1A 2026 rate cut has been priced out, and the trade that leaned on it broke
Kevin Warsh's first meeting as Fed chair was notable for what it withheld. Rates were held, the statement was cut to roughly 130 words, he declined to submit a dot of his own, and he stood up five review task forces instead. The dot plot did the talking: of 18 forecasts, one showed a cut this year and around half showed hikes.
The prediction markets took him at his word within hours. Polymarket's "How many Fed rate cuts in 2026?" contract — the deepest on the board at $3.2m of liquidity — prices zero cuts at 78.6%, one cut at 15% and two at 3.8%. The separate "Fed rate hike in 2026?" contract sits at 48.5%, close to a coin flip on tightening rather than easing, though on $145k it is a thin market and should not be read with the confidence the first one earns. The ten-year backs the direction if not the magnitude: 4.48%, up 10 basis points on the week.
What broke on the back of it was the debasement trade. Felix, on Felix & Friends, ties the gold, silver and bitcoin selloff — , and as the listed proxies — to three forces arriving together: a higher opportunity cost for holding assets that yield nothing, a stronger-dollar narrative, and geopolitical fear draining away, which is the subject of the next section. Bitcoin took the worst of it because leverage and margin calls forced the selling. He is explicit that he will not chase the bounce in the miners: none of roughly 50 major gold and silver names passes his breakout screen, the sector sits 20% to 60% below its highs (, , , ), and his framing for a higher-for-longer world is to upgrade quality and avoid the zombie companies.
He supplies his own counterweight, and it is worth holding alongside the call. Prices have already recovered 6% to 10% off the lows, June payrolls at +57k argue against the hike that the thin contract is flirting with, and the banks he tracks still carry year-end gold targets 25% to 50% above spot — he reads the whole episode against the 2022 hawkish scare, which crashed first and recovered after. So the near-term avoid and the medium-term case are both his, and they do not actually contradict each other. What they do require is a source of recovery other than rate relief, because the deepest contract on the board says rate relief is not arriving this year.
2The geopolitical premium is being priced as a shock that already happened
Iran carried the day's news flow — Khamenei's death and funeral, a US posture Trump framed as a deal or finishing the job, OPEC+ agreeing to raise output, the Suez reopening. Crude traded flat through all of it.
Polymarket explains that flatness better than the headlines do. "Will the U.S. invade Iran before 2027?" prices at 11.5% on $533k of liquidity, so the market's base case is negotiation and de-escalation rather than another leg of the war. Put that next to a flat oil tape and a historic loss of Iranian supply that has already been absorbed, and the disruption is being treated as an event that happened and got priced, not one still coming. The risk that remains is in depleted inventories, which is a slower and far less visible problem than an invasion — and a market that has stopped pricing the loud version of a risk is not the same as one that has retired it.
That fading fear is the same force draining the metals bid in the section above. One de-escalation does double duty: it takes the war premium out of crude and removes a reason to hold gold at the same time.
3The AI capex argument has moved from what it is worth to who absorbs the losses
Prof G Markets spent its episode on a thesis that the argument over AI spending has changed shape. The question is no longer only whether roughly $7trn of committed capex earns a return against the thin receipts booked so far; it is who is left holding the loss if it does not. The evidence offered is that OpenAI has reportedly floated a US government backstop and a stake of around 5%, with the Financial Times covering a discussion of a government guarantee. Stated plainly, that is an attempt to socialise AI capex losses onto taxpayers, and the episode's charge is that this is what positioning to be too big to fail looks like before anyone says the phrase out loud.
The financing structure underneath is what makes the worry systemic rather than a valuation quibble. is described as funding its own customers, with hyperscalers and venture capital subsidising demand from companies that lose money serving it — a circularity the episode maps onto both 1999 and the 2008 banks. The names arrive as exposure rather than as calls: more than any other at 39 mentions, alongside , , , and , plus the private labs OpenAI and Anthropic. The stated concern is systemic and valuation-level, explicitly not single-name timing.
Running the other way, in the same day's news, was the chip rally's own momentum — a near-term bullish read on semiconductors sitting directly on top of a bearish structural one. Both can be true at once, and that is the point. A bailout question is a policy question, and a tape can rally straight through a policy question for a long time. What it cannot do is settle one.
4The robotaxi thesis has moved past the technology to regulation and insurance
Equity Mates gave its "decade ahead" episode to autonomous vehicles, and the scale is the headline: a robotaxi fleet going from roughly 7,000 vehicles today to a projected 6 million over the coming decade. Their survey of ways to play it names 's Waymo as the current leader, 's unsupervised Houston fleet as real but small, 's Zoox in Los Angeles and San Francisco, as distribution with in-car advertising as the optionality, and alongside GM, Toyota and Mercedes-Benz as the less obvious automaker exposure. They were bullish on the structure and, in their own word, lukewarm on the near-term rollout pace — a survey, not a set of buy calls.
Where they put the risk is the interesting part. The three things they say to watch are a regulatory patchwork, with an EU framework due at the end of 2026 and China further ahead; the disruption of auto insurance as autonomous vehicles prove statistically safer than human drivers; and the scrutiny that follows a single bad incident, judged against a harsher standard than the one applied to human crashes. Not one of those is a question about whether the driving works.
Two names cut across the day. and are structural winners in this episode and bubble exposure in the one above, published within hours of each other. Same companies, same day, two frames that never meet — which is a fair description of where the AI-adjacent equity complex sits at the moment.
What the sources recommended
Two calls came out of the four videos, and both were reasons to stay away rather than reasons to own something. Two of the eight channels could not be reached this run and two had nothing new, so four videos is the day's sample, not the field.
. Ross Givens argues SpaceX's near-term support runs out this week. It joins the NASDAQ-100 tomorrow pre-market under a new 15-day fast-entry rule that skips the usual seasoning period, and the forced index buying — around $4.3bn, by some estimates as much as $7bn — is concentrated into the final ten minutes of today's cash close. That is the last catalyst on his list, and he reads the run into it as bought on the rumour. What comes next is supply: the first insider tranche, 20% of the total, unlocks roughly two days after the first earnings report in the first week of August, which he sizes at up to $400bn of potential supply against an $85bn IPO float. At 80 to 100 times sales he models $150 as the floor that gives way, then $135 — the IPO price — then $100 or lower during 2026, with a multi-year bottom and a bullish ten-year view after it. The video carries a hard sell for a $5 membership, which is worth knowing when weighing it; the call itself is dated and specific enough to be marked.
, , , . Felix's don't-chase-the-bounce line on the gold and silver miners, covered in the first section — a wait-for-a-setup stance on the sector rather than a view that it is finished.
The index bid gets tested at tomorrow's open
SpaceX's inclusion goes through pre-market tomorrow, which is the moment the mechanical buying described above stops and nothing steps in to replace it — the cleanest and nearest test on this page. Samsung reports the same day, cast in the day's news as the taster ahead of the chip earnings feast, and it is the first real read on whether the semiconductor cycle supports the momentum side of the AI argument or the balance-sheet side. The consumer and travel reads follow: and on 8 July, on 9 July, on 10 July. And agreed to buy Ultra Maritime for $3.45bn — a defence re-rating showing up in cash rather than in a chart.
