Week In Review: The Machine Didn't Move. Everything Inside It Did.

Friday July 24 close to Friday July 31 close.

The Russell 2000 moved three cents in five days. Three cents. Small-cap America spent an entire week doing an impression of a parked car, which is its own kind of statement about who actually benefits when the only thing working is four companies with their own power plants.

Had $1,000 riding the S&P on Friday the 24th? You have $1,010.96 now. Had it in Microsoft instead? $1,217.50. Had it in Micron? $893.68. Same week. Same country. Pick better, apparently.

Wednesday afternoon the Federal Reserve left its target range at 3.50% to 3.75%, which is exactly what everyone expected and absolutely not what anyone talked about afterward. The interesting part was the vote: 9 to 3. All three dissenters, Cleveland's Beth Hammack, Minneapolis' Neel Kashkari and Dallas' Lorie Logan, wanted rates to go up.

That is a rare thing. Dissents usually come from the dovish side, from officials who think the Fed is being too mean and should let the economy breathe. Three people asking for tighter policy in the same meeting is the committee telling you, in the only language it has, that the inflation fight is not filed away.

Chair Kevin Warsh then did something that made it worse. He gutted the forward guidance. The statement came out noticeably shorter, with the section that normally tells markets roughly what to expect next simply not there. Forward guidance is the Fed's practice of telegraphing its next move so nobody panics. Warsh removed the telegraph and left everyone staring at a blank wall, and a market handed a blank wall will fill it in with the scariest thing it can imagine, every single time.

It imagined immediately. The S&P fell 1.54% Wednesday, the Dow dropped 2.18%, and the VIX closed at 20.66 after starting the week at 18.58. We covered the whole ugly afternoon live on Thursday.

Then Logan went on the record Friday and removed any remaining ambiguity: "Without any policy restraint, inflation will likely continue to trend above target until there's an unanticipated shock. Modest action in the near term would reduce the likelihood of needing to take sharper action later." That is central banker for "we should raise now so we don't have to raise a lot later," delivered in the tone of someone reading a weather report.

Big Tech earnings landed all at once and the market did not grade on a curve.

Microsoft finished the week up 21.75%, including a 15.51% single day on Thursday. Azure, its cloud computing business, grew 43% after stripping out currency effects, and management said customer demand is still running ahead of the capacity they can physically build, with the shortage expected to run into fiscal 2027. That is a rare and enviable problem: not "will anyone buy this" but "we cannot pour concrete fast enough."

Amazon went up 17.00% on the week and 15.32% on Friday alone. AWS did $42.23 billion against the $40.57 billion Wall Street wanted, growing 37%, which the company says is its quickest pace in more than four years. Operating profit at AWS rose 64% to $16.6 billion and the margin on it widened to 39.4% from 32.9%. Alphabet tagged along at +11.38%.

Then there is Apple, down 7.24% on the week and 7.35% on Friday. Services revenue came in at $30.74 billion against $31.36 billion expected, Greater China did $18.82 billion against $19.58 billion expected, and the forecast for next quarter landed soft. Meta lost 6.47% after dropping 7.95% on Thursday. Two of the most valuable companies on earth reported into the same week as Microsoft and got the exact opposite reception, which we walked through Friday morning.

Worth noting before anyone frames these earnings: Alphabet's headline profit of $112.1 billion included $77.4 billion of unrealized gain on its SpaceX stake, and Amazon's $62.6 billion included $53.4 billion of non-operating income mostly from marking up its Anthropic investment. Unrealized means nobody sold anything and no cash moved. Those are the two biggest profit numbers of the quarter and the majority of both is a spreadsheet entry about what a private company might be worth.

Seven days ago this newsletter ran a gold-rush story: the AI spenders were getting punished for their spending and the chip suppliers were getting rich selling them the equipment. Miners broke, shovel guys prospered.

Every single name flipped.

Six for six. Not a drift, not a rotation, a clean inversion inside five trading days. The chip unwind started on Monday and memory hit the floor by midweek. Micron went from the best name on the board to the worst, dropping 10.63% after a week where it gained 8.48%, and it managed to do that while rising 18.36% on Thursday. Whatever conviction is supposed to look like, this is the other thing.

