Newsletter · · Ashutosh Agarwal

Cheaper Gas, a Hawkish Fed, a Cracking Consumer - US Macro Recap - Week of July 14, 2026

US Macro Recap for the week of July 14, 2026. Cheaper June gasoline should flatter the first inflation print of the Kevin Warsh era even as the Fed openly debates a hike, while Apollo's Torsten Slok warns that about half of US growth now rests on AI spending and hyperscaler free cash flow is falling toward zero, and a 50% one-month jump in rent delinquencies signals the bottom of the K-shaped economy is cracking.

US Macro Recap

Week of July 14, 2026: Cheaper Gas, a Hawkish Fed, a Cracking Consumer

This week the whole macro world is holding its breath for one number. The June inflation report lands in the next couple of days, and it is the first one new Fed Chair Kevin Warsh actually gets to react to. The setup is strange: gasoline fell hard in June, so the headline should look tame, maybe even negative, yet the Fed is now openly talking about raising rates, the exact opposite of where it started the year. Underneath that, two quieter stories got louder. Apollo's chief economist laid out just how much of America's growth is now a single bet on AI (about half), and the free cash flow funding it is heading toward zero. And for the first time in a while, the strain on the bottom half of the household economy showed up as a hard, ugly number: rental late-payments jumped 50% in a single month.

TL;DR

  • June's inflation report (due this week, with producer prices right behind it) is the first of the Warsh era. Gas prices fell about 9% in June, so the headline should tick down, but the "core" reading that strips out food and energy is still running near 3.4%, and the market is now betting the Fed's next move is a hike, not a cut.

  • Apollo's Torsten Slok put hard numbers on it: of roughly 2% US growth this year, about half comes from AI spending, with the rest from a one-time tax-refund boost and reshoring, and the big AI spenders' spare cash is "dropping down toward zero," forcing them to borrow and issue stock to keep building.

  • The bottom of the "K" is fraying in real time, a landlord-software firm tracking tens of thousands of properties says rent delinquencies jumped from 7.5% to 11.5% between May and June, "a very clear sign that the consumer is stretched."

What's new

1. The first inflation print of the Warsh era is coming, and cheap gas will flatter it. On Bloomberg Daybreak: US Edition, Bloomberg Economics' Stuart Paul said to expect "a decline in the monthly headline CPI reading" because "gasoline prices fell about 9% during the month of June," while the core (the version that strips out volatile food and energy) comes in "about 0.2% month on month," nudged up by World Cup travel prices and a chip-shortage bump in appliances and electronics. His bigger point: markets are "pricing in a hike well before the end of the year," and he thinks they're wrong, "the market is over-anticipating a hike," because there are "no inflation pressures coming from the job market, from wages, from income." Nomura's The Week Ahead put similar math on it: core CPI "relatively stable at 0.215%," headline "slightly negative" on falling energy, but core PCE, the Fed's favorite gauge, "moderated slightly to 0.27%. That's still 3.3% annualized," well above the 2% target. Why it matters: this is the first inflation data Warsh can actually respond to at a meeting, and the read-through is a headline that looks soft on top of a core that refuses to break.

2. Half of US growth is now an AI bet, and the cash paying for it is drying up. The single richest interview of the week was Apollo chief economist Torsten Slok on The Real Eisman Playbook (July 13). His breakdown of roughly 2%+ growth this year: "AI about 1%, the Industrial Renaissance about 0.3, and 0.9 from the One Big Beautiful Bill" (Trump's tax-refund boost, the average refund rose from about $3,000 to about $4,000). The catch, in his words: the tax-refund piece "is indeed one time," so 2027 looks weaker. On the money behind AI, he was blunt: Google "spent 80 billion on AI" last year "and basically funded it from its own cash flow. And this year they're spending 190 billion, and they just raised 85 billion in equity." Across the big cloud builders, "the free cash flow for the hyperscalers has absolutely gone from being very, very high and literally is dropping down over the next six, twelve months towards zero." His rate call was equally direct: the odds of a Fed cut this year are "zero… the economy is too strong," and the market is now "pricing that the Fed will be hiking rates in September and in December." In his companion Weekly Wrap (July 10), Steve Eisman drove home why this matters for everyone's portfolio: tech is "38% of the S&P 500," north of 50% once you add Google and Amazon, and "15% of all corporate existing debt is AI-related, and 50% of all newly originated corporate debt in 2026 is AI-related." His conclusion: a plain 60/40 stock-bond mix "does not create real diversification. It's all one trade. It's all AI."

