Is the risk worth the reward at this point? Part 2

The question I keep getting is not about a level. It’s about whether to be in this market at all. Stay invested while indexes print records, or take a guaranteed yield in short-term paper and sit this out until it works itself out. It sounds like the conservative choice, and it is — but conservative is not the same as costless. Every decision in this tape charges you something, including the decision to do nothing.

Good news was bad news, and that is the whole story.

The tape got a jobs number three times better than anyone asked for, and sold it. When a market can’t rally on its own good news, the question stops being what the economy is doing and starts being what the discount rate is doing.

Friday’s August payroll report printed 162,000 against a Dow Jones consensus near 53,000, with July revised up to +21,000 from −23,000. Unemployment held at 4.1% — 4.14% unrounded — average hourly earnings rose 3.1% year over year, the slowest since May 2021, and the labor force grew by 683,000 as participation ticked up to 61.6%, its first increase in nine months. On paper that is close to the best report of the cycle. The S&P 500 closed down 0.38% at 7,718.60, the Dow gave back 272 points, and the two-year finished at 4.374%, a fifty-two-week high. That reaction is the most important thing that happened all week, and it is worth sitting with before a single level gets drawn.

Wharton’s Jeremy Siegel read the report as the bulls’ version: “a supply response, not too much demand… wages completely under control,” and then said the quiet part — “if the midterm elections were not on the horizon, the Fed would raise rates.” Goldman’s Jan Hatzius took the other side of the same data: the print “does remove an obstacle to a hike,” but “it doesn’t provide an affirmative case for why the funds rate needs to go up again,” with unit labor costs growing below 2% and the funds rate already “in the zone of neutrality.” Ed Yardeni called it a Goldilocks report — U-6 down to 7.7%, hiring breadth the best since early 2024 — and concluded the Fed is now free to focus entirely on inflation. Fundstrat’s Mark Newton called it a double-edged sword, which is exactly how it traded.

Start with the bull column, because it earned its space. The AI order book is no longer a story, it is a backlog. Dell posted record revenue of $47 billion, up 58%, with $60.9 billion of AI server orders in a single quarter, a $95 billion backlog and a full-year guide raised by $25 billion — the stock added 15.8%. Broadcom grew AI semiconductor revenue 221% year over year to $16.7 billion and laid out a roadmap running to roughly $115 billion in fiscal 2027 across six custom-silicon customers including Google, Anthropic and OpenAI. Snowflake raised its full-year product guide to $6.07 billion and gained 16.6%. Nvidia agreed to buy Hugging Face for about $12.9 billion. Citi’s Scott Chronert kept his 8,100 year-end target and called the remaining upside “about a percent a month… certainly doable,” arguing the market would look through a preemptive September hike. Renaissance Macro’s Jeff deGraaf led his note with “stay long” — trends intact, and critically, “credit’s intact, even though bond yields are pushing higher, we’re not seeing it in the public corporate credit markets.” Tom Lee stayed contrarian at 8,000, calling it “Sup-tember” and saying the bottom “might even be today” on Tuesday. It was, by 116 points.

Now the other column, and again it is arithmetic, not valuation. The market has stopped paying for good quarters. Broadcom, with that 221% AI number, fell 2.7% because total Q4 revenue guidance came in at $34.8 billion against $35.0 billion expected — $200 million light. Zscaler beat and raised for fiscal 2027 and traded down 2.1%. MongoDB lost 13.5%, Palo Alto 9.3% after naming memory costs as a gross-margin headwind, Credo 20%, Guidewire 14%. Lululemon cut its full year from $11–11.15 billion to $10.35–10.5 billion and lost a fifth of its value in a morning. The rule all week: an upside surprise to the guide gets bought, an in-line guide gets sold, a decelerating guide gets punished. Beneath the surface it was worse than the index — industrials fell more than 4% across five sessions with four gainers in eighty-three names, new 52-week lows outnumbered highs midweek, and the ten-day advance-decline line hit its most negative reading since May before Thursday’s bounce repaired the optics. Schwab’s Kevin Gordon put numbers on the leadership problem: Nvidia’s contribution rank in the S&P has slipped to third after holding first place for years, Meta and Tesla are anchoring the bottom of the Mag 7, and “not necessarily thinking about it as a monolith anymore is going to be really key.” Chronert framed the same quarter as “a software surge and a semiconductor sell-off” inside a mega-cap growth cohort that is 35–50% of the index. deGraaf’s momentum work says the acute phase is over and the grind isn’t: the 65-day tactical reversion in semis reached the 2nd percentile of forty years, but the longer-term reversion is only around the 88th percentile — “the grind is still in front of us… call it six months” — with frustrated momentum holders rotating into healthcare. His level is 7,300 inside a consolidation, then a decent fourth quarter. Newton independently carries 7,300; BofA’s Savita Subramanian is Street-low at 7,100.

The rates leg is where the week was actually decided, and it was not decided in Washington. The ten-year touched 4.814% on Wednesday, its highest since January 2025; German yields hit 2011 levels, gilts 2007 levels, and Japanese ten-years crossed 3% for the first time since 1996. Nobody is discussing a cut. The entire argument is hike-or-hold, and the probability ran a full round trip inside four sessions: roughly a third a week earlier, two-in-three by Wednesday, a coin flip Thursday afternoon after Governor Christopher Waller said he would be “inclined to support holding” if the next two weeks of data cooperate, and back toward 60% after Friday’s payrolls. Hatzius framed the decision as “counting heads” — three dissenters who already voted to hike, against Waller and New York’s John Williams leaning hold — and named his threshold: two-tenths on core PCE is consistent with a hold. Warren Pies of 3Fourteen named the same line from the other direction: “0.25 is kind of the line in the sand… anything above 0.2 after this report, as long as oil stays up here around 90, we’re probably going to get a hike.” Gordon agreed a hotter CPI and PPI “probably locks in a rate hike for September,” and said outright that a September hike is “absolutely” possible. Both prints land inside the Fed’s blackout.

