The August jobs report showed payrolls up 162,000 against expectations near 55,000, the best month since March, with June and July revised up by a combined 55,000. Unemployment held at 4.1%. Wage growth slowed to 3.1%, which is the one soft spot. Yields rose, the S&P 500 slipped 0.4% to 7,719, and futures now price a September hike as roughly a coin flip leaning yes, up from about a third before Jackson Hole.
That completes a reversal. On August 10 the July print of minus 23,000 had the market convinced the hike debate was over. It was one number, and it has now been revised to a gain. Chair Warsh, speaking at Jackson Hole on August 28, said the summer's inflation readings did not tell him underlying trends had improved and that otherwise the Fed has work to do. Core PCE was 3.3% in July with a six-month annualized rate above 4%. The Fed meets on September 15 and 16. There is nothing in the data that stops them.
The Market Is Still Pricing an Easy Fed. It Should Not Be.
Look at what has led since the July 29 meeting: Microsoft's record day, the memory rebound, Bitcoin's 22% week, gold at $4,600, Tesla up 18% in August on robotaxi headlines. Every one of those is a bet on liquidity, and several of them were made on the back of a jobs report that has since been revised away. A Fed that hikes into a 30-year yield already above 5% is the opposite of liquidity.
Here is the test for anything in a portfolio this week: does the thesis need the Fed to stand still? If the answer is yes, the position is exposed. Three groups pass the test and one fails it badly.
What Survives a Hike
- Memory. Micron rose 6% on Friday, on a day the market fell, after saying it will double AI memory output in 2027. Nvidia guided its current quarter to $108 billion and described fiscal 2028 as supply-constrained. When a supplier is sold out through next year, its earnings do not depend on the funds rate; they depend on the number of wafers it can ship. This has been the through-line since June and it is intact.
- Banks. Higher for longer is margin. The second-quarter trading records were driven by exactly the kind of volatility a hawkish surprise creates. A 25 basis point hike is a tailwind for net interest income, and the group has not re-rated for it.
- Short Treasuries and cash. Bills above 4% with a Fed that may raise them is not exciting, and that is the point. It is where money waits while the long end reprices.
The group that fails is the one that led August: long-duration megacaps and the liquidity trades. Broadcom just reported AI revenue up 221% with a $230 billion target for fiscal 2028 and fell 6% because the near-term guide and margin were merely fine. The market's July rule, that only growth with controlled spending gets rewarded, is still in force, and a hike raises the bar again.
This Week's Three Tests
- August CPI, the last inflation reading before the meeting. July printed 3.4% headline and 2.5% core. Gasoline rose in August as the Hormuz truce lapsed and WTI climbed back above $90, so the headline is at risk of ticking up. A core reading above 2.6% alongside Friday's payrolls makes a hike the base case rather than a coin flip.
- Apple, Wednesday. The iPhone 18 event is John Ternus's first as chief executive. The product will be fine. The number that matters is pricing. Apple cut its guidance in July because memory was too expensive, and Tim Cook said the impact was increasing. If iPhone 18 prices rise, it confirms the memory shortage is being passed to consumers, which is good for Micron and Sandisk and a warning about the second-half consumer. If prices hold, Apple is eating the cost, and the margin guide in October gets harder.
- Oracle, due in the coming days. In June the stock lost 11% on a $40 billion financing plan and negative free cash flow. Every question from that report is still open, and a hike makes the debt-funded data center model more expensive. Watch the funding commentary, not the revenue.
Bitcoin and the Liquidity Trades
On August 10 the case for Bitcoin rested on liquidity turning: exhausted ETF outflows, a Fed that had stopped getting more hawkish, and gold moving first. Two of those three held, the flows and the gold signal, and Bitcoin ran from $64,000 to about $79,000. The third did not. A September hike is a liquidity withdrawal, and the honest reading is that the August rally front-ran an easing that the data has now postponed.
That does not break the thesis. The invalidation level was a weekly close below $55,000, and price is 40% above it. But a position that ran 22% in a week into a hawkish Fed is not one to add to, and the third tranche that was meant to confirm the breakout above $70,000 should be treated as the last purchase until the meeting is behind us. The same logic applies to Tesla after its 18% August, where Friday's mixed Cybercab reviews took 6% back in a day.
Key Levels Into the Fed
| Probability of a September hike, current pricing | Roughly 55% to 65% |
| August core CPI, level that makes a hike the base case | Above 2.6% |
| 10-year yield, level that pressures megacaps again | Above 4.80% |
| 30-year yield, level where the long-bond rout resumes | Above 5.30% |
| Micron, support on any macro sell-off | $950 |
| Bitcoin, invalidation unchanged | Weekly close below $55,000 |
| S&P 500, first support from the record | 7,600 |
My Take
The August rally was built on one bad jobs number that has now been revised to a good one. With core PCE at 3.3%, oil back above $90, and a Chair who has said the Fed has work to do, a September hike is live, and the market has not fully priced it.
Playbook: favor what survives a hike, memory, banks, and short Treasuries; lighten long-duration megacaps and the liquidity trades that led August; treat this week's CPI and Apple's pricing as the two signals that settle the September call.