3. The bottom of the "K" just cracked, rent delinquencies jumped 50% in a month. On One Rental At A Time (July 12), Dion of the landlord-software firm Hemlane, which sees payment data across tens of thousands of rentals, reported that delinquencies "have gone from about 7.5% to 11.5%" between May and June, a 50% jump, and "it was just overall nationwide." Veteran landlord Michael Zuber, who's rented for 25 years, said June "is not historically a spike in delinquencies" (that's usually January, after the holidays), so this "is a very clear sign that the consumer is stretched… renters just are so close to the edge that they're choosing to be delinquent and to pay the penalty." Slok's data (above) is the macro version of the same story: the bottom 20% of households, those earning under $25,000, have savings that in dollar terms are "exactly the same as where it was in 2019," while the top 20% have "a trillion and a half more." He described "a K-shaped situation" now visible in wealth, wage growth and inflation at once, with "delinquency rates on auto loans… credit cards… and student loans" all rising for the middle and bottom. The saving grace: the top 20% do "40% of consumer spending" versus "8%" for the bottom 20%, so aggregate spending "is actually still okay." The economy holds up because the wealthy carry it, not because the bottom half is fine.

4. A veteran strategist says there's been zero net job creation in a year, and policy has quietly turned tight. On Excess Returns (July 11), Jim Paulsen walked through the charts that turned him cautious. Averaging the two job surveys, "they average out to zero in the last year", payrolls up 0.3%, the household survey down 0.3%. "Nowhere in our history that I've been in the business going back to the early eighties was zero job creation over any 12-month period considered something that was okay," he said, and yet "most people thought the Fed is going to raise rates in the face of it." He flagged part-time work and the unemployment rate rising together ("things that only occur in the middle of recessions"), housing starts "about as bad as it was in the worst of the housing crisis in 2009," real disposable income (excluding government transfers) "deeply in negative territory," and Atlanta Fed GDP tracking "now down for this quarter to one and a quarter percent." His overarching point: the Iran conflict quietly re-tightened policy, the 10-year yield up "about 70 basis points," the dollar up "5% to 6%," real money-supply growth back "to about a quarter of a percent", and "with a lag, that's going to hit."

5. The bond market can't agree with itself on inflation. A "breakeven" is simply the inflation rate baked into bond prices, the gap between a normal Treasury yield and an inflation-protected one. Last week the story was breakevens crashing (screaming disinflation). This week the picture got muddier. On The Nasdaq Dorsey Wright Podcast (July 9), the team showed the 10-year breakeven rising even as oil fell, markets "pricing in higher average inflation over the next 10 years… not directly attributable to the commodity spike," pinned partly on tariffs, citing New York Fed research that "roughly 45%" of firms "are not finished passing through higher prices" from tariffs. Cutting the other way, on The KE Report (July 7), Craig Hemke pointed out the round-trip in oil, crude "went from 65 to 110" from March through May and is now "back to 68," which drove the PCE energy component up "21%" and is now set to reverse: "those break-evens from the 2-year and the 5-year and the 10-year, they're all falling." Same instrument, opposite readings, which is exactly why this week's CPI print is being watched so closely.

The debate

All three camps showed up in this week's podcasts, so here's each one, steel-manned from the audio that actually carried it.

