A 10-year US Treasury yields 4.5%. Inflation runs at 3%. Tax takes 25% of the interest. Your net real return after all three: 0.4%. The asset class the world calls safe — the one that dominates most pension funds, most institutional portfolios, most retail savings — is delivering near-zero purchasing power growth while frontier technology compounds at 20-25% annually. This is not a marginal miscalculation. It is the defining financial error of our era. And most of the world's savings are inside it.
Not investment advice. Not a recommendation to buy or sell. Research and long-horizon thinking only. Consult a qualified financial advisor before making any investment decision. This letter presents a point of view, not a prediction. The counter-arguments are real and are given a full hearing.
The 10-year US Treasury bond is the world's benchmark safe asset — the instrument against which every other investment is implicitly priced, the refuge capital seeks in every crisis, the foundation of most institutional portfolios on earth. In June 2026 it yields approximately 4.5%. That yield sounds, on the surface, like a reasonable return for a risk-free asset. It is not. What it actually delivers, once the three factors that determine real purchasing power are applied, is something considerably less reassuring.
US 10-Year Treasury · June 2026 · After all deductions
Let that number sit for a moment. For the privilege of lending your money to the most powerful government in the history of human civilisation, you receive a net real return, after inflation and tax, of 0.4% per year. Over ten years, your purchasing power barely moves. Over twenty years, it declines — because this calculation assumes inflation stays at exactly 3%, when the historical record suggests it rarely stays precisely where it is measured today, and often surprises to the upside. The asset the world calls the ultimate safe haven is, in real terms, a slow, quiet, almost imperceptible erosion of the wealth you worked to accumulate.
"The downside of bonds is a 100% loss of purchasing power over 20–30 years as inflation eats away at principal. The downside of frontier tech is an 80% drawdown — but with the potential for 10–50× returns. The expected value of optimism is mathematically superior."
If the maths is this clear — and it is — why do most investors, including many sophisticated ones, remain heavily allocated to bonds, cash, and other fixed-income instruments? The answer is not ignorance. It is psychology. The human brain is wired, through millions of years of evolution, to weight losses far more heavily than equivalent gains — what behavioural economists call loss aversion. A 30% portfolio drawdown feels catastrophically worse than a 30% gain feels good, even when the dollar amounts are identical. The volatility of equities is visible, dramatic, and emotionally painful. The erosion of purchasing power through inflation and low real yields is invisible, gradual, and emotionally painless — until it isn't.
This asymmetry in emotional weight creates a systematic preference for nominal certainty over real return. An investor holding a bond that yields 4.5% knows, with certainty, what they will receive in dollars. The fact that those dollars will buy 20-30% less in twenty years is abstract and distant. The human brain is exceptionally poor at processing slow, cumulative losses — which is precisely why inflation and low real yields are such effective wealth destroyers. They operate below the threshold of emotional alarm.
The most useful analytical tool for the bond-versus-frontier-tech decision is not a valuation model but a decision matrix — a framework that assigns probabilities to different futures and computes the expected outcome of each strategy across all of them. When you run this matrix honestly, the asymmetry between the two strategies becomes stark.
The matrix reveals the core asymmetry: in the collapse scenario, no strategy wins — bonds do not protect you from true systemic failure any more than equities do. In the moderate and strong optimism scenarios — which together represent the 80% of futures where civilisation continues — bonds deliver near-zero real returns while growth assets compound meaningfully. The expected value of the bond strategy is approximately 0.4% real across all scenarios, because it wins nowhere. The expected value of the frontier strategy is positive in 80% of scenarios and catastrophic in only 20% — the same 20% where bonds are also catastrophic.
In 2008, Amazon had a price-to-earnings ratio that was effectively infinite — because it had no earnings. It was burning cash building logistics infrastructure. Traditional analysts called it a speculation. Its stock price fell to $35 during the financial crisis. By 2020 it had reached $3,500. The investors who held through the volatility, who understood that Amazon was not a bookstore but the logistics backbone of global e-commerce, received a 100-fold return on the period. The investors who sold in 2008 and bought bonds at 3.5% received, net of inflation and tax, approximately zero.
The same pattern is repeating today — at an order of magnitude larger scale. AI companies are losing billions annually on compute costs. Fusion startups have no revenue and will not for a decade. Synthetic biology platforms are spending billions on research with uncertain commercial timelines. This is not evidence of irrationality. It is the rational pricing of survival infrastructure — technology that, if it succeeds, will generate cash flows so enormous that the companies owning it will pay dividends in the 2035-2045 period that dwarf anything available from fixed income today. You are not buying earnings. You are buying a claim on a future that fixed income explicitly does not participate in.
The case for bonds in 2026 is not zero — and this letter would be dishonest if it pretended otherwise. The liquidity argument is real: bonds can be sold at known prices in any market conditions, while equity and frontier tech positions may be illiquid precisely when you most need liquidity. The deflation scenario — which the bond maths above does not model — would make bonds genuinely attractive: if AI-driven productivity causes deflation rather than inflation, real bond yields improve dramatically while equity valuations compress. And the sequence-of-returns risk for investors near or in retirement is genuine: a 40-50% equity drawdown early in retirement is structurally different from the same drawdown for a 35-year-old investor with a 30-year horizon.
The 0.4% real return calculation also assumes current inflation persists at exactly 3%. If inflation falls to 2% — which is plausible in a world where AI drives productivity gains — the real yield on the Treasury improves to approximately 1.1% after tax. Not compelling, but not quite the trap framed above. And if rates rise significantly — as they did in 2022 — existing bondholders face capital losses on top of the low real yield, worsening the picture further.
The honest position: bonds are not the asset class for investors who understand the mathematics of purchasing power and have a long enough time horizon to absorb frontier technology volatility. They are, however, an appropriate allocation for investors with genuine short-term liquidity needs, genuine deflation exposure, or genuine sequence-of-returns risk from proximity to retirement. The error is not holding any bonds — it is holding the proportion of bonds that most institutional portfolios and most retail investors currently hold, which was designed for a different inflation and interest rate environment than the one we are actually in.
The bond trap is not a conspiracy. It is not a failure of the financial system to function as designed. It is the predictable outcome of a system built around nominal certainty in an era when purchasing power is what actually matters, and nominal yields are insufficient to preserve it. The investors who escape the trap are not those who found a secret — they are those who did the arithmetic, accepted the volatility that comes with genuine real returns, and gave their investments the time horizon required to realise them. 0.4% per year is not safety. It is the slow, quiet, mathematically inevitable transfer of purchasing power from those who mistake nominal certainty for financial security to those who accepted the discomfort of owning the future instead.
Long-horizon thinking on capital, technology, and the forces shaping the next decade of wealth creation. Written from first principles. Not consensus. Not noise.