1A collapsed ceasefire has put a 2026 Fed rate hike back to roughly even money
The US–Iran ceasefire broke 22 days after the memorandum was signed. Iran struck three tankers in the Strait of Hormuz; the United States hit more than 80 targets and revoked the waiver on Iranian oil sanctions; Iran answered against US bases and Jordan's Azraq base. Brent rose 7% to about $80, a two-week high, and the 10-year Treasury yield moved with it — 4.48% on 6 July, rising on inflation fears rather than on growth. No payrolls or CPI printed inside the window, so the move is running on expectations rather than on data: the last employment reading is May's, with unemployment at 4.3% and payrolls up 129,000.
The rate contracts have taken the oil move literally. Polymarket prices a 2026 Fed rate hike at 52.5% on about $3.6m of volume, close to the roughly 56% on Kalshi's hike-before-year-end market cited on Prof G Markets — a coin flip on tightening, in a year that began with hikes priced near zero. The deeper book says the same thing from the other side: "how many Fed rate cuts in 2026" prices zero cuts at 78.6% on about $41m, the heaviest contract in the day's set. The expectations data points the same way. The New York Fed's one-year inflation expectation is 3.7%, the highest in years, the 2s10s spread is mildly steepening at +0.36 as of 7 July, and the IMF has cut 2026 global output to 3% while projecting global inflation of 4.7%. Housing is already carrying strain, with June sales disappointing even as prices printed at record highs.
What the oil contracts do not say is runaway. Polymarket gives WTI a roughly one-in-three chance (32.5%) of tagging $80 during July and only 3.4% of reaching $100, on about $2.2m — elevated and sticky rather than spiralling. The chokepoint, though, is priced to stay shut. Hormuz traffic returning to normal by 15 July sits at 0.45% on about $8.4m, rising only to 16.5% by 31 August and 62.5% by year-end: eventually, not soon. A US–Iran nuclear deal by 31 July prices at 1.45% on about $9m, and a US invasion of Iran before 2027 at 15.5% on about $40m. For the record, President Trump posted "OIL PRICES COMING DOWN!" on 6 July, citing a Saudi price cut; the 8–9 July move went the other way inside two days.
Brian Katulis of the Middle East Institute, speaking on Prof G Markets, argues there is no articulated political end-state on either side, expects a prolonged "not-war-not-peace" limbo, treats Hormuz as the lynchpin for years, and says the geopolitical risk is not priced in. The contracts are the counter-argument, and they are not a straight refusal: they price the chokepoint staying shut and the deal not happening, exactly as he describes, while pricing crude capped well short of $100. So the disagreement is narrower than it looks — it is about the size of the shock, not its duration. The WTI strike ladder is where that gets settled, and a $100 print would say Katulis was reading it correctly.
2The June AI selloff is being framed as a dip, and one memory quarter is doing the work
June took the chip index down 10%, nuclear names down 30% and quantum down 50%. Brian, on Business with Brian, argues that was a dip rather than a top, and the evidence he leans on for that reframing is a single quarter: Micron's revenue up 346% year on year at roughly 85% gross margin, AI memory sold out into 2026, and a shortage running into 2028. The mechanism matters more than the growth rate. If the memory that feeds the accelerators is spoken for two years forward, the buildout is accelerating rather than digesting — which is the opposite of what a 10% drawdown in chips implies. rebounded about 5% pre-market on the print.
The names he scales into it are chosen to sit at different points on that supply chain rather than to diversify away from it: for data-centre electrical, on data-centre orders up 240%; as, in his description, the only profitable nuclear name, carrying Navy reactors, small modular reactors and isotopes against an $8.6bn backlog; on the CUDA moat at a PEG near 0.5; on custom AI silicon and networking, also at a PEG near 0.5 on next year's earnings; and as the cheapest mega-cap at about 24x, on TPUs, Cloud and Waymo, and the most oversold of the five.
The disagreement sits on the very name the case is built from. Michael Burry is doubling down on his Micron short on an AI-chip-bubble thesis, reported by TheStreet, at the same time as the print and the rebound. That is why is undirected here rather than a call: the day's material contains both the blowout and the short, and what separates them is a single testable claim — whether the shortage really runs to 2028. If it shortens, the dip-buying case loses its anchor and Burry is early rather than wrong.
