Jay Singh — Weekly SSR: CPI/PPI week, CENTCOM disables three Iranian tankers, the 40-year bond bull market declared over, private-credit gates re-imposed — and two new event-driven names, Entravision (EVC) and Himax (HIMX) (Sep 7 2026)

Ticker	Name	Research	View	What Singh said	Ref
EVC	Entravision Communications	QT · SA · STK · FA	Positive	The week's featured value special situation — a broadcaster the market still prices as a broadcaster. "One of the names we are interested in is a company called Entravision, which is sensitive to advertising, but it is quite an interesting name going into midterm elections because its advertising spend should explode given its key Hispanic demographic in specific states that are the most important in the US when it comes to ad spending during the midterms." The transformation is in the numbers: the Advertising Technology and Services (ATS) segment did revenue +230% to $182M and operating profit +673% to $40M in Q2 2026, against a legacy TV/radio media segment running a $3.3M operating loss. "There's unrivaled access to the US Hispanic demographic… a dominant media footprint targeting the fastest growing demographic and voting bloc in the US, giving it significant pricing power during political election cycles," plus "over 83 million of cash and marketable securities" and a model blending recurring digital spend, sticky retransmission-consent fees and "cyclical political ad windfalls." Why it is cheap: management "guided for ATS revenue to decrease sequentially from Q2 to Q3 due to normal customer spending timing. This pause in EBITDA growth resulted in a big sell off in the stock," taking consensus for Q3/Q4 down to $18-25M a quarter. His mark: "2026 adjusted EBITDA should be around 100 to 115 million… based on consensus forward EBITDA estimates, EVC trades at a forward multiple of about seven and a half times to eight and a half times, moderately below media industry averages of 10 to 12 times" — versus a trailing EV/EBITDA of 12-12.4×. The verdict: "at eight times, EVC is priced at a discount compared to pure play ad tech platforms which often trade above 12 to 15 times forward EBITDA, but at a premium to legacy radio and broadcasters which trade at only four to six times. If ATS maintains double digit growth while political ad dollars flow in, the forward EBITDA could surprise to the upside, lowering its effective valuation floor." Risks named: the TelevisaUnivision affiliate contract expires at the end of this year, sales commissions and cloud costs grew 85%, and large digital clients "can create quarterly lumpiness" — but on AI disruption, "ad tech volatility when it comes to TV is not as sensitive to AI trends as internet. So this is more safe than internet advertising."	53:55 · 57:39
HIMX	Himax Technologies	QT · SA · STK · FA	Positive	The week's featured growth special situation, and the only name he says outright he is looking to add. "HIMX is another very cheap name to play the AI tailwind that we're seeing in the market. And frankly, you can hedge with any of the AI ETFs to take out the beta from that name." Asked in Q&A whether digital infrastructure is too pricey: "we're just waiting, and looking at less followed names like Himax to add to for now." What it is: "a Taiwan-based fabless semiconductor company… instead of manufacturing its own silicon chips in factories, Himax designs the integrated circuits and outsources production to foundries like TSM," selling as a tier-two supplier into panel makers (AUO, BOE) and tier-one auto electronics, and from there into Apple, Dell, Samsung and the global automakers. The moat is automotive: "Himax is the global market share leader in automotive display driver integrated circuits… Himax owns 40% of the global market share for automotive displays." The optionality is edge AI and glasses — WiseEye ultra-low-power always-on computer vision, LCoS micro-displays and wafer-level optics — and "all of the Meta Ray-Bans use Himax. So as you see those sales increase, it's positive for the company." The setup: Q2 revenue $227M, +14% sequentially and +5.9% y/y, snapping a four-quarter stretch of declining revenue growth, EPS $0.11 (a penny light) on 33.1% gross margin, exceeding guidance, with small/medium display drivers +19.6% to $162M and now 71% of sales. Guidance: +7-11% q/q to $243-250M, +23.8% y/y at the midpoint, gross margin +90bp to 34%, inventories flat at $150M — "so no red flags… but the market doesn't seem to be rewarding this strong beat." His read on why: "its last quarter beat and its stock is down simply because of the AI sell off." On valuation: "while Himax does look expensive when viewed through trailing multiples, on a forward basis it looks a lot more reasonable — Himax's PE on a forward basis is around 30, but the business has so much potential. The TAM in the smart glasses is so high… and the fact that the business is growing and auto has bottomed, means that there could be upside from here."	1:06:46 · 1:29:37
