July 2026 Market Update: Rising Yields, AI Under Scrutiny, and the Liquidity Question in Private Markets

The Bottom Line

July was a difficult month in public markets and an instructive one. The S&P 500 finished essentially flat, the Nasdaq fell more than 3%, long-term Treasury yields reached a 19-year high, and oil briefly crossed $100 before retreating. Behind each of those moves sat a single question the market is now asking insistently: is the enormous capital being committed to artificial intelligence going to earn a return?

That question is usually framed as a technology-stock story. For families holding meaningful private allocations, it isn't. The same capital wave is reshaping infrastructure, power generation, and commercial real estate — and it is arriving at a moment when private markets have a liquidity problem of their own.

Both things deserve attention this month.

Key Market and Economic Drivers in July

• The S&P 500 fell 0.1% and the Nasdaq declined 3.2%, while the Dow Jones Industrial Average rose 0.3%.
• Volatility rose mid-month, with the VIX reaching 21 before settling near 16.
• International developed markets gained 1.9% (MSCI EAFE, in U.S. dollar terms), while emerging markets fell 3.3% (MSCI EM).
• The 30-year Treasury yield closed near 5.28%, a 19-year high; the 10-year ended at 4.74%. The Bloomberg U.S. Aggregate fell 1.3%.
• Brent crude briefly exceeded $100 before closing near $90; WTI ended around $85.
• The Dollar Index slipped below 100, the yen weakened to roughly 157, and gold was little changed near $4,050.
• Second-quarter real GDP grew at an annualized 1.5%, down from 2.1% in the first quarter.
• The Federal Reserve held rates at 3.50%–3.75% in a 9–3 vote.

The Market Began Auditing AI Spending

Second-quarter earnings turned the AI conversation from enthusiasm to arithmetic. The largest technology companies are committing hundreds of billions to data centres and computing infrastructure, and investors spent July scrutinizing free cash flow rather than applauding ambition.

The scale is genuinely difficult to absorb. Big-technology capital expenditure exceeded $400 billion in 2025 and is expected to rise roughly 75% in 2026. Goldman Sachs Research projects $765 billion of AI-related capital spending this year, reaching $1.6 trillion by 2031. Data-centre construction alone now exceeds every other category of office construction in the United States.

Three developments sharpened the scrutiny. Global semiconductor shares corrected hard — South Korea's KOSPI 200 fell 24% in July. Moonshot AI released Kimi K3, an open-weight model reportedly competitive with the leading proprietary systems, raising questions about how durable any single company's advantage will prove. And Fitch flagged what it called major credit risk across the AI ecosystem, citing the interconnected nature of financing among the principal players.

None of that means the investment is misplaced. It means the market has stopped taking the returns on faith — which is a healthier posture than the alternative, and a noisier one.

A Divided Fed and the Highest Long Yields in Nearly Two Decades

The Federal Open Market Committee held rates at 3.50%–3.75%, but the vote was 9–3, with three officials favoring an increase. Dissent at that level is rare; the last comparable split was in September 2016.

Chair Kevin Warsh has deliberately narrowed the Fed's forward guidance. The post-meeting statement is shorter, and he has declined to speculate about responses to hypothetical scenarios. Less guidance means more interpretation, and the market interpreted July by pushing yields to their highest levels in years. Futures now imply a possible increase by October and potentially two by mid-2027.

For families holding substantial fixed income, the practical consequence is twofold: more volatility in yields, and materially better income than has been available for most of the past fifteen years. Both are worth planning around.

Oil, the Strait of Hormuz, and a Wider Conflict

Renewed U.S. airstrikes on Iranian military sites slowed traffic through the Strait of Hormuz, and the conflict widened when Houthi forces struck Saudi tankers in the Bab al-Mandeb Strait. Brent moved from roughly $72 in early July to above $100 before closing near $90.

Gasoline remains near $4.10 a gallon nationally, which keeps upward pressure on headline inflation at precisely the moment the Fed is debating whether it has done enough.

The Part That Matters Most If You Hold Private Assets

Public markets are the visible half of the picture. For families with private equity, private credit, direct real estate, or energy and infrastructure exposure, four things happened this year that deserve more attention than July's index moves.