The mechanism is not mysterious. Last week the spenders reported enormous capital budgets with no proof of return, so the market bought whoever was cashing those checks. This week the spenders came back with the proof, and the money walked straight back across the street. Amazon then raised its 2026 capital spending plan to roughly $220 billion from about $200 billion and explicitly blamed rising memory prices, which is a fascinating thing to do to Micron's stock in the same breath.

Thursday morning the Bureau of Economic Analysis reported that the economy grew at a 1.5% annual pace in the second quarter, against the 2.1% economists expected and the 2.1% it managed the quarter before. That is a real miss.

The same release carried the GDP price index, which measures what everything in the economy cost. That came in at 6.3% against 3.6% expected. Nearly double. So the economy produced less and charged more for it, which is the arrangement nobody wants and every central banker is paid to prevent.

Now the honest complication, because a number this scary deserves the fine print. The quarterly figures look at April through June. The most recent month tells a gentler story: core PCE, the inflation gauge the Fed actually targets, ran at 3.3% in June, down from 3.4%, and the headline monthly reading was -0.1%. So inflation in the rear-view quarter was ugly and inflation last month was fine. Both are true and the Fed has to make policy standing in between them, which explains a 9 to 3 vote better than any statement could.

Consumers, meanwhile, refuse to be surveyed consistently. The Conference Board's confidence index slipped to 90.8 on Tuesday against 92.4 expected. Then the University of Michigan's sentiment index jumped to 55.2 from 49.5 on Friday. Two surveys, same population, same week, opposite directions. Michigan also showed people expect 4.2% inflation over the next year, down from 4.6%, so Americans are simultaneously more cheerful and still budgeting for prices to rise twice as fast as the Fed's target.

Seven of the eleven S&P sectors fell this week. The index went up anyway, because the four that rose included the one holding Amazon.

Utilities lost 4.19%, last place out of eleven. Real estate lost 1.92%. Materials lost 1.62% and industrials lost 1.54%. These are the parts of the market people buy when they want a dividend and a quiet life, and they got wrecked for a boring reason: they compete with bonds. When a government bond pays 4.74% risk-free, a utility stock paying something similar while carrying actual business risk becomes a much worse offer. Every basis point the 10-year climbs, that trade gets uglier.

$1,000 in utilities on Friday the 24th is $958.09 now. No CEO said anything, no product failed, no scandal broke. The math simply moved.

Oil went the other way and then came back. Crude fell hard early on a pause in US and Iranian strikes, leaving the oil fund down 5.50% on the week, before recovering about 1% Friday. Energy stocks finished essentially flat at -0.12%, which is a remarkably calm result for a sector whose underlying commodity had that kind of week.

Live odds pulled Friday, July 31. Ranked by money actually at risk.

Forty-six million dollars says the Fed does not cut once this year, which as recently as spring was a fringe opinion. The Hormuz pair is the bleak one: the July market settled at under 1%, meaning shipping traffic did not return to normal, and the December market sits at a coin flip. Twelve million dollars of combined volume and the best answer anyone can produce is "maybe."

Gold closed the week at $4,075 an ounce, down about 0.46%, which after this week qualifies as a personality trait. A year ago it was $3,362. That is +21.2% over twelve months.

Here is the part that should bother people. Gold pays no dividend, no interest and no rent. It sits there. Its entire pitch is that it is not a promise from anyone, and it has beaten most portfolios over the past year while the smartest money on earth argued about cloud margins. Speculators trimmed their bets slightly this week, down to 182,100 contracts from 183,900, which is not an exit so much as a shrug. Grandma's coffee can is up 21% and did not have to read a single earnings call transcript.

The 30-year mortgage rate moved up to 6.66% from 6.58%. Eight basis points does not sound like much until you translate it: on a $350,000 loan that is roughly $19 more a month, or about $6,800 over the life of the thing, for the crime of waiting a week.

People responded accordingly. Mortgage applications fell 6.4%. Refinancing fell 9.9%. Purchase applications fell 3.6%. Home prices, for their part, are still creeping up, with the Case-Shiller index showing 1.6% annual growth in May.

So: prices rising slowly, borrowing costs rising faster, and a Fed that three members want to make more expensive. The real estate sector fell 1.92% this week and it was not being dramatic. If September brings a hike, every number in this section gets worse before it gets better, and the "wait for rates to come down" strategy that millions of people have been running since 2022 turns four years old with nothing to show for it.

Retail and the professionals spent the week disagreeing loudly.

