1The AI build-out's stress showed up in credit before it showed up in earnings
Torsten Slok, the Apollo Global Management chief economist, on Prof G Markets, gave the mechanism plainly. The hyperscalers shifted the financing of their capital spending from equity to debt; that surge of investment-grade supply widened credit spreads; and the cost of insuring that debt against default has now hit record highs on , , and, most recently, . The single hardest number underneath it is that Google printed its first-ever negative free cash flow.
The equity market is reading the same accounts and arriving second. On The Compound's Animal Spirits, both hosts pointed at sitting in an 18% drawdown after a quarter with cloud revenue up 82% year on year — the drawdown is not a demand event, it is capex going up again and free cash flow going negative. Their estimate of where spending would actually capitulate is around a 40% drawdown, which is a long way from here. So the discipline credit is asking for is nowhere near being imposed.
Slok's own conclusion follows from that: tonight's and prints and Thursday's hyperscaler numbers matter more than the Fed statement, because they are what resolves the capex-versus-return question credit is already pricing. Hold one caveat against his prescription — Slok's employer is a large private-credit manager, so "diversify into private equity and private credit" is not a disinterested destination. The spread and default-swap data stands on its own; where he sends you does not.
The argument has moved. It is no longer whether AI spending earns a return, it is who is funding it while the question stays open.
2The memory complex fell apart without taking the index with it
The Nasdaq 100 fell a fifth straight session and briefly entered correction at −10%, with the PHLX semiconductor index down 6% and down 9%. It began overnight in Asia — SK Hynix off 15%, the KOSPI off 11%. It matters more than a routine chip selloff because 9 of the 12 biggest contributors to the S&P 500's first-half return were semiconductor stocks, which is what makes "where does this market go without chips" a real question rather than a rhetorical one.
Measured from their highs the memory names are far worse than the index suggests: −27%, −34%, −46%, −26%, −41%. And the S&P 500 still sat roughly 2% below its all-time high. Both Animal Spirits hosts read that gap as the leadership "de-bubbling" without the market breaking — good news rather than bad. Their supporting detail is a Monday session where semis fell 4% while software rose 3.3%, an outlier pairing on a scatter going back to 1989, and the round trip in the momentum factor: went from −10% year-to-date in April, to +40%, and back to +20%.
Valuation has not become the objection. The forward multiple is 19.5× against 18× at the Liberation-Day lows despite a much higher index, because the S&P 500's earnings-guidance-momentum score is the highest in the series — the earnings are moving faster than the price. So the de-rating so far is concentrated in one complex. Whether that is rotation or the first leg of something wider is a question only the index can answer, and the index has not answered it.
3Equity supply is coming back, and that ends the bid that carried US stocks for a decade
Rupert Mitchell of Blind Squirrel Macro, on Excess Returns, described his core position as long , the equal-weight S&P 500, against short — on since early June, up 10% since the third week of June, and explicitly framed as a 12-to-24-month trade rather than a tactical one. What makes it interesting is that the thesis is supply, not sentiment. The SpaceX listing absorbed $86bn in a single June day, Google raised $85bn, and mega-caps issuing shares reverses the buyback-driven de-equitisation that has bid US equities for ten to fifteen years. The valuation leg is that more than 51% of the Nasdaq 100 trades above ten times sales.
Ross Givens reaches the breadth half of that conclusion from the tape instead of the flow. looks flat since May only because it is cap-weighted: has kept making progress, net new highs have stayed positive, the share of stocks above the 200-day average has not deteriorated, and the advance-decline line is at highs. Slok gets there from a third seat, arguing for value over growth and S&P 400 midcaps on the grounds that "growth is crashing, value is skyrocketing" on his factor models. Three of the day's six videos, then, pointing the same way from flow, tape and factor data.
One caution belongs with it. Mitchell runs the portfolio he is describing, so every position named is a book he is talking, and he is not selling the trade as riskless — record-low implied correlation plus equity leverage has him adding volatility hedges into the summer and preferring assets that are already cheap. If the supply argument is right, the gap between equal-weight and cap-weight is a multi-year adjustment rather than a few weeks of rotation.
4Into the July decision, the priced tail was a hike, not a cut
Polymarket's July Fed contract — $135.6m lifetime volume and $22.3m of it in the last 24 hours, the deepest thing on the board — priced no change at 75.4%, down 3.7 points on the day, against a 25bp increase at 24.3%, up 3.9. Cuts of any size sat near 0.3%. So the tail widened into decision day, on the hike side, and the separate 2026 contract prices a hike landing at some point this year at 76.5%. That caps how dovish a hold can be read.
The data underneath is genuinely two-sided. The headline CPI index actually fell 0.42% month-on-month in June, to 332.568 from 333.979 — a disinflationary print that predates this week's oil move entirely. But core PCE rose 0.32% in May, roughly 3.9% annualised, and that is the series the Fed weights. Unemployment ticked down to 4.2% from 4.3%, which removes the dovish cover a softening labour print would have given. The 10-year eased six basis points into the meeting to 4.65%.
Justin Wolfers, of the University of Michigan, on the same Prof G Markets episode, supplied the textbook frame: both the tariff regime and the Iran war are supply shocks, and a central bank can look through supply shocks. The tariff piece is not small — the new Section 301 regime took effect Friday across 80 countries covering more than 99% of US imports at 10% and 12% rates, taking the average applied rate to 11.1% and, on Yale Budget Lab numbers, to 11.8% by the end of 2026. Wolfers flagged the roughly 70/30 pre-decision odds as unusually unsettled and declined to out-forecast the market, which is its own signal about how much this one is worth guessing.
