A couple of days ago I argued that the impact of AI (in the context of LLMs) on the economy will be sanguine. This will come to the possible relief of some and to the disappointment of others. There will be no AGI, takeoff, ramp-up or other hyped-up scenarios in which AI somehow escapes its confines, becomes sentient, or sees a recursive sort of compounding of its capabilities. 2027 and beyond will come to pass without the anticipated ‘phase shift’, but rather continued incremental improvements at common LLMs tasks, and more computational resources to throw at these tasks.
But what if I am wrong. If AGI or takeoff does somehow happen, it will be reflected in the equity and bond markets in anticipation of its arrival. We would expect stock prices to suddenly surge in anticipation of significant real GDP growth, and a concurrent crash in long duration bond prices in anticipation of high inflation, although it’s also possible bonds can surge in anticipation AGI lowers inflation. This surge would come without warning, and of the likes of which have never been seen before.
The S&P 500 could conceivably gain 200% or more in a year in anticipation of AGI (the record was 50% in 1943), although the window would be a narrow one, meaning that you would have a limited opportunity to participate and reap those huge returns, after which it would be too late and subsequent returns would be more in line with average historical returns for the S&P 500. Consistent with the efficient market hypothesis, once the anticipated growth is factored in, prices will stop rising as much, so the chart would resemble a step-pattern, in which the chart abruptly goes vertical and then also just as abruptly levels off. The market would appear massively overvalued by the sudden surge, and then the rapid growth would come later, bringing valuations in line with long-term averages.
Millionaires and billionaires would be minted seemingly overnight. Anyone with call options could see returns in excess of 1,000x, similar to Bitcoin during its early days. It’s also possible that counterparties would become insolvent, creating confusion and liquidity problems in certain markets. Because of the impossibility of pinpointing when AGI or takeoff happens, a prudent move could be to allocate some small percentage of one’s account for far-out-of-money (OOM) long-dated index call options. If you think there is a 1% per annum chance that the market comes to the ‘mass realization’ that AGI is inevitable, then you would allocate 1% of your account balance annually for such OOM options.
The farthest out-of-money SPY call option which expires in a year, the SPY $900 strike, as of 8/13/2025 trades at 35 cents. If the S&P 500 were to double over the next year, this option’s value would increase by roughly 1,142-fold.
Alternatively, under a continuous rather than discreate model, the market’s confidence in AI boosting productivity rises, but short of AGI or takeoff. This would be reflected in the stock market rising more gradually instead of a step-pattern, with annual returns in excess of 30-40% for the S&P 500, similar to the post-War expansion during the ’40s and ’50s–a period of historically strong GDP growth and stock market returns. In this case, a general diversified portfolio would work well (e.g. QQQ, SPY), which is what many people are already doing, although I would add some leverage such as QLD, which is the 2x version of the Nasdaq 100.
As I correctly predicted, avoid emerging/ex-America markets, as those have not only historically significantly lagged US markets, but also are not dominant players in AI either, except maybe China:
Or better yet, combine this with the OOM strategy. One strategy could be to diversify equally among AAPL, AVGO, PLTR, NVDA, META, TSLA, AVGO, AMZN, and MSFT all of which are potential key players in AI. Each of them fills some sort of niche or role, like Meta’s Llama, Google’s Gemini, Microsoft’s Open AI investment and Copilot, Broadcom ASICs, Nvidia’s GPUs, or Tesla’s robots and autonomous cars (there is no way to invest in Grok directly, but Tesla is close enough, I suppose). Apple and Amazon does not have any offerings, but they can quickly scale. It is too bad there is no way to invest in Anthropic in a similar manner though.
But right now it would seem as if the market’s confidence in the likelihood of AGI is low, with the possible exception of Nvidia stock. The S&P 500 is only up 18% year-over-year. Nvidia long-dated out-of-money call options are appropriately very expensive, meaning a very high implied volatility. So the returns will not be as good. This is why I mention index call options, as those have a much lower implied volatility.