Gold's Bloodbath Reversing as Dollar Crumbles | Clive Thompson

Thompson argues that AI’s technological progress may outpace its investment returns: more than $5 trillion could be spent by 2030 against roughly $100 billion to $200 billion in current gross revenue, while competition may compress pricing and favor lower-cost specialized services. He also says Anthropic, OpenAI, and Microsoft Copilot are credible contenders, but he would not yet bet on a single long-term winner.

Rising sovereign borrowing costs are being linked to worsening fiscal positions: the U.S. 30-year Treasury yield reached 5.41% amid a roughly $2 trillion annual deficit, while France’s 30-year yield reached 5.2% versus 3.845% for Germany because of budget and credit concerns. Thompson views these developments as evidence of broader debt-system risk, though he says any transition could involve financial repression, capital controls, or a new currency and its timing is unknown.

Thompson says elevated equity valuations can persist or rise further because government borrowing may channel savings into stocks, but the U.S. market’s roughly 27-times earnings multiple is well above normalized historical levels of about 16 to 18 times and therefore implies lower future returns. He cautions that valuation alone does not trigger a crash, while noting that low-strike reverse convertibles can add downside exposure despite offering roughly 7% to 8% interest.

Thompson favors a portfolio framework of 60% equities, 20% gold, and 20% other assets rather than a traditional 60/40 allocation, with bond exposure concentrated in short-dated government securities. His preference reflects concerns that long-duration bonds could be vulnerable to rising yields and that assets held through banks, insurers, or governments may face taxation, confiscation, or forced sales, as illustrated by Russian shares sold for 30% of market value.