly, as new fabs, advanced nodes and capacity expansion require years of investment. SK hynix’s ₩40T buyback and Samsung’s ₩90–110T shareholder-return plan are also significant. They suggest these companies are generating enormous cash and are increasingly willing to share the AI windfall with investors. However, the biggest risk is still AI capex. If hyperscalers slow spending or memory prices peak, earnings expectations could change rapidly. For now, I remain bullish, but I’d watch memory pricin