Daily Macro Brief
Long-End Yields Rise Again as AI Rewards Delivery
U.S. ECI rose 0.9% for a second straight quarter and consumer sentiment improved while China’s manufacturing PMI fell to 49.2; the 30-year Treasury reached 5.27% intraday and Amazon gained 15.1%, as markets priced both sticky inflation and AI spending that is producing revenue.
This report is based on intraday data as of 12:56 PM ET and does not reflect closing prices. Markets may have moved since publication.
Revisiting Yesterday’s View
Yesterday’s conclusion was that softer monthly inflation could no longer pull down the long end, shifting the problem from the policy rate to the term premium. Today’s second straight 0.9% ECI reading and simultaneous increases in long U.S. and Japanese yields turn that clue from a one-day Treasury move into a broader question about fiscal and policy credibility.
Today’s Core View
The long end resisted soft inflation yesterday and rose again today after a 0.9% ECI reading and an unchanged BOJ rate. Markets are reframing the problem from monthly inflation to global fiscal and policy credibility, while AI rewards only capital spending already producing revenue.
Macro and Geopolitical Analysis
The U.S. data do not show a fresh acceleration. They show nominal demand becoming sticky again. The BLS reported a 0.9% quarterly increase in Q2 ECI, above the 0.8% consensus and unchanged from Q1. Wages rose 0.9%, benefits increased 1.0%, and total compensation was still up 3.4% from a year earlier. On the same day, University of Michigan consumer sentiment rose from 49.5 to 55.2, while the Chicago PMI increased from 56.7 to 57.6. Together, the releases show improving business activity and confidence without another step down in labor costs. Yesterday’s softer PCE data therefore failed to produce easier long-term financing conditions.
This was not an unambiguously bullish data package. Sentiment remains roughly 11% below last July’s 61.7, and the BLS reported a 0.4% annual decline in inflation-adjusted private-sector wages. Nominal activity is resilient enough to delay policy relief, but purchasing power is not strong enough to confirm prosperity. That is the hardest middle ground for bond markets: growth is not weak enough to force a policy turn, while inflation is not low enough for long-duration investors to ignore fiscal supply.
U.S. and Chinese manufacturing moved in opposite directions on the same day. China’s official manufacturing PMI fell from 50.3 to 49.2 in July, below the 50.0 consensus, as the Chicago PMI rose to 57.6. China’s large-cap ETF is still up 15.0% over one month with an RSI of 77.5, meaning prices are anticipating policy support before current data confirm a cyclical turn. Copper was almost unchanged intraday and crude remained down 5.2% over five days, another sign that the global growth impulse is not synchronized. Sticky U.S. services and labor costs cannot be translated directly into a broad industrial recovery.
The apparent improvement in Hormuz remains an improvement in measurement, not normalization of the route. Kpler’s narrow count recorded only two transits on July 30, CNN’s broader 24-hour snapshot counted roughly 15 commercial vessels, and Bloomberg observed at least seven vessel pairs conducting transfers off Sohar, Oman. The figures can all be true: crude can leave the Gulf through offshore transfers while regular, insurable direct voyages remain far below prewar levels. WTI was $84.68 intraday, up 1.3% for the day, down 5.2% over five days, and up 21.8% over one month. Markets are discounting the marginal improvement in getting barrels out, but they have not removed the physical constraint from prices.
That defines the geopolitical Kill Switch. Hormuz can be downgraded from a structural constraint to an event premium only when direct traffic, insurance coverage, navigation rules, and military de-escalation improve together. More offshore transfers reduce scarcity risk; they do not prove that the strait has reopened.
Devil’s Advocate: today’s increase in long yields may still contain month-end flows rather than new regime evidence. Consumer sentiment remains low in absolute terms, Chinese manufacturing is back in contraction, and crude has declined over five days. Those facts support a counter-scenario in which slower growth eventually pulls yields down. It would require the 30-year Treasury to move back below 5%, the JGB long end to decline with it, and subsequent ECI or core inflation readings to shift clearly lower. Until then, treating today as temporary noise still lacks price confirmation.