Disinflation, hold-then-cut (economists and analysts). The cleanest version came from The MUFG Global Markets Podcast (July 10): headline CPI at 4.2% but expected "to trend lower toward 2%" as rents lag lower, tariff effects roll out of the year-on-year math "by Q1 next year," and energy disinflates, so "the Fed's next move will be rate cuts rather than hikes." On RiskReversal Pod (July 13), strategist Liz Thomas expected June CPI at "3.8%… down from 4.2%," and argued "CPI probably has already peaked" if oil stays de-escalated, adding the case against hikes plainly: "you hike rates when you're trying to contain an economy that's overheating… I don't think that's the situation we're in. You don't hike rates into a slowing economy." Craig Hemke (above) is the market-priced version of the same view.

Sticky / reflation, hold, maybe even hike (the Fed itself, plus sell-side). This is the camp that hardened this week. Slok (above) sees "zero" chance of a cut and a market pricing "two hikes." On Kyle Talks (July 11), the host walked through the latest core PCE at "3.4% annually… the highest since October 2023," with headline at "4.1%," calling it "sticky inflation" that "keeps interest rates elevated and borrowing expensive," with oil (Brent "at $78.55") a key culprit. CFA Society Chicago (July 7) flagged services inflation still sticky on labor costs and tariff-driven supply-chain restructuring that could keep inflation from returning to 2%. And Market Maker (July 9) noted the Fed's own minutes had officials warning that if "AI-driven demand, tariffs and Middle East oil shocks keep price pressures high, some policy firming would likely be warranted." One rung further out, on Get Rich Education (July 13), Keith Weinhold made the "permanent inflation" case: CPI "reheated to 4.2%," now above the 2% target for "63 consecutive months," which he views as structurally embedded rather than transitory.

Stall speed (labor and consumer data). The most data-grounded bear case ran through Jim Paulsen (above), zero net job creation, housing at 2009 levels, GDP tracking at 1.25%, and Rebel Capitalist News (July 6), which flagged Atlanta Fed GDP tracking at "1.2%" and predicted it drops "below 1%" once June's weak 57,000 payrolls (versus ~115,000 expected) are folded in. On The Julia La Roche Show, Danielle DiMartino Booth's earlier warning still frames it, labor-force participation at a 50-year low, "720,000 Americans" leaving the workforce in one month, and the rent-delinquency spike (above) is the household-level evidence. The through-line: the headline unemployment rate (4.2%) looks fine mostly because the labor force is shrinking, not because hiring is strong.

Worth naming the honest tension for readers: the same economy is being called "too strong to cut" (Slok) and "zero job creation, stall speed" (Paulsen) in the same week. Both can be partly right, the strength is concentrated in AI, capital spending and the top-income consumer, while the broad labor market and the bottom half look soft.

The labor and immigration mechanism sits underneath all of this: a shrinking labor force means the economy needs fewer net new jobs each month to hold unemployment flat, which is part of why a 4.2% unemployment rate can coexist with what Paulsen calls roughly zero net hiring over the past year.

The trades in play

Plenty of instrument-level positioning came through, and it clustered on the front end of rates, the dollar, precious metals, and the rotation out of crowded tech.

  • Sell the Fed's hike bets; short the dollar against EM; go long silver (operator). On The Macro Trading Floor (July 10), the guest, a discretionary macro trader who runs a roughly 10-day holding period, laid out a "Goldilocks" view (growth okay, inflation reverting to a bit above target but "not going out of control") and the trades to monetize it: receive 2-year rates near "4.2%" to collect carry against hikes he doesn't think will fully happen (a "negative skew, high probability" bet), short the dollar against positive-carry emerging-market currencies ("dollar Brazil… dollar Mexico"), and "long silver and gold" as "very bombed out," eyeing silver "back at 63, 64, maybe even up to 70 by August" from around 58.

  • Bullish carry, bullish dollar, but be selective (sell-side strategists). JPMorgan's FX team on At Any Rate (July 10) said their "bullish carry and bullish US dollar barbell" can "withstand even a more serious escalation in the Middle East" because the carry baskets are financed with energy-importing, low-yielding currencies. They favor high-beta Swiss-franc baskets over the Swedish krona now that "we've transitioned over to a carry regime," and see euro/sterling heading to an "84 handle" as UK political risk fades. Their big watch item: Japan floated using its giant GPIF pension fund to buy yen, a shift they estimate at "about 12 trillion yen," comparable to this spring's interventions, which "multiplies that intervention risk" for anyone short the yen.

