← News & Research
RESEARCH DIGEST · TUESDAY 4 AUGUST 2026 · 1:54 PM EDT
Written by AI, which can make mistakes. Not financial advice.

These pages are written by AI from podcast transcripts, market news, macro releases and prediction-market pricing. It can misunderstand what a speaker meant, attribute a view to the wrong person, or get a figure wrong.

Where a ticker carries a + or , that is our reading of the direction a source leaned — not a recommendation from AlphaDrift, and not necessarily a formal recommendation from the speaker either. Check the original before you rely on any of it, and speak to a licensed financial adviser about your own circumstances.

A $45bn fund's collapse was already a closed trade by Tuesday — and the September hike lost its lead the same day

4 videos5 news & macro sources5 prediction markets4 things worth your time

1A $45bn fund went to $10bn in six days, and the bounce its selling caused was fully traded in three and a half

Michael Green, of Simplify Asset Management, told Prof G Markets that Leopold Aschenbrenner's Situational Awareness fell from roughly $45bn to roughly $10bn in six days — days after a 24 July investor letter reported a 439% first half and invited fresh capital. The public book was fire-sold to Citadel. Green's mechanism is arithmetic rather than misfortune: about 400% leverage against high single-name volatility, where forced selling moves prices against the longs and the shorts at the same time. Holdings named in the segment included and .

His read-through is that the structure is not unique to one fund. US leveraged-ETF assets are at a record $218bn, up 60% since March, and they carry the same daily-rebalance volatility drag: at three times leverage against April–May semiconductor volatility, you needed more than 170% a year simply to break even. South Korea is his worked example — the KOSPI down 44% from its June highs, more than a million margin calls, exchange closures, and leveraged ETFs now banned there. He argues a US-domiciled, unlevered memory ETF exported the shock. One caveat travels with it: Green runs volatility-aware strategies, so this is a manager describing the risk his own product is built around.

The same forced selling was, to Ross Givens, a trade with an expiry date. He put $10,000 into each of ten names sold down in the liquidation and closed the whole basket on Tuesday, roughly $14,000–15,000 up, about 15% in three and a half trading days. He was explicit that it was a mechanical-overselling trade and never a hold, and his exit was price location rather than any change of view: the names had run straight back into still-declining moving averages, so he took the gain. Only four of the ten are disclosed — , where a double-entry wash sale distorts his displayed cost, and , both of which he flagged as having barely moved, plus a fourth the auto-captions render only as "core". CoreWeave is the closest fit there, though Core Scientific or Corning cannot be ruled out.

Brent Donnelly, of Spectra Markets, gave Excess Returns the general form of both halves without naming a stock. His first rule is to avoid ruin before chasing edge: most blow-ups, he argues, are short-optionality structures that were identifiable in advance, and he passes on high-expected-value trades whose risk he cannot measure. His second is horizon discipline — pairing a multi-year structural thesis with a two-day stop is the recurring retail error, and valuation shorts are his worked example. Givens' basket is the inverse case: a three-day thesis held for three days.

This happened into a record tape. The Dow closed at a record, closed above a $3trn valuation for the first time, and was up 12.9% from the open on its second-quarter commercial revenue, with named alongside it as a leader in the session. The consequence is narrow and worth saying plainly: a fund can be liquidated out of existence inside a week without the index registering it, and with $218bn of leveraged-ETF assets outstanding, the mechanism that did it has not been retired.

2Crude is pricing a ceasefire, not a settlement — and the shipping lane is priced to lag both

The largest odds move in the day's set was the US–Iran ceasefire. Polymarket's two-week-pause contract for 14 August rose 11 points to 80.5%, on $139k of 24-hour volume and $72k of liquidity, and the 31 August tranche rose 8 points to 83%, on $130k of volume and $87k of liquidity. A flat-to-upward term structure like that is not a market pricing a one-week truce that lapses. Bessent's line that a Hormuz deal is close is being priced as credible rather than rhetorical, and a Tuesday reshare on President Trump's feed carried the same framing — a deal described as imminent as talks restart on denuclearisation and Hormuz.

A settlement is a different contract, and it is priced nowhere near the truce. A final US–Iran nuclear deal by 31 August sits at 3.75%, down 0.7 points, on $139k of volume; by 31 December it is 34.5%. So the day's oil move is a ceasefire trade, and ceasefires reprice in both directions.