Private equity: the question is liquidity, not marks

Exit activity remains suppressed. Deal volumes in June ran roughly 5.6% below June 2025 and substantially below January's pace. Distributions have slowed accordingly, and limited partners are feeling it.

The industry has responded by turning to continuation vehicles and the secondary market as the release valve. Secondary transaction volume reached an estimated $40 billion in the first quarter alone, split roughly 45/55 between LP-interest sales and GP-led transactions. Dry powder dedicated to secondaries has fallen below one year of transaction volume — a supply-and-demand imbalance that generally favors buyers.

The phrase circulating among allocators this year is that DPI has become the defining metric — distributions to paid-in capital, not internal rates of return. Investors have grown skeptical of paper marks and want to see cash returned.

What this means practically. If you have committed capital in funds approaching the end of their intended life, assume hold periods extend. Model your liquidity accordingly, and treat a fund's reported IRR as a claim rather than a fact until distributions support it. If you need liquidity, the secondary market is functioning — but pricing currently favors the buyer.

Private credit: compensated less for more risk

Private credit has been the most enthusiastically adopted alternative of the past decade, and this year it is being tested.

The U.S. private credit default rate stood at 6.0% for the twelve months ended May 2026, matching April's record. Proskauer's index, tracking 697 loans across roughly $189 billion, recorded a 2.73% default rate in the first quarter — up from 1.84% two quarters earlier. Defaults are not catastrophic, but the direction is unambiguous.

Meanwhile, competition has compressed both spreads and fees. Two-thirds of managers surveyed cite competition as the primary pressure on performance.

That combination deserves stating plainly: rising defaults alongside compressing spreads means investors are being paid less to take more risk. That is not an argument for abandoning the asset class. It is an argument for knowing precisely what you own — underwriting standards, covenant quality, sector concentration, and how your manager marks positions that stop performing.

Infrastructure and energy: where the AI money actually lands

Here the public and private stories converge, and it is the most interesting development of the year.

The binding constraint on AI expansion is no longer capital or chips. It is electricity. Global data-centre power demand is expected to rise 27% in 2026, reaching 132 gigawatts against 104 last year. In the United States, demand climbs from 31 gigawatts in 2025 to a projected 41 this year and 66 the year after.

Public grids cannot deliver that on the required timeline. The consequence is a strategic pivot toward dedicated on-site generation — and an enormous flow of capital into power generation, transmission, grid equipment, and the land and cooling infrastructure that data centres require.

For Texas families in particular, this is not an abstraction. It touches natural gas demand, grid infrastructure, land values in the corridors where data centres are being sited, and the economics of energy assets that many families here already own. If your wealth includes energy or mineral interests, the AI capital wave is arriving in your portfolio whether or not you own a single technology share.

Real estate: a genuine recovery, very unevenly distributed

Commercial real estate has moved past crisis into recovery, though the average conceals enormous variation. CBRE forecasts investment activity rising 16% in 2026 to $562 billion, near the pre-pandemic average; Colliers projects a 15–20% increase in sales volume. Cap rates are expected to compress modestly, on the order of 5 to 15 basis points, and lending has returned as banks, insurers, and debt funds re-enter.

The dispersion is the story. Prime, well-located assets are leasing and financing well. Commodity office and older buildings remain the most likely source of discounted sales, recapitalizations, and genuine distress. Data-centre and industrial assets sit in a different market altogether, driven by the demand described above.

If you hold direct real estate, this is a year to be specific rather than directional. “Real estate is recovering” is true and nearly useless. Which asset, which submarket, and what the debt maturity schedule looks like will determine your outcome far more than the sector trend.

How We're Approaching It

We are not making a call on artificial intelligence, or on whether the Fed raises in October. Both are genuinely uncertain, and portfolios built on a single forecast tend to be fragile.

What July reinforced is more prosaic. Rates are high enough that fixed income does real work in a portfolio again. Private allocations need liquidity planning that assumes longer holds than the offering documents suggested. Private credit warrants harder questions than it did three years ago. And the infrastructure and energy build-out is a genuine, multi-year theme with implications for families who already hold those assets.

The families who navigate this well will be the ones who know what they own, in detail, and have planned for the possibility that some of it takes longer to become liquid than they expected. For a fuller picture of how we see the second half, see our mid-year 2026 outlook.