The global economy is entering an unusual moment: borrowing is becoming more expensive, government debt is expanding, yet equity markets continue to march higher.
Across major economies, hawkish central banks and relentless government borrowing have pushed bond yields toward levels not seen in decades. Japan, long associated with ultra-low interest rates, has become one of the clearest symbols of this transformation.
With its 10-year government bond yield reaching 3% for the first time since 1996. Normally, rising yields would be expected to cast a shadow over stocks. Higher government borrowing costs raise the price of capital.
Register for the next Tekedia Mini-MBA.
Register for Tekedia AI in Business Masterclass.
Join Tekedia Capital Syndicate and co-invest in great global startups.
Increase financing expenses for companies and make relatively safe bonds more attractive compared with equities. But markets have so far refused to follow the traditional script. Stocks remain green, even as the foundations beneath them become more expensive.
Part of the explanation lies in the extraordinary concentration of the current rally.
Fewer companies are carrying the broader market, with investors increasingly clustering around businesses perceived to have durable earnings power and exposure to structural technological trends. At the center of that enthusiasm is artificial intelligence.
The AI boom has developed into something larger than a technology story. It has become an industrial spending cycle involving semiconductors, cloud computing, electricity, networking equipment and enormous data centers.
Every new model requires more computing power, and every leap in capability appears to create another appetite for infrastructure. Anthropic illustrates the scale of this spending frenzy.
The company has reportedly committed to enormous amounts of computing capacity, including a reported $35 billion cloud agreement with Nvidia-backed Lambda for a massive new data center in Texas.
That comes only days after Anthropic reportedly committed another $45 billion to Nscale for compute infrastructure in West Virginia. These numbers are staggering because they reveal the new economics of artificial intelligence.
The race is no longer simply about who can build the smartest model. It is about who can secure enough chips, power, cooling, data-center capacity and computing resources to operate those models at extraordinary scale.
The irony is difficult to miss. On one side of the global economy, governments are borrowing aggressively, pushing bond markets to demand greater compensation for holding public debt.
On the other, technology companies are committing tens of billions of dollars to an infrastructure race whose future returns remain difficult to measure. Yet investors continue to believe that AI could generate enough productivity, revenue and economic value to justify the spending.
That optimism has become a powerful counterweight to rising yields. Still, the market’s resilience should not be confused with immunity. If yields continue climbing, valuation pressures could eventually become harder to ignore.
Expensive capital can change corporate behavior, reduce investment outside the strongest sectors and expose companies whose business models depend heavily on cheap financing. For now, the financial world appears caught between two machines.
The first is the government borrowing machine, steadily issuing debt into a market demanding higher yields. The second is the AI investment machine, consuming capital at breathtaking speed in pursuit of the next technological frontier.
And both machines are humming. The question is not whether the money is moving. It clearly is. The deeper question is whether the productivity and profits eventually arrive quickly enough to justify the extraordinary capital being deployed today.
For now, the market’s answer is simple: keep spending, keep building and keep betting on AI. Brrr goes the money machine.