1Large-cap tech ran six standard deviations rich, and Friday's crack came partly from the plumbing
Nick Kohus and Jessica Rae of Data Track Research, on The Compound, put a measurement on how far the tech trade had travelled. , the large-cap tech sector, outperformed the S&P 500 by 29.3 percentage points over the 50 days to 2 June — a six-standard-deviation reading, the most extreme in a dataset running back a decade, and one they say still stands with Friday's selloff included.
Friday itself took 6.66% out of in a session, and their mechanism is worth separating from the trigger. A hot jobs report and a two-year Treasury yield at a twelve-month high supplied the reason to sell; market-making algorithms then widened spreads after such an anomalous run, which amplified the move rather than started it. A crowded trade does not need much of a reason to unwind. It needs thin liquidity, and it got some.
The gap inside tech is where their call sits. Semis beat software by 44 points over the same window — four standard deviations, again the widest in a decade — so they want semi exposure reduced, software bought against it, or tech held at index weight through to cut single-theme concentration. Composition matters to that trade: , which they say "came back with a vengeance", is a heavy weight on the software side, cyber names (which Kohus owns) and ran hard, and "sat things out".
What stops this becoming a call on the whole market is the index itself. The S&P 500's own 50-day run is 15.3% against a 2.4% average — two standard deviations, not six. The five prior two-sigma crossings since 2015 were each higher 50 days later, by an average of 7.3%; five observations is a tendency rather than a law, but it points where the rest of their work points, which is constructive on the index and cautious on what has been leading it. The wider tape reads the same way, with at a new all-time high while fell 1.89%.
The awkward part for the rotation call is that semis were bouncing as it was made. rebounded 6–7%, joining the $1trn club with a Wells Fargo $1,220 target, rose 11.2%, and , which led Friday's drop, was back up 1.73%. If a six-sigma stretch unwinds, it should show up as semis giving ground to software rather than as the index breaking — and on this tape it has not started.
2The rate question has flipped from when the Fed cuts to whether it hikes
Kohus and Rae call rates "a new dimension" in the tech story, and the change they describe is directional rather than incremental. Goldman Sachs has removed its year-end cut. The two-year Treasury yield sits at a twelve-month high. An argument about the timing of insurance cuts now has a hike inside it, and Reuters and Bloomberg items in the day's news feed carried the same thread.
The date to hold is 16–17 June: the FOMC's next meeting, and the first chaired by Kevin Warsh. A first meeting under a new chair offers no reaction function to lean on, which matters more than usual a session after a hot jobs print and a twelve-month-high two-year yield took 6.66% out of the most crowded part of the market in a day. Equities that re-rated on the expectation of easing have to carry the possibility of the opposite for another week.
3The bull case for the biggest IPO ever rests on forced index buying, not on the valuation
SpaceX prices at $135 a share, roughly $1.775trn, and trades on Friday 12 June from 8:30am Central — the largest listing on record. The video making the buy case is titled "SpaceX Isn't Worth $135 - But I'm Buying It Anyway", and Ross Givens means the first half of that literally: his own fair value is $42–43 base, $16 bear, $84 bull, against Morningstar at about $60. His sum-of-parts has Starlink ($11.5bn of revenue, up 50% year on year, $4bn of operating income) carrying a rocket business that loses money and is all-in on Starship, plus xAI and Grok at roughly $3bn of revenue and a $6bn loss.
He buys it anyway, and the reasoning is mechanical rather than fundamental. The deal is oversubscribed; retail gets 30% of it at Musk's insistence, through Robinhood, SoFi, Fidelity, Schwab and E*Trade; and index funds then have to buy at whatever price exists — Russell 1000 fast entry around five days after listing, NASDAQ-100 after fifteen, the S&P later, which he sizes at $15–30bn of forced demand. He thinks $200 in week one is reachable. The caveats travel with the call: he has filed an indication of interest at $135, says he will sell into strength if he is allocated, and will not chase at $150–160 without an allocation — an allocation trade with an exit, not a position. The same video pitches a $5 paid service twice and runs on lines like "everything Elon touches turns to gold", which is context for how much weight the call carries.
Kohus and Rae reach the valuation from the institutional seat and land close to him: $1.7–1.8trn is a stretch, institutional bids show up nearer $1.2–1.5trn, and retail sets the day-one price. Two very different desks agree the price is wrong and disagree completely about what follows — one leaves it alone, the other buys the flow.
That flow argument has a live analogue on the tape. Marvell rose 7–9% on its S&P 500 inclusion, effective 22 June, per the day's Yahoo and Google News items — index demand moving an already-listed name on the announcement alone. Nothing in the SpaceX case is tradeable before Friday, and if the mechanics are right, the first evidence is a price that ignores every fair-value estimate attached to it.
What the sources recommended
Two episodes produced four calls, and the largest of them is not tradeable until Friday and carries a warning from the person making it.
The buys. SpaceX, from Ross Givens, explicitly as a momentum and index-flow trade and not a value one — he puts fair value at roughly a third of the offer price and plans to sell into strength if allocated. — software, on a horizon of weeks, bought as the other side of the semis-software gap. — tech at index weight, from Kohus and Rae, as the low-concentration way to hold the theme; the same pair flagged tech broadly as stretched enough to "be cautious, keep looking for the next car", so it is a hold with a caution attached rather than an endorsement.
The sell. — reduce semi exposure on rotation risk, the day's only outright reduce, and the direct counterpart to the software buy.
Both of the day's arguments get tested inside a week
Six companies report today — , , , , and — with the consumer names carrying the most information after Campbell's stuck to its targets while describing US shoppers as tightening their spending. SpaceX lists on Friday and the Russell 1000 fast entry follows about five days later, so the forced-buying claim gets marked almost immediately. The FOMC meets 16–17 June under a new chair, which is where the rate question stops being rhetorical, and 's S&P inclusion takes effect on 22 June. Underneath all of it, the Iran–Israel ceasefire recurs through the day's Reuters and Google News items and is holding, which is what keeps a war premium out of crude — the Strait of Hormuz is why that tail never fully closes.