Between those two sits the shipping lane. Hormuz traffic returning to normal by 31 December prices at 62.5%, up 6 points, on $215k of volume and $290k of liquidity — the deepest liquidity of the Iran-linked contracts checked. That sits 18 points below the ceasefire odds, and the gap has a named cause: Reuters reports Iran demanding inbound control and outbound oversight of the strait. Fighting is expected to stop well before the lane is expected to work.

Crude sold off on it. was down 1.3% from Tuesday's open even as the broad tape rallied to records. The distinction worth keeping is that a truce caps the escalation premium building, while only a settlement returns the strait to normal working order — and the market prices the truce at four-fifths and an August settlement at under 4%. reported quarterly profit doubling to more than $5bn on the war-driven oil surge, which is a backward-looking quarter if Hormuz reopens.

3The September hike lost its lead, on a coin-flip margin

The deepest market checked on Tuesday was Polymarket's September Fed decision, at $1.54m of 24-hour event volume, and it flipped. No change prices at 49.5%, up 7 points; a 25bp increase at 47.5%, down 8. A day earlier the hike side led at 55.5% and was falling. A 25bp cut sits at 1.75%, with 50bp moves either way near 1%. The argument in the price is hike-versus-hold, and easing is not in it.

Two prints did the work. Job openings fell to 7,359k in June from 7,537k in May, released Tuesday, which is labour demand cooling. And the ceasefire repricing took crude with it, weakening the energy-driven inflation impulse the hawkish case leans on. Neither settles anything: the hawkish tail survives at 47.5%.

That is the whole finding, and it is a modest one. A hold at 49.5% against a hike at 47.5% is not a forecast — it is a coin flip waiting on Friday's payrolls.

4The long end sold off in the same week the front end turned dovish

The 10-year Treasury yield sits at 4.75%, up 14bp on the week, with a 2.47% real yield — a rise into a week whose data pushed the front end the other way. Katie Martin, of the Financial Times, told Prof G Markets why she thinks the two can move apart. The long end, on her read, is selling off because market participants cannot follow the Warsh Fed's reaction function. That makes it a term-premium problem rather than an inflation one, and term premium does not reverse on a softer job-openings print.

The same week produced the sharpest currency intervention in fifteen years. Tokyo sold roughly $59bn on Thursday, and the US Treasury joined on Friday — the first US-joined intervention since 2011 — paying for it, unusually, by selling euros. Martin's read of the American motive is not about the yen at all: it is about stopping Japan from dumping Treasuries into an already-weak long end. She is careful about the premise too, saying it is not clear speculative accounts were attacking the yen in the first place, so the framing of the move as a defence against an attack is contested.

Felix, on Felix & Friends, describes the plumbing built for exactly that problem. The Fed's FIMA repo facility — created in March 2020, made permanent in 2021, with a standing limit of about $60bn per foreign central bank — is being opened to Japan at scale, with Bessent pushing for the cap to be raised. The mechanism is the point: Japan pledges Treasuries for dollars instead of selling them, so it can defend the yen without adding supply to the US long end. His precedent is the coordinated central-bank dollar-liquidity action of March 2023, around the Credit Suisse collapse, and his pattern is that the dollar taps open when something is cracking. His positioning conclusion — that cash is the wrong place to sit when a liquidity pipe opens, because asset prices respond to dollar supply rather than company quality — arrives inside a heavy promotional funnel: a free companion report, a free beginner seminar, a paid programme and a screening-app trial are pitched in the same video. The mechanism and the sales pitch travel together there.

One long-horizon counterweight came from Brent Donnelly on Excess Returns, who argues that permanent bearishness is "the biggest leak in finance", and that deficit spending alongside inflation below roughly 3.5% has empirically been bullish for equities rather than bearish. He presents that as an empirical regularity and cites no source for it in the interview, which is how it should be weighted. The consequence for the week is narrow: two policy actions, an FX intervention and a repo line, are aimed at the same pipe — the US long end — and if the term-premium read is right, that is the pipe softer data cannot fix.