There is a real fight underneath that about what a hike does to the long end, and it matters more to equities than the funds rate does. Siegel argues a credible move “might actually lower long term rates.” Pies took the other side explicitly: “Volcker thought the same when he started hiking in 1979. He later called himself foolish,” and points to bear flattening every time hike odds rise. Chronert sits in the middle and is honest that his target depends on it — “all eyes from our team’s perspective is really on where the ten year and out the curve is going.” The buyer side is thinning independently: Norway’s $2.3 trillion sovereign fund has proposed cutting government bonds from 70% to 50% of its bond book, roughly $75–80 billion of Treasuries phased out, joining China, Japan and the Gulf on the list of historically reliable buyers with less reason to buy.

The tell nobody wanted to talk about was gold. Three sessions of open escalation in the Gulf, crude through $90 for the first time since July on a 5.2% day, and gold fell on every one of them into a two-week low before recovering to $4,513.20 by Friday. Gold traded its rates beta, not its geopolitics beta. That is what a real-yield shock looks like. WTI settled at $90.55 and Brent at $95.04, capping a week of roughly +10% and +7.6%; Morgan Stanley’s Martijn Rats moved the house Brent call to $100. El-Erian flagged diesel at a record $5.85 a gallon. Sierra Alpha’s David Keller had crude breaking out of a multi-month coil with $92 and $98 as objectives, and rotation running into energy, utilities and healthcare. On the AI side, Pies argues the risk isn’t fundamental — “we see rental rates staying strong” — it’s political: “the semi group, which is the beating heart of the AI trade, has started to move with the Republican odds of holding the Senate.”

Two structural items belong on the calendar rather than in an argument. Roughly $6.2 trillion of index option exposure expires on the 18th, the same morning the Bank of Japan meets — two days after the FOMC — with a Japanese hike close to fully priced and the yen already having its biggest week of the summer. The last time a BoJ hike met a crowded carry position, in August 2024, the S&P lost 6% in three sessions and the VIX traded above 60. That is not a forecast, it is a date with a history. The other is the corporate bid: Citadel Securities’ Scott Rubner has buyback authorizations going quiet around the 12th, with retail’s September dip-buying historically running at half its normal rate — his advice all week was to use strength to trim and add inexpensive protection into a tactical reset, with a constructive turn back around mid-October.

So where does that leave the tape. The index put in a one-month low Tuesday at 7,611, reclaimed all of it by Thursday, and finished the week roughly where it started, still without a 1% down day in more than twenty-five sessions — the VIX has now held a 14-to-17 range for twenty-five straight days, the longest such streak since May 1992. Protection ended the week about the cheapest since 2024. Keller notes AAII bulls at 40% with only two points separating bulls from bears, and the survey crowd flipped net bullish for the first time in five weeks in the same week the index printed its low. deGraaf sees ETF flows and futures positioning showing a crowd “pretty gunned up”; Lee sees shrinking margin debt and negative sentiment and says “we don’t have excess positioning.” They are looking at the same market and reaching opposite conclusions, which is usually what a range looks like from the inside.

I am biased toward the tape. I follow what the market does, not what I think it should do. Right now the tape is a percent off its highs with volatility at a year-to-date low, the front end pricing a hike, and breadth thinning underneath — and I will keep watching the ten-year, the credit spreads and the equal-weight index to tell me when that changes.

So the tape hands us a divided picture, not a verdict.

On one side: the best labor print of the cycle — 162,000 against a 53,000 consensus, July revised up to +21,000, unemployment at 4.1%, wages at 3.1% and the slowest since May 2021, with participation rising for the first time in nine months. The AI order book turned into a backlog: Dell’s record $47 billion quarter with $60.9 billion of AI server orders and a $95 billion backlog, Broadcom’s 221% AI semiconductor growth toward a roughly $115 billion fiscal-2027 roadmap, Snowflake raising to $6.07 billion. deGraaf’s note led with “stay long” and, more importantly, “credit’s intact.” Chronert is still at 8,100, Lee at 8,000. On the charts the levels are doing their job: SPX held the 7,626 prior high, NASDAQ and the Dow both found the 50 DMA, the Russell held its 100 DMA, crude’s $67 line is still the low, and the Fib numbers are working cleanly across IGV, silver, SOX and Bitcoin.

On the other: a market that has stopped paying for good quarters. Broadcom fell 2.7% on a $200 million light guide, Zscaler beat-and-raised and traded down, MongoDB lost 13.5%, Palo Alto 9.3%, Credo 20%, Guidewire 14%, Lululemon a fifth of its value in a morning. Industrials fell more than 4% with four gainers in eighty-three names, new lows outnumbered new highs midweek, and the ten-day advance-decline line hit its most negative reading since May. The ten-year touched 4.814%, the two-year finished at a fifty-two-week high of 4.374%, and the entire argument is hike-or-hold with CPI and PPI landing inside the blackout. Gold fell through three sessions of Gulf escalation and crude above $90 — a real-yield shock, not a geopolitics bid. Roughly $6.2 trillion of index option exposure expires on the 18th, the same morning the BoJ meets, two days after the FOMC. And gold gave back the 200 DMA, ORCL’s Fib levels at $132.14 and $121.76 did not hold with earnings Thursday, and the grains are stretched far above the 20 DMA.

I’m biased toward the tape. I follow what the market does, not what I think it should do — and right now the tape is a percent off its highs with volatility at a year-to-date low, the front end pricing a hike, and breadth thinning underneath. I’ll keep watching the ten-year, the credit spreads and the equal-weight index to tell me when that changes.