  • Overweight small caps and "large-cap growth as value"; light on bonds and Europe (operator). On Bloomberg Surveillance (July 13), Federated Hermes CIO Steve Auth, running real money, said small caps "are widely outperforming everything else, so we're overweight there," argued "the large-cap growth stocks have become value stocks now," and is underweight "bonds and European equities" while overweight emerging markets as "an alternative tech investment at a cheaper price." His three-year target: "the S&P at 9,000… against an earnings number of 450."

  • Rotate from expensive semis toward the cheaper mega-caps (analysts). DataTrek's Nick Colas and Jessica Rabe on The Compound and Friends (July 13) noted tech has "lagged the S&P by 5.6 points" since a "six-standard-deviation" outperformance peak on June 2, and see mean reversion ahead: the top five semis trade at "52.5 times forward earnings" versus "25.9 times" for the rest of the Magnificent 8 (ex-Tesla), "a double." Their call: expect this year's tech laggards to play catch-up in the second half, "more mechanical than fundamental."

  • A defined-risk energy bet, and metals-miner call options (operators). On Macro Voices (July 9), the show's options segment framed an energy trade in the XLE ETF (~$55.25) using a collar, long a $50 put, short a $65 call for a small net cost, for "$10 upside potential and $5 downside hedge," while flagging crowded long positioning in S&P futures and the dollar. And Craig Hemke (above) said he "bought 10 $80 GDX calls" for August, betting gold-miner earnings "are going to be a lot better than people might be expecting" once the energy-cost scare reverses.

Read-throughs

  • The AI trade is quietly becoming a credit trade. If hyperscaler free cash flow is heading "toward zero" (Slok) and half of newly issued corporate debt this year is AI-related (Eisman), then the risk migrates from stock prices to bond spreads. Slok pointed to one specific pressure point: software, where "$500 billion" of the roughly "$2 trillion" private-credit market is loans made in 2021–22 that hit a "maturity wall" in "2028 and 2029." You can already buy some of that paper "in the open market" at "12%" yields. His reassurance is that, unlike 2008 subprime, this debt sits on lightly levered balance sheets, business-development companies are "only levered twice", so it's a real problem for the lenders, but "not anywhere near those systemic levels."

  • The Fed is deliberately hard to read, right as data starts to matter most. Both Slok and Nomura noted Warsh has dropped precise forward guidance and shifted the minutes to "scenario-based" language. With no official signal to anchor to, every data point, starting with this week's CPI, will swing rate expectations harder. As Slok put it, "when you don't give forward guidance, the market has to start guessing."

  • Bank earnings kick off the season this week. Eisman flagged that "earnings season begins with the banks, as usual, reporting first." With the hyperscalers reporting later this month and their stock prices no longer rewarding capex (Meta and Oracle were both cited as pulling back), the DataTrek crew raised the 1999 parallel: when the spenders' shares stop reacting to good news, the component suppliers, the semis, are usually the last to look down. That's the fear priced into the semi complex right now.

What changed

Last Friday's recap was built around two market-priced signals: breakevens "crashing" (the bond market screaming disinflation) and the labor force shrinking to a 50-year low. Three things shifted this week.

First, the disinflation-vs-reflation call is now genuinely contested inside the bond market itself, where last week breakevens were falling hard, this week the 10-year breakeven was shown rising even as oil dropped, pinned on tariff pass-through rather than energy. Second, the reflation/hike camp hardened from "the Fed is split" to "the market is pricing two hikes" (September and December, per Slok), a notable move from a week ago. Third, the consumer-stress story finally got a hard, fresh number: rent delinquencies up 50% month-on-month, the clearest single data point yet that the bottom of the "K" is buckling. And the "AI is the economy" theme sharpened from a debate about GDP math into a specific, uncomfortable fact set, half of US growth, hyperscaler cash flow to zero, and half of all new corporate debt tied to the same trade. The one big known unknown that reframes all of it, this week's CPI print, is still ahead of us as this goes out.