1A hot inflation print pulled the chip complex apart rather than selling it as one trade
The 10 June CPI print came in hot, and the first reaction was a broad futures selloff led by semiconductors — but inside the sector the selling went one way and the buying went the other. Google News' pre-market wrap has (Super Micro) down 12% in that move, while the memory and storage names diverged upward: (Micron) up 4% and (SanDisk) up 5%, with Micron joining the trillion-dollar club on the same theme. This morning's pre-market took most of the index-level damage back — S&P futures up 0.74%, the Nasdaq up 1.22%, the VIX at 20.6 and down 7.4% — and Yahoo Finance's wrap puts the bounce on memory and storage strength plus a ByteDance deal at (Qualcomm), not on a general chip recovery: was still down 3.7% pre-market. Dispersion around the print was wide in both directions, with (Marvell) up 9% on pending S&P 500 inclusion and (Oracle) up on a broker upgrade, against (uniQure) down close to 50% pre-market on a gene-therapy collapse.
Mohnish Pabrai, on My First Million, supplies the frame that fits that split, though from a back-catalogue episode uploaded on 22 May rather than from today. His argument is that the hyperscalers are playing a high-capex game whose payoff is genuinely uncertain, but that whichever of them wins, the money still has to cross the same toll bridges — , and . The caveat is his own and it is a large one: he owns none of the three, filing them under too hard or too expensive. So the person making the structural argument is not taking the structural position, and the episode is three weeks old, which makes it positioning philosophy rather than a read on this week's tape.
On The Compound's Animal Spirits, the AI-bubble-versus-AI-productivity argument ran the length of the episode without landing, with as the contested name. What the tape settled, at least for one session, is narrower than "chips sold off": memory and storage were priced as a bottleneck that gets paid whoever wins, and the AI-server assemblers as the part of the chain that absorbs a rate shock. If that split survives the next inflation print, it is a more useful fact than the index-level move it was hiding inside.
2The year's biggest listing was refused by the S&P 500, and the index buying it triggers is a rounding error
Brian, on Business with Brian, takes the SpaceX listing apart on its index mechanics. The S&P 500 refused it outright — index rules require profitability and SpaceX loses money — which removes the roughly $14bn of forced buying that inclusion would have required. The Nasdaq-100 and the total-market funds (, ) went the other way and rewrote their own rules to fast-track it inside five to fifteen trading days, counting founder shares toward the weighting. Even so, the total forced buying comes to $8–12bn, about 0.01% of a $75trn US market, and no existing constituent gets bumped to make room — the index simply carries more than 100 names for a while.
The price is the harder part. The listing is being discussed at roughly $1.75trn, about 95 times sales; the same multiple applied to would value it at $24trn, which is the cleanest way to see what is being asked. Morningstar's fair value is less than half that ask. And a growing share of what is on offer is not a rocket business: SpaceX now houses xAI's Grok data centres, roughly $2bn a month of revenue, with tenants including Google and Anthropic. Brian's own answer is to skip the debut and watch the December lockup expiry instead.
Two things to hold against that. It is one channel's reconstruction of index-inclusion rules, and nothing else in the day's material tests it — Yahoo Finance corroborates only the timing, calling the listing a day or two away. And the same video carried a paid promotion for an unrelated microcap sponsor, disclosed mid-video, which is a reason to weigh the explainer on its arithmetic rather than on the presenter's standing. Taken on the arithmetic, the consequence is that mechanical index demand is not what holds this listing up, and the float that matters arrives in December rather than on day one.
3The bear case on US equities is about the entry price, not the economy
Pabrai's other argument is aimed at the index rather than at any company inside it. At a price-to-earnings ratio near 23, and citing Howard Marks, he points out that the historical record for forward ten-year returns from that starting multiple runs between minus 2% and plus 2% a year — the valuation is the handicap, before the economy gets a vote. His own book sits deliberately outside it, in what he calls the "hated and unloved": (Constellation Software), on the argument that entrenched vertical software is far stickier than its detractors assume and that AI will not dislodge it, alongside Turkish warehouses in Reysas, the airport operator TAV, and coal. He prefers gold to Bitcoin, and puts both in his too-hard pile.
The near-term reading from The Compound's Animal Spirits points the other way, on a longer run of evidence: the economy has shrugged off inflation, rate hikes, tariffs, fuel prices and successive recession calls, which is what the episode's Teflon framing is about. The latest jobs report then came in much stronger than expected, and under a Fed still weighing cuts that strength is the problem — a good labour market reads as a bad stock market — with the 10 June CPI print stacking straight on top of it.
The two are not in direct conflict. One is a ten-year claim about the price of entry, the other an observation about several years of resilience, and neither says what would end the streak. The Animal Spirits episode does name the candidate: oil feeding into inflation expectations. That channel is live right now — Finnhub's feed ran US–Iran exchange fire for a second day undermining a shaky ceasefire, cited alongside hot US inflation as the twin overnight risk drivers, and (Kosmos Energy) rose 10% on the escalation. Gold is rising in both accounts. So the escalation is not only an energy story; it is the shortest route from a fragile ceasefire to a Fed that stops cutting.
What the sources recommended
Two buys came out of four videos, and the day's one bearish view was aimed at the index rather than at any single name.
(Outlook Therapeutics), from Ross Givens, is an explicit catalyst trade on a dated clock. The FDA has reversed two prior rejections, and the company's 1 June Class-1 resubmission forces an answer within 60 days — a decision expected by 1 August, against a normal six-month review. Its wet-AMD eye injection is already approved and selling in Germany, Austria and the UK, which is why he frames US approval as the last domino. The size of the thing is the point: an $87–90m market capitalisation and under $1 a share, against $40 three years ago and 20c last month. He would buy only on a chart breakout, and he names both risks himself — another rejection, and dilution from a late-May raise while the company burns cash. Weigh it knowing the same video is a repeated hard sell for the presenter's own paid subscription service.
(Constellation Software) is Pabrai's long-standing holding rather than a new idea, it comes from a three-week-old episode, and it is not US-primary-listed — it trades in Toronto, with only an over-the-counter line in the US.
The resilience read gets its first test after tonight's close
(Adobe) and (Lennar) report after the close, and Finnhub's calendar has them as the only large caps in a day that sits inside 160 earnings events through 25 June. Adobe is the nearest thing on the calendar to a test of the software-durability argument in the third section; Lennar is a direct read on what a stalled rate-cut path is doing to housing demand. The listing mechanics resolve faster still — Yahoo Finance has the SpaceX debut a day or two away, alongside a Quantinuum quantum listing and a wider space-stock frenzy — so the arithmetic in the second section gets marked almost immediately. The one dated test that cannot resolve this week is the FDA clock, which runs to 1 August.