AVGO	Broadcom	QT · SA · STK · FA	Positive	Bought on the print, with the arithmetic stated. "We bought a little bit of Broadcom. Hock Tan's actually guiding to $30 of EPS by fiscal 28. On a 20 multiple, that implies AVGO could almost double. The company's growing at 50%. Now imagine they do 40 or 50 of EPS by 2028, 2029. This stock is basically pricing in that growth is peaking." The quarter: revenue ~$30B vs $29.25B consensus, EPS $3.32 vs $3.21, free cash flow $13.7B at a 46% FCF margin. The guide is the story: "the hyperscaler lock in — Broadcom secured commitments to enable over 20 gigawatts of compute capacity over its top customer base including Google, Anthropic, Meta and OpenAI through 2028," with FY26 AI revenue guided at $58B, FY27 doubling to $115B, and doubling again to $230B by 2028 — "which is quite aggressive when it comes to guidance." Q4 guidance of $34.8B implies a 93% sequential step up with Q4 AI revenue of $21.7B, +236% y/y. His reaction to the tape: "so we bought some Broadcom. I don't know why the market was not happier." Later, on the supply-constraint theme: "Nvidia is guiding to 70% growth. AVGO is guiding to 100% growth. Both are supply constrained and saying real demand is higher, but the market is not giving them credit." (Morgan Stanley's caveat, quoted: the $115B FY27 number came "slightly below the firm's $120 billion estimate," still in line with management's prior "well above $100 billion.")	51:16 · 1:18:32
SPGI	S&P Global	QT · SA · STK · FA	Positive	The spin-off leg of the special-sits book, and a clean multiple argument. "S&P Global shares rallied last week on news that it was looking to spin off Capital IQ, which is the most AI-exposed part of the business. So that's obviously positive because Moody's trades at a higher multiple than SPGI." Shares "traded around 450" at the end of the week and "closed on the fourth around 443 after hours." The comp set: "FactSet is probably the best comp for the SPGI Cap IQ spin, which trades at around 15 times PE and 12 times EBITDA. So the Cap IQ business should trade at a 25% lower multiple. SPGI trades at 22 times forward EPS versus MSCI at 26 and Moody's at 27." The deck adds the strategic logic: coming two months after the Carfax/Mobility separation, a Cap IQ spin "would further weight the business towards the higher growth / margin portions of its business, namely Ratings & Indices."	06:08
LPSN	LivePerson	QT · SA · STK · FA	Positive	A closed winner, and the clearest example this week of the microcap-spread edge the product is built on. "There's also a microcap spread which we talked about in the Discord, which we thought was an easy layup, which was LPSN. As you recall, we bought it in the 280s. It rallied up to 316 and it looks like the deal has now closed. So that was a very quick, almost like a two-day trade — and now the shares were halted, but it looks like the deal is done." A ~13% gross move on a sub-$5 name in days, sized and alerted to the Discord rather than the weekly call — the type of spread institutions cannot trade and the reason he keeps a microcap sleeve at all.	06:08
PYPL	PayPal Holdings	QT · SA · STK · FA	Positive	The same range trade as last week, executed again, with a new bid rumour on top. "There's also a new bid for PayPal possible according to Betaville, which is an M&A blog." The history: PayPal "rallied about 40% from its unaffected price from the summer up to August. Then there were rumors that the Stripe deal failed. We didn't own it at that point. The shares sold off. We bought at 51, we sold at 55, and now they're back at 54.96. As you recall, closer to 50, we said we would buy shares and then sell them again, because we do think that there's a chance there could be another offer." The valuation floor is the reason the trade is repeatable: "the board had rejected the initial offer of 60 to 50 a share, which was 53 billion, as inadequate, signaling a price target closer to 70. We think that without a bid, shares should trade around 53, so there's not a lot of downside versus upside if there's a new bid. However, there is no proof of any new bid." The bidder map, with each candidate's obstacle named: PE consortiums (KKR, Blackstone, Silver Lake) — "interest rates are so high that doing such a large LBO would be quite difficult"; big tech (Alphabet, Amazon, Google) — "they're spending so much money on AI that this is likely not going to be a deal this year," though PayPal brings "400 million active payment accounts"; and the networks/banks (Visa, Mastercard, JPMorgan), "which have suffered a little bit on valuation due to AI… but we don't see any evidence of that." Cheap versus Adyen or Stripe; if no deal comes, the fallback is self-help — "operational restructuring, layoffs, larger share buybacks, some asset sales, carving out something like Venmo or Braintree to sell."	06:08
