Parallels and Intersections | Sept ’25

Thoughts are welcome via emails to zhenghan236@gmail.com, and will be featured in the next Tab

---

If you’ve been following my tabs, you would be privy to my rather conspicuously grim outlook on the AI boom. I still stand by that — with a key caveat. That is, if you’re thinking of entering the AI market as a long-term investor. With the sky-high valuations, the value investing ship has long sailed. Yet, the 2 competing schools of thoughts in the AI rally, coupled with the Trump administration's scattershot trade policy, has created a conducive volatility that traders can exploit. Success in such a market still rests on the same fundamentals: careful stock picking, disciplined entry and exit, and judicious use of investing instruments, but not by the book. 

I shall draw upon the example of Microsoft and Global Crossing during the dotcom bubble, the latter of which you’ve likely never heard of (well... survivorship bias). During the dotcom bubble, both Microsoft and Global Crossing poured money into infrastructure. But while Global Crossing bled red ink, Microsoft never suffered an operating loss — a crucial difference. Amazon, meanwhile, raised liquidity through convertible bonds and survived the storm with a negative cash conversion cycle. Evidently, nimble, decisive leadership and financial prudence determined who lived to tell the tale. Still, it took Microsoft 16 years to recover its dotcom peak. Traders, however, care less about that kind of timeline.

The AI surge invites parallels to dotcom, but also key differences. Dotcom-era firms often had weak or no revenue, let alone profits. Start-ups like Global Crossing relied entirely on plentiful capital markets to fund their expansion. Today’s AI players — Nvidia, Microsoft, Amazon, Alphabet — are highly profitable, cash-rich incumbents that fuelled their capex with cash. These giants make up a fifth of the S&P 500’s market value but only 2% of the investment-grade bond markets. Financially, their balance-sheets look rock solid. Even if AI turned out to be a bust, their core business models and revenue streams can keep them afloat for years. The risk of default, then, is low. 

That doesn’t mean that your capital is secure during a downturn. The real risk is valuation reset, exacerbated by leverage and the asymmetry of losses and gains. A retracement of 10% at 2x leverage wipes out 20%, requiring a 25% gain to break even. A 25% retracement becomes a 50% loss, needing a 100% gain to recover. Beyond that, it’s effectively Game Over.

It’s not all grim. This analysis assumes passive market participation, which most successful traders rarely stick to. Conservative position sizing during inflated prices, robust risk management and trade discipline can tilt the odds in your favour, and the prevailing volatility is the best place to gain that edge. In this case, profitability and market conditions can seem to have no meaningful correlation. 

Notes from the author: Conversations with my dad’s friend, Mr DJ, highlighted the importance of distinguishing between theory and execution, as well as investing and trading. I’m grateful for his insights and learnt a great deal from our exchange.