5The Gulf escalation bought crude a higher floor, not a spike
The tape re-escalated overnight. US and Saudi forces struck Iran-backed groups in Iraq, Iran fired ballistic missiles at US forces, and Iran publicly ruled out regional management of the Strait of Hormuz. Crude jumped about 7%, on Reuters' figure. The prediction market moved further on this than on anything else: a US-Iran effective ceasefire by 31 July collapsed 37 points in 24 hours to 22.5%, by 14 August to 41.5%, and by 31 August to 52.0% — $1.69m of volume in a day against $6.6m lifetime. A day earlier those same contracts had been pricing that tail away.
The shape of the oil repricing is the part worth keeping. On the August WTI contract, ≥$90 rose 14 points to 73.5% and ≥$100 rose 13 points to 46%, while ≥$110 fell 13 points to 17%. The near strikes bought a floor; the far strikes were sold into the same headlines. That contract is thin — $61k lifetime, $38k in a day — so the exact levels are weak evidence and only the direction is solid.
That shape is precisely what Mitchell described independently as the "China collar": China has become the price-maker through storage, effectively writing a put under crude in the high $60s while capping the right tail. His conclusion is that a collared oil price makes energy equities more investable rather than less, because capital spending becomes underwritable — he likes global super-majors, refiners (trimmed slightly this week), offshore services and Canadian producers with long reserve lives, and says outright that he does not need triple-digit crude for the trade to work. Hormuz traffic returning to normal by 31 December still prices as a coin flip at 53.5%, and it drifted down three points on the day. With Kalshi unreachable for a second consecutive day, this pricing comes from a single venue.
What the sources recommended
Ten calls came out of six videos, and they are less independent than that count suggests: three of the buys are one cybersecurity basket from a single eight-minute technical video, and the two sells are one leg of a pair trade plus one brand judgement.
The buys. (Stryker) is Ross Givens' highest-conviction name, and the peg is not Stryker at all — Johnson & Johnson's Ottava robot won an FDA De Novo authorisation, creating a brand-new soft-tissue robotics device class, which he reads as proof the category is approvable rather than as a trade. His argument for Stryker is that it sits in bone rather than soft tissue and so is not in the // fight at all, with 3,000-plus installed Mako systems locking in the implant revenue behind them; the discount comes from a Q1 cyberattack that halted production for three weeks and cost about $375m of revenue without management cutting full-year guidance. The caveat is his own timing — the thesis rests on an earnings print landing days after he published it.
(Rapid7), (Check Point) and (CrowdStrike) are the cybersecurity basket, all pullback entries on a group he says is up around 50% in six weeks, using as the strength read rather than as a call. These are technical entries with no fundamentals presented, and in particular is a $9 stock down from $60 that had just run 114% in a fortnight. Both Givens videos also push a paid membership service twice, with a QR code.
against is Mitchell's pair from the third section. (Corning) comes from Josh Brown on Ask the Compound, who names it as a stock he has a buy on and wants to be long after his firm's rules-based momentum sleeve sold it — that episode carries sustained promotion of the firm's own momentum product, an infomercial risk the hosts flag themselves. comes from Animal Spirits, at a roughly $2.5bn market cap with "a long runway" and reported takeover interest, on the back of Nolan's The Odyssey taking $140m to date, 73% ahead of Oppenheimer at the same point; the hosts say explicitly that they are not making calls, and the takeover line is press reporting rather than their own work.
Two ideas carry no ticker here, because neither symbol could be confirmed: Mitchell's LSE-listed depositary receipts in the Templeton-managed Uzbekistan privatisation fund, which listed in May and is up 30% since, and SpaceX, which was discussed at length in three of the six videos — a new low near $107 a share, down 52% from its high and lower in 14 of 18 July sessions.
The sells. is the short leg of Mitchell's pair: more than half the Nasdaq 100 above ten times sales, into an issuance wave. is Animal Spirits, where both hosts decline to buy the drawdown, and the stated reason is that the brand has lost its hold on Gen Z rather than that the multiple is wrong — a structural objection, not a valuation one.
The capex question gets answered after the close, not at two o'clock
The Fed statement lands around 14:00 ET with a press conference from Chair Warsh after it, priced at 75.4% for a hold. But and report after the close, follows on Thursday 30 July in Tim Cook's final earnings call as chief executive, and 's third quarter lands with AI's smartphone and datacentre pull as the focus. On Slok's framing those prints, not the statement, are what settles the credit story in the first section — and they are the reason a 2pm hold could be the less consequential event of the day.
Two of the day's own calls also get marked quickly. reports its second quarter this week, which tests Givens' thesis within days of him publishing it, and reports on 5 August, which Brown named as the specific trigger for putting fresh money into a position he already owns. Elsewhere in the print flow, fell 14% on a revenue miss, beat and rallied pre-market, and topped estimates while holding guidance.
On the geopolitical side the watch list is escalation plumbing rather than headlines: new US sanctions on tankers, the Houthis weighing transit fees for the Red Sea, and a drone strike on a gas storage tanker at an Egyptian Mediterranean port. None of those is a production outage, which fits a market that bought the floor and sold the far strikes.