Four buys and four sells came out of two of the day's four videos, and neither set is quite what a table makes it look like: one buy had not triggered, one was bought live on camera with a stop attached, and the four sells are a single profit-take rather than four judgements about four companies.

The buys. , from Felix & Friends, on a price/earnings ratio of 19 that he calls the cheapest in seven years — with cloud revenue around $12bn a quarter growing 80% year on year, a backlog near $500bn, and a $15bn Texas data-centre deal with Anthropic that reframes the capex as pre-sold demand. He adds that the chart has cleared three resistance lines after a 14% bounce. , from the same video, is the deep-value case: a $3bn market capitalisation against $2.6bn of cash plus roughly $600m generated this year, a $35 cash bid from Hapag-Lloyd against a share price near $26, and Red Sea rerouting removing 8–10% of global container capacity against 5% trade-volume growth. The caveats belong to the call. The video itself says the market prices that bid as dead on Israeli golden-share politics, those politics are described rather than handicapped, and the negative-enterprise-value figure ignores vessel-lease and debt obligations.

, from Ross Givens, is a copper-miner call on a base-on-base breakout in copper futures, on the argument that miners give leveraged earnings to the metal — a 20–30% move in copper driving 40–60% profit growth. He bought it live near $66 with a stop around $60.50, about 9% of risk, and named , the copper ETF, only to pass over it. is conditional and had not fired: he asks for a break above roughly $383 on the fund, equivalent to gold futures above $4,190–4,200, and was trading $375.65 mid-session. Gold has based after a 29–30% drawdown from its January peak, with dips getting shallower into that resistance.

The sells. , , and the unverified "core" name are the disclosed quarter of Givens' ten-name liquidation-bounce basket, closed together for the gain. They are a profit-take on price location, not a view on the businesses, and six of the ten names were never said out loud — so what is visible here is a fragment of the trade rather than the trade.

Both channels that made single-name calls run paid subscription funnels alongside them: a paid academy and a screening-app trial in one case, a $5-a-year trading service in the other, pitched with a performance claim about what one video's advice was worth. Neither was flagged for undisclosed sponsorship or fabricated performance, disclaimers are present in both, and the framing is promotional throughout.

Friday's payroll print is where the 49.5/47.5 split gets settled

The July Employment Situation lands on Friday 7 August — the first payroll print since Tuesday's job-openings softening, and the swing factor for a September decision currently split 49.5% hold against 47.5% hike. July CPI follows on Wednesday 12 August, and it is the first inflation print to capture the post-peak crude move, which is the same mechanism the ceasefire pricing rests on; July PPI comes a day later. Earnings are thin either side: , , and on Thursday, then , , , and on Friday. The Iran talks run underneath it, with Qatar confirming progress and Iran's demand for inbound control of Hormuz still the open item — the same demand the shipping-lane contract is discounting.

Two company prints landed away from the day's themes: beat on the quarter and raised the low end of its revenue guidance on non-Covid products, and had the 737 MAX 7 certified by the FAA.

Who called what4 buys · 3 exits

Every single-name call the day's sources made, as they made it. The caveats attached to each one are in the article above — a ticker in this table is not a recommendation from AlphaDrift.

Buys

TickerCompanyHorizonConvictionWho said it
AlphabetmonthshighFelix & Friends
ZIM Integrated Shipping ServicesmonthshighFelix & Friends
Freeport-McMoRanweeks-monthshighRoss Givens
SPDR Gold SharesweeksmediumRoss Givens

Exits

TickerCompanyWhyWho said it
Jack in the BoxProfit-take, not thesis break — closing the whole 10-name Aschenbrenner-liquidation bounce basket after +15% in 3.5 trading daysRoss Givens
Micron TechnologySame basket close; flagged as one of the two names that "haven't really done much"Ross Givens
Western DigitalSame basket close; same flat-performer note as `MU`Ross Givens
Mentioned today

How to read the tickers
TICKERa source leaned toward buyingTICKERleaned toward selling or trimmingTICKERmentioned, no direction givenNAMEunlistedprivate or pre-IPO — no symbol to chart

Sources scanned · Tuesday 4 August 2026

More research

How the analysis works

Views attributed to named sources are theirs, linked to the original in every case. AlphaDrift holds no position on the basis of anything published here.