Summer bubbles, autumn blues

WE have never had it so good. The S&P 500 and Nikkei 225 just hit record highs. SpaceX had its spectacular initial public offering (IPO), raising an eye-popping US$75bil (the largest IPO in history), with a sharp rise in share price to a mind-blowing valuation of US$2.5 trillion.

The possibility of a US-Iran deal to re-open the Strait of Hormuz has caused petrol prices to drop back to below US$4 per gallon.

Global stock markets are creating wealth for portfolio investors that seems unstoppable.

Bull investor Ed Yardeni is calling the S&P 500 Index at 8,250, meaning a year-on-year return of 33% after three consecutive years of double-digit returns. US companies are still recording bumper profits, driven by the artificial intelligence (AI), technology, financial and consumer giants.

Let’s connect some dots of boom-or-bust noise that cloud underlying investment signals.

Firstly, the doomsters are also flagging warning signals.

The World Bank’s latest global projections show that global gross domestic product (GDP) growth will slow to 2.5% in 2026, down from 2.9% last year, mainly due to greater tensions in the Middle East and higher oil prices.

Veteran monetarist Steve Hanke thinks that a financial crash is inevitable based on advanced-country debt levels and monetary stance, but he doesn’t know when it will happen.

Last Wednesday, the Fed Open Market Committee, under its new chairman Kevin Warsh, held Fed funds rates at the current 3.5% to 3.75% range, but took what appeared to be a hawkish stance on maintaining price stability amidst strong productivity growth.

With 30-year US Treasury yields hovering at around 5% per annum – even as consumer prices spiked to 4.2% in May – the average interest costs of US sovereign debt of US$39 trillion are roughly 3.4% per annum, estimated at over US$1 trillion, or just under 19% of federal government revenue.

US Fed monetary policy is not independent of monetary policy in two of the largest surplus economies, Europe and Japan.

The European Central Bank just hiked interest rates by 0.25% on June 11, given higher inflation rates. The Bank of Japan hiked interest rates to 1%, the highest since 1995, with concerns over inflation and yen weakness.

Japanese intervention to defend yen devaluation beyond 160 to US$1 inevitably means selling US Treasuries, which will keep long-term Treasury yields high.

Japan’s holdings of US securities fell by US$73bil in May, even as 30-year Japanese Government Bond (JGB) yields hit a record 4.16% per annum in May. That means that the differential between US and Japanese sovereign debt 30-year yields is narrowing to around 1%, compared to 3% as late as October 2023.

The narrowing gap suggests that long-term investors, such as Japanese insurers and pension funds, would prefer to go back to yen holdings to avoid foreign exchange risks.

This has augured well for the Japanese stock market, but convergence of global interest rates will force a rerating of holdings in US-dollar assets.

As investors worry over US fiscal debt sustainability, global asset rebalancing will occur as interest rates normalise. This is because there is a credit-risk premium now attached to US government debt, which used to be zero.

Secondly, unlike the bond market, the US stock market is compensating investors in terms of returns for holding US dollar assets but its spectacular growth is more recently narrowly focused on 10 AI-related stocks.

The US stock market capitalisation is US$62.2 trillion, just under half of global capitalisation.

The combined valuation of SpaceX (just listed) and Anthropic + OpenAI (waiting for IPO) is an estimated US$4.5 trillion, lower than Nvidia, with a market valuation of US$5 trillion.

Putting things into perspective, Nvidia has a larger market capitalisation than the London Stock Exchange (US$4.4 trillion).

How the market is valuing these high-tech companies is nothing short of amazing. In its prospectus document, the estimated net revenue of SpaceX in the year to March 2026 is negative to the tune of US$4.9bil.

Investors are valuing Anthropic using its annualised run revenue, based on revenue estimated at US$47bil for 2026.

This allowed Anthropic to raise US$65bil to achieve a post-money valuation of US$965bil in May 2026.

Using the same logic, OpenAI raised US$122bil to reach a market valuation of US$852bil in March.

The market is betting that these three companies will deliver superior future net profits to justify their valuations, but in essence, they are betting on the ability of Elon Musk and the key entrepreneurs of Anthropic and OpenAI to deliver on such promises.

The mission statement of SpaceX is breathtaking in its ambition.

It aims “to build the systems and technologies necessary to make life multiplanetary, to understand the true nature of the universe, and to extend the light of consciousness to the stars”.

For such audacity, the market believes their starry dreams. But the secret sauce of all financial bubbles is excess market liquidity, because if liquidity is scarce and the price of money is high, bubbles based on high leverage cannot form.

Financial bubbles are Ponzi schemes whereby one borrows from Peter to pay Paul at promised higher returns (or interest rates) until the music stops. What goes up cannot defy gravity forever.

The gap between promised expectations and reality cannot be infinite.

As inventor Thomas Edison once said, “Genius is one percent inspiration and ninety-nine percent perspiration”, commonly translated as “vision without execution is hallucination”.

We all know that, technically, AI models can hallucinate, because if the data is wrong or incomplete, the result is garbage. Financial hallucination floats on excess government spending and deficits based on excess debt, enabled by loose monetary policy.

As long as money is made available cheaply through fiat monetary printing, financial engineering genius comes from promising the stars.

Shrewd investors understand from bitter experience that you can never buy at exactly the bottom nor sell at the top. When greed overcomes fear, hallucination happens. We know AI hallucinates; will you join the crowd?