1A two-month semiconductor parabola broke, and the rotation trade named against it is the electrical grid
The tape supplied the first half of this argument. The semiconductor index fell 11% intraday on Tuesday and Marvell fell 18.5%, after the Nasdaq had already dropped 4.8% on Friday. Ross Givens, on his own channel, reads that sequence as a top rather than a dip — back-to-back gaps, a buy-the-dip Monday, then a Tuesday that rips down through it. His instruction is to take semiconductor profits now, naming , , and , while conceding in the same breath that the group "could go higher" and that he is working from topping signs, not a broken fundamental.
The second half is where he says the money goes. The destination is electrification: (Quanta Services) on its first pullback to the 50-day moving average, (Johnson Controls) on a breakout that held a 2% gain on a red day, and (Eaton) on the same pullback logic but flagged as the choppier of the three and a roughly $400 stock. At group level, and for the power and grid build-out, and for materials — which he defines by its holdings, Linde, Newmont and Nucor, rather than as a broad-economy bet. The macro claim underneath all six is that US electrical output has to double or triple over five to ten years, funded by federal grid spending and hyperscaler capex.
Two things in the day's news cut against the sell side. Marvell , the worst-hit name in the dump, joins the S&P 500 on 22 June — scheduled, mechanical buying arriving straight after the fall. And Micron spent the day in headlines about approaching the "$1T club." Neither is a fundamental rebuttal of a topping call, but both are the kind of flow that punishes an early exit. Worth weighing against the source too: the episode runs a disclosed paid-service promotion throughout, a $5 subscription with a repeated call to action and a QR code. The named stocks are not themselves the product, but the calls arrive inside a sales funnel.
The exit and the entry here rest on different kinds of evidence — the exit on price behaviour over three sessions, the entry on a decade-long spending programme. Only one of those can be marked wrong by Friday.
2Index plumbing, not fundamentals, is the argument for mega-cap selling over the next fifteen trading days
Felix, on Felix & Friends, makes a case with no company analysis in it at all. He counts roughly $350bn of equity supply landing over the coming weeks: a triple listing — SpaceX at about $75bn, OpenAI about $60bn and Anthropic about $60bn, so roughly $200bn between them — on top of mega-caps printing new stock, with Alphabet's $85bn already done and Meta reported to be next. Meta fell 5–9% on that dilution rumour.
The third leg is what turns it from a sentiment call into a forced one. Under the Nasdaq's fast-entry rule, index and Nasdaq-100 funds have to buy SpaceX within 15 trading days of its listing, and to fund that they must sell what they already hold. The easiest-to-sell names go first, which is how he arrives at his list: , , , , , and . The concentration figures are why he thinks it matters — the top ten names are about 40% of the S&P 500 and delivered 72% of its gains this year — and why he treats , and as the exposure rather than the safe harbour.
He is careful to say he is not alone in the direction, citing Pabrai as holding no S&P or mega-cap exposure, Berkshire as having sold its S&P index fund, and Tom Lee — a standing bull — now warning of a 20% correction on a mix of Fed-leadership uncertainty, AI repricing and trade-war headwinds. One piece of corroboration arrived from the tape rather than the pitch: Broadcom was being dragged in the day's news by weak AI guidance, and Intel was down 2.1%.
Read the seven-name list as one call, not seven. And read the packaging with care: the episode is built as a fear funnel — a free Saturday seminar on the "index fund trap," an app trial, an academy — and it shows no fundamental work on any name it tells you to sell, nor a single ticker on the buy side. The mechanism it describes is checkable within fifteen trading days, which is more than most macro theses offer. The sales structure around it is not evidence either way, but it is the reason to check rather than take it on trust.
3A forecast for US growth 50% above consensus arrives with a case for leaving mega-cap tech
Joe Davis, Vanguard's global chief economist, gave Excess Returns the most bullish growth number of the three videos and the least intuitive positioning to go with it. His projection is about 3% US GDP growth in 2027, roughly 50% above consensus, and he attributes it to AI automation and augmentation — with demographics and globalisation contributing nothing to the forecast. He places the economy at "1996-97": a build-out phase with one to two years still to run.
