The Cautionary Tale of Financial Euphoria | Aug ‘25

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

---

It is rather clear that, as of Aug ’25, markets are exuberant, greed is pervasive, and retail investors are riding high on the AI hype train. Credit is abundant, exacerbated and masked by a fast-growing private equity sector. Palantir is trading at 600 times earnings and a valuation-to-sales ratio of 120. Meta is pouring over $68B into AI CAPEX, up 74% YoY, and returning only 21% revenue uplift. Market sentiment is largely positive as well, with core CPI rising lower than expected, especially factoring in tariffs. A rate cut by the FOMC is therefore expected at an 82.6% probability.

There is, however, a key insight: CPI figures for tariffs lag policy implementation, as tariffs are paid by importers upfront. It would take a considerable period before these costs are eventually passed on to consumers and reflected in the CPI data. Internal communications at Ford suggest a potential 1.5% increase in prices, while Apple expects a $900M rise in costs for the current quarter. In the near term, mitigation of Chinese production for Apple seems unlikely without support from the Indian government and its current lack of production infrastructure.

The buoyant attitude towards a rate cut motivates the shorting of the USD in the FX market, which expects depreciation as capital is moved out of US securities. Given FX’s high leverage, a correction in CPI and a subsequent rate hike poses significant risk of margin calls to bearish FX traders. Capital from equity markets may be shifted to cover losses. Tech (especially AI) earnings expectations were elevated at the previous earnings calls to good results; much is riding on continued performance by tech to sustain the high valuations. Should earnings miss expectations, it is reasonable to expect additional momentum on the FX trigger to correct the overvaluation.

We are likely in the euphoric extreme of the credit cycle, where too much capital is chasing too few deals. AI CAPEX is reaching an alarming level; AI startups are crawling in cash without returning proportionate value, and AI has yet to prove itself (only 5.1% of US companies are using AI in production). If history is anything to go by, the AI hype is increasingly resembling the dot-com bubble.

Notes from author: When I wrote this Tab, my main scepticism stems not from the utility or potential of AI. In fact, I’m a rather strong proponent of it. The main concern, however, is excessive infrastructure spending fueled by lose credit, which draws a striking parallel to the telecom era. Rosy projections have to be tempered with realistic expectations — it is appealing to think of a future where AI proliferates, but that is merely one out of a million different possibilities. Remember flying cars? There may be a chance that today’s investment pays off, or that the company you back defaults. Without analysing the fundamentals, the latter seems more likely.