WY	Weyerhaeuser	QT · SA · STK · FA	Positive	A third idea, deliberately floated with less conviction than the other two. "We think Weyerhaeuser could also be a contrarian REIT pick. I'm not as bullish on this one because the stock will not likely move too much, but it looks like it's bottoming." The frame: it "operates at the intersection of a cyclical trough — depressed single family home building — and structural undervaluation, trading at a significant discount to its underlying private land values. Unlike traditional equity REITs which collect recurring office or apartment rents, Weyerhaeuser is a timberland REIT, meaning its earnings function as a call option on housing activity, single family construction, and commodity lumber pricing." The assets: 10.5 million acres of US timberlands plus a wood-products arm making lumber, OSB and engineered wood. The leverage: "just a $10 MBF change in lumber prices adds about 50 million in annual EBITDA," and "a $10 MSF change in oriented strand board adds about 30 million" — and "single family home construction requires roughly three times more lumber per unit than multi-family apartments," so a rate-driven return of single-family starts expands margin upstream and downstream at once. The valuation: "Weyerhaeuser trades at a 30-40% discount to its private land NAV. Investors who buy ahead of a housing market turnaround effectively buy millions of acres of prime real estate at a steep markdown relative to private transactions." The self-help: strategic land solutions EBITDA guidance raised to $450 million from "selling and leasing acreage for solar and wind development, carbon capture and higher-and-better-use real estate sales." And the payout mechanism: a 3.5% quarterly base dividend plus a variable dividend targeting 75-80% of annual free cash flow — "so the dividend could go up dramatically by this formula if the EBITDA were to inflect higher… you could see massive special cash payouts in a housing bull market." Timing, in his own words: "if you were to see a cyclical change in housing demand after interest rates peak, going into 2027 — so we're early on Weyerhaeuser." Note the crosscurrent from his own macro section: mortgage rates "could easily go above 7%" per First American's Mark Fleming, and private construction outside data centres is declining.	1:01:07
CVX	Chevron	QT · SA · STK · FA	Positive	The Venezuela franchise, stated as a durable competitive position rather than a trade. "Chevron is finalizing a deal that will significantly expand its Venezuela footprint. It's adding two giant oil fields in the Orinoco belt to its operations. It has a first mover advantage. And we think Chevron will continue to be the biggest player in Venezuela going forward." The policy backdrop he reads alongside it: the White House published the terms of the US-Venezuela oil deal, under which "the US government received a 35% equity stake at zero cost to taxpayers with the right to purchase 20% of all current and future production at cost, and a first right of refusal on the remaining 80%" — a state-backed structure that favours the incumbent operator already on the ground.	1:15:52
VST	Vistra	QT · SA · STK · FA	Positive	An insider buy flagged from the new Form 4 tracker, in the sector he keeps naming as the multi-year opportunity. "VST CEO actually bought shares last week. So this was an interesting insider buy… a significant purchase for about a million. So not a small purchase. President and CEO James Burke bought 270k of VST, 2,000 shares at 135. It's his eighth largest trade out of 13." It sits directly under the thesis he restates later from Musk's G20 remarks: "AI growth is now being constrained by power… I think that power will be the biggest opportunity over the next few years. There's a potential 15 gigawatt shortfall just by next year alone, with AI chip production growing 40, 50% annually, while power outside China is only growing at 10 to 20%."	49:30
MU	Micron Technology	QT · SA · STK · FA	Positive	The memory bottleneck, tightening on two sides at once. Supply: "China's CXMT reportedly stalled small-scale production of HBM3E, the advanced memory used in AI processors, so that could be positive for Micron and SK Hynix and Samsung," while "Micron accelerates its HBM growth by about 60k wafers per month." Pricing: "Samsung reportedly allocated 70% of its memory production capacity through 2031 to long-term contracts, similar to SK Hynix and Micron, which was the first one to do it," with contract customers now including Nvidia, Microsoft and Google — and spot running far above contract: "HBM3E 36 gigabyte spot is now 2100, roughly four to five times LTA pricing. HBM4 16 stack spot is at 3500. DRAM export unit prices are up 36.6% from May to July, even as unit volumes fell 13%." Samsung "is considering converting its Pyeongtaek S5 foundry line to memory production as early as next year because of demand, which is a positive sign."	48:49 · 1:18:32