The positioning does not follow the growth. Near-term momentum favours mega-cap US growth, he says, but for risk management and for the payoff from AI's second phase he would rotate outside it — into the adopters with unmet need, which he lists as healthcare, financial services including advice, education and business services. That is a value orientation reached from an AI-optimistic premise, which is the part worth sitting with.
He also puts a number on the downside most AI bulls leave unquantified. If AI only automates rather than lifting productivity broadly, fiscal pressure bites: he assigns a 20% probability to the 10-year Treasury yield exceeding 9% within five to ten years, which is why he prefers shorter-duration fixed income. He expects a significant drawdown or consolidation within two to three years, on the precedent that big technologies have had them before, and declines the word "bubble" on the grounds that this is intrinsic economic rewiring rather than a mania.
Set the three videos side by side and they point the same way out of mega-cap tech on entirely unrelated evidence: a chart top, an index rule, and a decade-long adoption thesis. They also run on three different clocks — weeks, fifteen trading days, five to ten years — and only the middle one can be settled this month. Agreement between horizons that far apart is weaker corroboration than it looks, because none of the three can falsify another. Two of the three episodes carry disclosed promotional flags; the one that does not is the one making the slowest claim.
4Oil is bid on strikes around Hormuz, and this morning's inflation print is too old to speak to it
The most-repeated story across the day's news was the US–Iran conflict and strikes around the Strait of Hormuz, carried six times over the Reuters wire — oil up, broad equities down, energy and oil services bid. Kosmos Energy was up 10% pre-market on it.
May CPI lands this morning, billed in Yahoo Finance and CNBC coverage as the "scariest inflation report," with uncertainty about a new Fed under Warsh layered on top and volatility already elevated at a VIX of 20.4. But the print measures May, and the energy move is happening now — so a benign number cannot settle the question the oil bid is asking. Two smaller macro prints sit alongside it: China's factory-gate inflation ran at roughly a four-year high, and gold sat at an eleven-week low on a firmer dollar.
That leaves a two-sided morning — a backward-looking inflation reading arriving into a forward-looking energy shock, with an unusually wide range of plausible Fed reactions sitting behind both.
What the sources recommended
Six buys and thirteen sell lines came out of three videos, and the sells are two theses rather than thirteen independent judgements — while two of the three channels disclosed a paid promotion running through the episode.
The buys. The six come from one video and amount to one trade: electrification. (Quanta Services) and (Eaton) on first pullbacks to the 50-day moving average, with Eaton flagged as the choppier of the pair; (Johnson Controls) on a breakout that added 2% on a red tape. At group level and for power and grid infrastructure and for materials, that last one defined by Linde, Newmont and Nucor. Horizon: weeks for the single names, weeks to months for the ETFs.
The sells. Ross Givens' semiconductor top gives , , and , all framed as taking profits rather than shorting. Felix's forced-unwind gives , again, , , , and , plus on the share-printing report, and extends to the index wrappers themselves — , and . is the only name both arrive at, and they arrive from unrelated reasoning.
The gap worth naming: Felix's buy side carries no tickers at all. He points at coal and energy, oil services, transportation, biotech and basic materials, and names nothing inside them — so the actionable half of that episode is the exit list alone.
The mechanical half of this week's argument has a date; the rest does not
May CPI prints this morning into that VIX of 20.4, and it is the only event on the calendar that can move all four arguments above at once. Retail earnings follow today from , , and — a read on the consumer that none of the three videos touched.
After that the tests separate sharply. The index-rebalance case carries its own deadline: fifteen trading days from SpaceX's listing, which either shows up in mega-cap flow or does not. Marvell enters the S&P 500 on 22 June, putting scheduled buying into the most-sold semiconductor name in the group — the cleanest near-term test of the topping call. The electrification and second-phase-AI cases have no date attached at all, which is what makes them hard to be wrong about in either direction.
Elsewhere in the tape, the moves were idiosyncratic rather than thematic: Nuvalent rose 39% on a reported acquisition above $10bn, Qualcomm jumped on a ByteDance deal, American Airlines added 3.6% on a Starlink in-flight wifi agreement, and Boeing posted another rise in deliveries. On the other side, Super Micro fell 7.6%, Applied Optoelectronics fell 17.2%, and SanDisk sat in a memory complex down 3–6% — the same corner of the market the first argument is about.
