JPMorgan's cheap-capital warning is not just a bond-market note. If rates settle around 4 to 5 percent, a lot of startup valuations from the last cycle were built on the wrong world.
JPMorgan's six-D framework has a simple message for anyone still waiting for money to get cheap again: stop waiting. The bank has been pointing clients to six forces shaping 2026 and beyond, deficits, deregulation, de-carbonization, de-population, de-globalization, and de-dollarization. Two do most of the work here. Governments are borrowing heavily, and the labor force is aging. That's enough to change the price of capital.
The debt numbers are ugly. The IMF's April 2026 Fiscal Monitor said global public debt rose to just under 94 percent of GDP in 2025 and is on track to hit 100 percent by 2029, one year earlier than it expected last year. The IMF had already warned in 2024 that public debt was expected to exceed $100 trillion. You don't need a heroic macro model to see the pressure. When states borrow this much, central banks are no longer the only actors setting the rate environment.
The demographic piece is just as structural. For decades, rising working-age populations in developed and emerging economies helped keep labor abundant, wages contained, and rates lower than they might otherwise have been. That dividend is fading in Japan, Germany, China, and the United States. Fewer workers mean tighter labor markets and heavier pension and health obligations. There's no easy policy fix for that. You can't print people.
If you accept that framework, 4 to 5 percent rates don't look like a temporary mistake by the Fed. They look like the new floor.
AI debt meets higher rates
The technology sector is where this stops being an abstract macro argument. JPMorgan Asset Management's Bond Bulletin, published this week, said AI-related bond issuance is becoming one of the most important areas in corporate credit, with hyperscalers turning more often to bonds, private credit, and off-balance-sheet structures to fund data centers. Its market materials also say the hyperscalers sold $121 billion of bonds in 2025 and are expected to ramp issuance to $250 billion in 2026.
That is not distress financing. The big cloud companies still have strong balance sheets, and JPMorgan Asset Management says average net debt-to-EBITDA for the group is near 0.8 times, well below the 2.6 times typical for investment-grade issuers. But don't confuse balance-sheet strength with immunity from repricing. If you raise hundreds of billions for assets whose payback depends on AI demand lasting for years, the rate you pay matters.
You can already see investors asking for more. The Financial Times reported today that Meta's latest $12 billion data-center financing in El Paso is expected to yield more than 7 percent, about 0.4 percentage points above bonds tied to its earlier $27 billion Hyperion facility. The Wall Street Journal also reported this week that BlackRock is leading the financing for that El Paso project, with Meta owning 20 percent and leasing capacity from the site once it is built. That is a real structure, with real rent, real debt, and maturities stretching into 2048.
Here's the thing founders should notice: the largest companies in the world can absorb this shift, and they still care about 40 basis points. A startup with no profits and a 2021 valuation should care a lot more.
The repricing problem for founders and VCs
For venture capital, JPMorgan's thesis is confirmation, not revelation. Hurdle rates moved when rates moved. A fund that deployed capital in 2020 and 2021 at exit values priced on 8x or 10x forward revenue was assuming the discount rate would stay close to zero. It didn't. The markdown cycle that followed wasn't a mood swing. It was arithmetic catching up.
Founders raising today face that arithmetic directly. The companies that raised at peak multiples in 2021 and are now trying to grow into those valuations are doing it in an environment where capital has a real cost again. Extension rounds, flat rounds, down rounds: the labels are softer than the underlying fact. The valuation was built in one rate regime, and the business is living in another.
The harder implication sits upstream, in how VCs size funds and model returns. A fund promising 3x net when the risk-free rate is 1 percent is asking investors to believe one thing. The same promise at 4.5 percent asks for something much harder: better exits, fewer write-offs, or both. Neither appears because a partner wants it in the model.
JPMorgan's six-D framework won't be the last word. Some of the D's are cleaner than others, and big macro theses always draw pushback. Fair enough. But debt and demography are not side issues. They are the machinery underneath the cost of money. If you built a company, a fund, or a data-center boom on the assumption that cheap capital was merely away for a while, the burden is now on you to prove the old world is coming back.
Also read: Tesla posted a $1.1 billion free cash flow deficit and investors finally stopped giving Musk the benefit of the doubt • Trump's crypto fortune may be the thing that kills the crypto bill he champions • China's CXMT lists on July 27 in Asia's biggest IPO of 2026 and it's already moving markets