News volume on Amazon spiked to 141 articles Friday, the loudest name on the board, and sentiment on it stayed strong all week. Apple ran 109 articles Friday with the weakest sentiment reading of any megacap we tracked at 0.575. Micron's sentiment bottomed at 0.297 on Tuesday, which is close to the floor, and then climbed back to 0.735 by Friday even as the stock finished the week down 10.63%. The mood recovered. The price did not.

Meanwhile the professionals are positioned in the other direction. Futures traders held a net short position of 17,200 contracts on the S&P 500 and got slightly more short, while the fear index collapsed 13.94% to 15.99. So the crowd is calm, the pros are hedged, and both are reading identical headlines.

The genuinely strange one: speculators piled into crude oil, lifting net long positions 47% to 120,100 contracts, during a week the oil fund fell 5.50%. Somebody bought that dip with real enthusiasm and real money, and OPEC+ met on Sunday.

The amuse-bouche was a durable goods report on Monday that came in at 0.3% against 2.5% expected, served cold and sent back barely touched.

The appetizer was Wednesday's Fed decision: a dish everyone had already read the description of, which arrived with three of the nine chefs publicly refusing to sign off on the recipe. The dining room lost its composure. Someone knocked 2.18% off the Dow reaching for the bread.

The main course was Big Tech, and the kitchen sent out two completely different plates. Microsoft, Amazon and Alphabet arrived hot, plated beautifully, and the room applauded. Apple and Meta arrived at the same table, from the same kitchen, and got scraped into the bin.

Dessert was Friday's consumer sentiment print, unexpectedly sweet, and by then nobody had room.

The bill was the 10-year Treasury at 4.74%, its highest in about eighteen months, presented quietly at the end the way expensive things always are.

Your washing machine finishes exactly where it started. It does not travel. It sits in the same six square feet of your house for a decade, and if you judged it purely on displacement you would conclude that nothing ever happens in there.

The S&P 500 gained 1.10% this week and ended at 7,489.72. That is the machine sitting still. Inside the drum, Microsoft gained 21.75%, Micron lost 10.63%, three Fed officials voted to raise interest rates in a meeting where rates did not change, the ten-year yield hit an eighteen-month high, and every AI position from the previous week ran exactly backwards. A 32-point spread between the best and worst megacap. A ten-point spread between the best and worst sector. Seven of eleven sectors red on a green week.

Nothing about "up 1.1%" tells you any of that, and this matters more than it sounds. Most people check their portfolio, see a small green number, and conclude the week was uneventful. This week the small green number was the least informative fact available. What actually happened is that the market stopped rewarding AI ambition and started demanding AI revenue, and repriced roughly a trillion dollars of market value across five days sorting out who has which.

The one question that decides next week arrives Friday at 7:30 AM CT, when the July jobs report lands. Forecasters expect 91,000 new jobs against June's 57,000, and unemployment to tick up to 4.3% from 4.2%. If hiring comes in hot, the September hike stops being a 60% probability and starts being a scheduling detail. If it comes in cold, the Fed is stuck holding an inflation problem and a jobs problem at the same time, with three members already voting to make one of them worse.

Either way, check the drum, not the machine. The machine never moves. It is not supposed to.

Markets don't wait, and neither do we. Follow Tracking the Trade for fast takes, the numbers that matter, and the occasional meme that hits a little too close to your portfolio.

Everything above is provided for informational, educational and mildly therapeutic purposes only. It is not investment advice, financial advice, tax advice, or a signed affidavit regarding the behavior of the Federal Reserve, which has demonstrated this week that it cannot even get nine people to agree with each other. We are not your financial advisor, your fiduciary, or the reason your utilities position had a rough Tuesday. Past performance does not predict future results, and neither does a washing machine metaphor, however satisfying. Do your own research, consult a licensed professional before moving real money, and please do not make portfolio decisions based on a newsletter that opened with laundry.

Your customers are searching in places your strategy doesn’t reach.

So before your business is buried and left behind, you need to understand the new rules of SEO.

BELAY's SEO in the Age of AI report explains how search is changing, what AI means for your visibility, and the practical steps small businesses like yours can take to stay visible.

BELAY’s U.S.-based Marketing Assistants turn strategy into execution, helping your business stay visible, credible, and competitive in every search.

Download a copy for free