Bond Market
The long end delivered cross-market confirmation today. The 30-year Treasury reached 5.27% intraday, remained above the 5% threshold, and stayed near a 52-week high; the 10-year reached 4.74%. In Japan, the BOJ voted 8-1 to keep its policy rate at 1%, with the lone dissenter favoring 1.25%, while warning that core inflation could move clearly above 2% from September. The outcome did not protect JGBs: 10- and 30-year yields continued to rise.
| Market | Intraday Yield | 1-Day Move | Interpretation |
|---|---|---|---|
| UST 10Y | 4.74% | about +0.1 percentage point | the Treasury source rounds the daily change to one decimal place |
| UST 30Y | 5.27% | about +0.1 percentage point | 5% has shifted from warning line to persistent range |
| JGB 10Y | 2.801% | +4.4bp | the BOJ stayed put, but the long end still demanded more compensation |
| JGB 30Y | 3.971% | +3.0bp | approaching the 4% threshold again |
Bloomberg estimated from BOJ account data that the suspected July 30 currency intervention totaled about ¥8.45T, although the Ministry of Finance has not confirmed it. USD/JPY stood at 159.27 at 12:56 PM ET. That was still below the pre-intervention area but roughly 0.9 yen above yesterday’s intraday low of 158.34. The message is direct: currency operations can create time, but if the rate reaction function does not follow, pressure migrates to long government bonds and the next policy meeting.
Taken together, the U.S. and Japanese long ends are saying the same thing. Central banks can look through a supply shock and fiscal authorities can intervene in currencies, but long-duration capital will still demand compensation for inflation tails, bond supply, and policy credibility. That is closer to a tradable definition of Fiscal Dominance than any single meeting in either country.
Sector Spotlight
AI platforms and infrastructure: markets rewarded only capital spending that has already become revenue. Amazon gained 15.1% intraday after Q2 revenue reached $200.6B, up 20%, while AWS revenue increased 37% to $42.2B and AWS operating income reached $16.6B. Management raised planned 2026 CapEx from about $200B to $220B and still expects capacity to fall short of demand through 2027. Higher spending received a positive response because cloud revenue, profit, and backlog all validated the demand behind it.
At the same time, GOOG rose 5.8% and VRT gained 8.3%, although VRT was still down 15.2% over five days with an RSI of 29.2; MU declined 4.4%. This was not an indiscriminate recovery across the AI chain. Platforms received monetization evidence first, while infrastructure and memory suppliers still need delivery cadence to prove that supply is not running ahead of revenue.
Digital assets: weak results amplified a modest decline in the underlying asset into double-digit company moves. BTC fell 3.0% intraday, COIN declined 12.0%, and MSTR fell 5.3%. Coinbase reported Q2 revenue of $1.22B and a net loss of $359.5M, marking a third consecutive quarter below market expectations. With the Nasdaq 100 still up 0.5% and VIX down 1.1%, this was not broad Risk-Off. It was a specific reassessment of digital-asset activity and business-model leverage.
Japanese equities: a 4.0% gain is repair, not reversal. The Nikkei rebounded 4.0% intraday but remained down 3.1% over five days and 8.1% over one month, with an RSI of 32.5. The unchanged BOJ rate and Amazon’s results triggered a regional technology rebound, but the intervention boost began to fade while long JGB yields kept rising. Risk tolerance recovered today; the macro constraint did not.
What to Watch Next
August 2, OPEC+ core-seven meeting: Watch whether the roughly 188K bpd September increase becomes the final expansion of 2026. If further increases pause while direct Hormuz traffic remains depressed, the five-day decline in crude will look like temporary relief from offshore transfers. If expansion continues, producers will have delivered the first policy signal large enough to offset part of the shipping risk.
August 3 after market, Palantir Q2: The key test is not whether one quarter barely clears estimates, but whether commercial demand, backlog, and guidance show that the software layer can convert AI spending into revenue as AWS has. If the story persists without faster delivery, today’s platform-versus-supply split will extend into software.
August 4 at 8:30 AM ET, June U.S. trade data: Watch whether the Q2 import surge has begun to reverse. A decline before tariffs take effect would establish the starting point for a mechanical Q3 improvement in net exports. Persistently high imports would make the front-loading argument look more like continued inventory rebuilding than a one-time timing shift.
August 4 around 4:30 PM ET, API; August 5 at 10:30 AM ET, EIA: Focus on crude imports, distillate inventories, refinery utilization, and whether SPR stocks decline again. If offshore transfers increase without higher U.S. imports, the logistics improvement still has not reached physical U.S. flows. If imports and inventories improve together, there will be stronger grounds for another reduction in the Hormuz premium.
August 6, Constellation; August 7, Vistra: The power segment needs revenue and load evidence independent of a semiconductor rebound. If data-center demand lifts only the distant narrative without improving near-term contracts and earnings, this second-order AI segment will continue to follow rates rather than compute demand.
Disclaimer
This article is public market commentary and personal research notes. It does not constitute investment advice.
