Market episodes, capital cycles and mistakes that shaped investing history. Open a story for its context, turning point and a practical lesson.
Technology changes. Greed, fear and leverage keep returning in new clothes.
One of history's greatest minds was not immune to crowd psychology.
South Sea shares surged in 1720. Newton bought, sold at a profit, then watched prices keep rising.
He re-entered near peak euphoria and suffered heavily in the crash. The famous £20,000 loss is widely repeated, though historians note the surviving records do not make it perfectly certain.
Intelligence is no vaccine against FOMO. A good first trade can become a bad second one when envy replaces process.
He read the crash correctly — but could not preserve the fortune.
Livermore saw weakness beneath a euphoric market and built short positions into the 1929 collapse.
His profit is estimated near $100 million, yet subsequent oversized bets destroyed much of his wealth.
Reading one cycle correctly is not permanent mastery. Risk control and knowing when to stop matter as much as finding the trade.
A scandal crushed the share price, but not the franchise.
The 1963 salad-oil scandal hit American Express with major liabilities and a crisis of confidence.
Buffett checked whether customers still used the brand. In 1964 he ultimately built a roughly $13 million position, reported as more than 40% of partnership assets.
Separate a temporary financial wound from permanent franchise damage. Field evidence can be more useful than fearful headlines.
Elegant models, crowded trades and a world that refused to behave normally.
LTCM included Nobel laureates Robert Merton and Myron Scholes and used leverage to amplify small pricing gaps.
Russia's crisis broke correlations and liquidity. LTCM lost about $4.6 billion; 14 firms provided $3.6 billion in a private rescue coordinated by the New York Fed.
“Rare” risks become common when positions are crowded and liquidity vanishes. A model is not a margin of safety.
A fixed exchange-rate promise collided with economic reality.
Britain tried to hold sterling inside the European Exchange Rate Mechanism despite growing economic pressure.
The Quantum Fund shorted sterling at scale. Britain left the ERM on September 16; Soros's profit is commonly estimated around $1 billion.
Policy can resist markets temporarily, not indefinitely when fundamentals oppose the promise.
AI is also a capital chain spanning power, data centers, logic chips, memory and markets.
A strategic attempt to turn slower growth into a digital-infrastructure opportunity.
Thailand is promoting cloud, data centers and digital infrastructure to attract higher-value foreign investment.
Large commitments create opportunities for power, industrial estates and telecoms, but commitments are not revenue until projects connect and operate.
Trace the value chain and distinguish shovel sellers with real cash flow from stories that remain on paper.
When infrastructure spending grows faster than visible returns.
Hyperscalers are racing to buy chips, build data centers and secure power. Multi-trillion figures are multi-year estimates, not one bill.
Cloud revenue is growing, while investors question payback, depreciation and overcapacity. Unlike many dot-com spenders, today's giants have substantial cash flow.
Being right about a technology does not make every valuation right. Connect capex to revenue, margins and payback.
A foundry leader became both an economic engine and a geopolitical asset.
TSMC manufactures advanced chips for many leading designers, giving it exceptional influence in Taiwan's market and the AI supply chain.
AI drives advanced-node demand, while overseas fabs trade higher costs for geographic resilience.
Exceptional quality still carries concentration, capex-cycle and valuation risk.
US design and platforms, Taiwanese foundries, Korean memory.
US-designed GPUs require Taiwanese manufacturing, advanced packaging and Korean HBM.
Orders flow from hyperscalers through Nvidia, TSMC, SK Hynix and equipment vendors. One bottleneck can reshape prices across the chain.
Look beyond the final star. Find the bottleneck, pricing power, inventories and where profit actually accumulates.
SK Hynix moved first; Samsung is using scale to fight back.
HBM stacks memory close to GPUs and has become critical to AI servers.
SK Hynix became Nvidia's leading HBM supplier while Samsung accelerated new products. Counterpoint data cited by Reuters put SK Hynix near 57–58% recently.
Product leadership can outrun corporate size, while aggressive expansion can seed the next oversupply cycle.
Composite educational cases focused on the mechanism, not a single historical event.
A 10% gain becomes 50% — and so can the loss.
A young investor borrows several times their capital to chase a rising stock. Early wins increase both confidence and position size.
A 15–20% fall can erase the equity at roughly 500% exposure. Forced selling removes control at the worst moment.
Leverage shortens how long you are allowed to be wrong. Size risk against real equity, not hoped-for returns.
The market does not need permission before collateral is sold.
A concentrated portfolio uses margin and waits for a rebound as correlations rise.
Collateral falls below maintenance levels. Cash must arrive or positions are liquidated into weak liquidity.
Define invalidation before entry, keep reserves and assume correlated holdings may fall together.
Calm is not stubbornness; discipline is not inaction.
After several wins, an investor confuses self-worth with forecasting skill and treats contrary evidence as an insult.
Confirmation bias turns “conviction” into delayed loss-cutting after the thesis has failed.
Write down what would prove you wrong. The goal is not to be right; it is to protect capital and update probabilities.
Sometimes a rally is simply trapped holders searching for an exit.
A stock rallies intraday and social feeds celebrate, but heavy volume arrives at resistance and the close fades from the high.
The investor sees green while ignoring trend position, closing quality and supply at resistance.
Read price in context: trend, volume, base and invalidation. Sentiment is data, not an instruction.
Want to discuss a real portfolio situation? Connect with Amy or join the community.
Contact AmyJoin Zalo group