1The September hike needed a hot CPI to survive — and got an in-line one
July CPI printed at +0.07% month on month and 3.30% year on year, with core at +0.22% and 2.47% — in line, and a touch cooler than June on both. That mattered more than it reads, because an upside surprise was the only clean trigger the September-hike case had left, and it did not arrive. Polymarket's September Fed contract — the deepest on the board at $29.8m of volume — repriced hold up eight points to 66.5% in a single session, and the 25bp-hike leg down seven to 33.5%. The move matched exactly what Reuters read as "dollar slips as in-line CPI tempers rate-hike bets." Note the regime it is happening in: the live debate is still hold-versus-hike, not hold-versus-cut. A cut sits near 2%.
The caveat worth keeping is that the previous day's core-CPI contract had resolved on the "benign, not soft" outcome — core landed in the 0.2% bucket the market had leaned toward. So the print removed the hike trigger without handing the doves anything either. It bought time, not a direction.
2Nvidia's $500bn financing loop is the new bear thesis — and it explains a gold-miner pitch and a metals-rotation pitch on the same day
The most-discussed single item was Nvidia's agreement, first reported by the FT, with six asset managers — Apollo, Blackstone, BlackRock, Brookfield, Goldman and KKR — to raise $500bn that then gets lent to Nvidia's own customers to buy Nvidia's own chips. fell about 4% on its own announcement. Jay Goldberg, on Prof G Markets, read it as a tell rather than a triumph: GPUs have never been accepted as loan collateral, so neo-cloud debt has been backed by hyperscaler offtake contracts — and now that hyperscalers are pulling that guarantee to build their own capacity, Nvidia has to step in as the backstop. He is careful that these are memoranda of understanding, not finalised deals, and that the precedent — the OpenAI MOU — changed materially six months after its press release.
That same structure is what Felix, on Felix & Friends, uses to justify a gold-and-silver-miner trade: his argument is that the trigger for metals is not inflation or the Fed but the AI financing plumbing, with the cost of insuring Nvidia's debt having doubled since late May and the hyperscaler buyback bid flipping from a roughly $190bn-a-year buyer to a net $147bn seller. His way in is , and he wants a dollar of upside confirmation before buying; his non-gold buy is , bought heavily this week. He is explicitly avoiding and — "already own plenty of that through everything else."
Ross Givens arrives at the same rotation from the chart side, with money leaving a four-year-old AI trade — companies "borrowing at high interest rates" without returns on capital — and moving into natural resources, where gold, silver, copper, steel, aluminium and coal are all running. His trades come with mechanics attached: (Peabody Energy) on a buy-stop at $24.75 above the pivot, stop $22.85; (Century Aluminum) bought around $50.27 with a $47.50 stop on a breakout retest; and (Ramaco Resources) at $10.25, which he flags as the volatile one and not one of his two official trades. On the obvious names he is patient — miners like are already extended, and he wants a pullback rather than a chase.
Two things to hold onto. Goldberg and Farrar were the only voices on the Nvidia segment and no bull was invited, so it is a one-sided panel. And Felix's two load-bearing numbers — the buyback flip and "cheapest miners in 50 years" — are single-source and unsourced on screen; the Nvidia CDS doubling is the one piece Prof G's segment independently corroborates.
3The space complex de-rated 40% after SpaceX listed, and the earnings didn't win it back
On the same Prof G episode, Tim Farrar of TMF Associates called the whole space complex overvalued on current revenue and profitability — a sector that de-rated about 40% once SpaceX IPO'd in June and never got the money back. (AST SpaceMobile) missed on revenue with losses doubling, including a write-off for a satellite Blue Origin put in the wrong orbit; (Rocket Lab) posted record revenue up 62% but a wider-than-expected loss, with its Neutron rocket slipping from 2025 toward 2027. Asked whether they are all overvalued, Farrar's answer was "I think that's right." He rates the better-positioned of the two — buying Iridium swaps a high-flying stock for a real cash business, "a bit like AOL buying Time Warner" — but he does not exempt it from the call. Both rows are the same single answer, not two independent ones.
What the sources recommended
Nine single-name calls came out of three videos — five buys and four avoids — and several of the buys are conditional rather than market orders.
The buys. Ross Givens named three metals trades with mechanics attached: (Peabody Energy) on a buy-stop at $24.75 above the pivot, (Century Aluminum) bought around $50.27 on a breakout retest, and (Ramaco Resources) at $10.25 — which he flags as the volatile one and explicitly not one of his two official trades. Felix added as the way into gold and silver miners, but a wait-for-confirmation call rather than a market order, and , the international quality-factor fund he bought heavily this week.
The avoids. Felix is explicitly steering clear of and — "already own plenty through everything else." And on Prof G, Tim Farrar rated the space complex overvalued: (AST SpaceMobile) and (Rocket Lab), with the better-positioned of the two but not exempt from the call.
The reads that test all this land within a day
reports after the close tonight and (Applied Materials) after the close tomorrow, alongside July PPI — the wholesale-inflation print that sits underneath the CPI that just moved the Fed odds. is the cleaner of the two, since it reads directly on AI-capex order flow, which is the demand side of the financing loop in the second section.