3Humanoid robotics is being pitched as AI's second phase, on a build cost that just collapsed
Ross Givens builds a robotics case on three numbers: $16bn invested into robotics in Q1 2026, Morgan Stanley's projection of a billion humanoids by 2050 into a $5trn market, and Nvidia's open-source Isaac Groot operating system taking the cost of building one from about $250,000 to $30,000–100,000. The last of those is the actual mechanism — a shared software layer is what turns a bespoke research machine into a manufacturable product, and a cost collapse of that size changes who can place an order. Tesla converting a Fremont line to Optimus is the same claim expressed in factory floor space.
Three names carry it: (Allient — robot motors, gears and motion control, about $1.5bn of market cap, under-covered, profitable, with a roughly 7% daily range), (Hesai — LiDAR as "robot eyes", a record 471,000 units in the quarter, revenue up 30%, a Mercedes-Benz deal and an 8-for-1 split effective 10 July, with Chinese-ADR risk flagged alongside the call) and (the KraneShares humanoid-robotics ETF, up 29% year to date, pitched as the whole-supply-chain option).
Two caveats travel with this one and neither is optional. The channel carrying it also pitched a $5 paid trading service twice with a QR code, and framed the names with "some of these will go 20x" and "the biggest fortunes of the next 5-10 years" — the companies are real, the framing is promotional, and those are separate facts. And the auto-caption garbled two of the three symbols, rendering Hesai as "HSI" and the KraneShares ETF as "KID"; both were resolved by matching the spoken descriptions, so the symbols themselves warrant an independent check before anyone relies on them. The theme has a genuine cost curve underneath it and the day's only source for it is the day's only source flagged for promotion. Both are true at once, and the second does not cancel the first.
4Four US states are suing Meta for more than 90% of its market value, and its insurer no longer has to pay
The Meta segment on Prof G Markets puts the claim at $1.4trn across four state suits — more than 90% of 's market capitalisation — and frames it as a "big-tobacco moment". The detail doing the most work is the insurance one: the insurer has been relieved of coverage on the grounds that the behaviour was ruled intentional, which moves any eventual settlement off a covered liability and onto Meta's own balance sheet. is down 8% year to date, and the segment's reading is that settlements are not yet in the price. The headline number is a filing rather than a judgment, so it is not the figure to watch; the insurance ruling is, because it decides who writes whatever cheque is eventually written.
What the sources recommended
Eleven buys, out of two of the day's three videos, and they are two theses rather than eleven separate ideas. The two bearish-leaning names in the day's material — Meta's legal overhang and Michael Burry's Micron short — are contested reads rather than sell calls, and neither is counted as one here.
The AI-buildout basket, from Business with Brian, is a single "buy the dip" call scaled deliberately from mega-cap down. , , , and are the high-conviction core, described above, with the 200-day moving average named as the entry level on rather than the current price. Three higher-beta names carry conditions from him in the same breath: (Arista Networks) is the richest of the set at a PEG near 2, and he says to wait for the 50-day moving average rather than buy it here; (Rambus) is paid per memory module rather than per price, so the very shortage that powers the thesis can cut against it, and there is an outstanding Department of Justice document request; (Fabrinet) is thin-margin and capacity-constrained, with Nvidia and Cisco together around half of sales. Those three are explicitly a different risk proposition from the first five, by his own framing, and he says he sizes the basket heavy at the mega-cap end and scales down from there.
The robotics basket, from Ross Givens, is , and . It carries both caveats from the section above — the promotional framing and the two symbols recovered from garbled captions.
The oil-into-costs question gets its first earnings test this window
(PepsiCo), (WD-40) and (Delta) report in this window, and Delta is the most direct read on whether a 7% jump in crude has begun landing in operating costs — which is the mechanism underneath the whole hike-versus-hold argument in the first section. Q2 bank earnings begin on 13–14 July and carry the first broad read on credit. 's 8-for-1 split takes effect on 10 July, which changes the quoted price without changing the thesis. And (Corning) was the largest pre-market mover at +5.96%, on news rather than on any stated argument — as were , , , and . Those are prices without a thesis attached, which is worth knowing before reading anything into them.