DELL	Dell Technologies	QT · SA · STK · FA	Positive	The print and the call, both read as bullish — and the call is used as a component-shortage datapoint for the whole complex. "Dell delivered a massive beat and raise with record revenues of 47 billion versus estimates of only 44.9 billion, a 58% year over year. EPS was $7 a share versus 4.9 expected. So massive, massive beat for Dell." And separately: "if you listen to Dell's earnings call on memory and optical shortages, it was very, very bullish. Basically, they said the constraints remain the same across DRAM, NAND, spotty CPU shortages, disk drives, power integrated circuits, microcontrollers, drivers, ABF substrate, T-glass and optical." (HPE is used as the corroborating second print: Q3 revenue $12.2B vs $11.9B, "beating like Dell and up 34% year-over-year with EPS of 1.11 versus 0.93 expected.")	48:22 · 1:18:32
BIDU	Baidu	QT · SA · STK · FA	Positive	A mechanical-flow catalyst with a named size. "Baidu's stock connect — the Shanghai Stock Exchange announced that Baidu is included in the Stock Connect, which will be effective next trading day after HK market close. Baidu's CFO, Henry He, estimated that there will be an additional 10 to 15% of fund flows into Baidu's stock. It could be an interesting one to watch." It sits alongside his running observation that "Chinese open-weight models are taking over the market" — roughly 70% of OpenRouter share versus US LLMs, with the usage-weighted token price falling below $1 per million for the first time from $2.04 at the end of May.	1:18:32
TTWO	Take-Two Interactive	QT · SA · STK · FA	Positive	An overreaction call on a dated catalyst. "Take-Two fell 7% as GTA leaks continued ahead of the November 19th release. We think that GTA will still be released over the next couple months and it was a market over reaction." The event is a scheduled, contractual one — a leak does not move a ship date that the company has already committed publicly — so the drawdown is treated as sentiment, not information.	1:18:32
GLD	SPDR Gold Shares	QT · SA · STK	Positive	Long-term conviction, short-term hedge — the most explicit two-horizon answer of the call, given in Q&A. "Gold is obviously seeing some pressure with rates rising. Gold is inversely correlated with real rates. What we're seeing is gold slightly weak, down about 18 bips, but silver is higher by about 50 bips. Now, long term, I still think you're going to see a sharp depreciation of all fiat currency and reserves going towards gold versus Treasuries over time. However, in the near term, if we see a hike, you'll see some temporary pressure before the uptrend resumes. So, if you want to trim some gold, or sell covered calls on your gold, it would make sense. But I would use this as a buying opportunity for adding to gold going into 2027." The structural case is the deficit arithmetic: "the debasement trade will still be the single biggest trade of coming years, driven by giant out-of-control deficits all around the developed world… we have 18 trillion of OECD debt issuance this year with 4 trillion net new issuance. It's really unbelievable. Two-thirds of that is the US alone." The near-term offset: "the dollar could strengthen a little bit… because of a hike, and then subsequently I think we'll resume the downtrend in all fiat currency." Sentiment tell: "Modi is trying to curb gold buying in India, but that I think will just push people to want to buy gold even more as the central bank is buying gold."	24:42 · 1:33:12
VRNS	Varonis Systems	QT · SA · STK · FA	Neutral	A live take-private he grades as mostly played out. "There's a cybersecurity company called Varonis Systems… it rallied more than 10% intraday upon a takeout bid from a private equity firm called Thoma Bravo" — per the deck, Thoma Bravo-backed Proofpoint is in advanced talks, with a transaction possibly "announced in the coming weeks." "That spike took shares back to the highs of the year around 46, $47. No deal price was actually mentioned, but the reporting does seem to suggest a deal is likely at this point. At the highest they reached about 48 and then settled on the 4th of September around 46. From the unaffected price of 41, there could be some upside to $50 a share for this bid. The reporting comes 2.5 months after initial reports that VRNS was exploring a strategic sale." The honest mark: "from the lows, this was trading at 20 back in March, but from the time rumors started, this was trading at 30. So there's probably not much upside from here, but there could be a little bit of upside from that 46 price."	06:08
ATAI	ATAI Life Sciences	QT · SA · STK · FA	Neutral	An arb reaching its terminal event, kept in the book only as a calendar item. "ATAI is set to close on Wednesday." The deck adds the mechanics from the NASDAQ trader alert: the merger is tentatively scheduled to close before the market open on September 9, 2026, with the stock halted immediately following the after-hours session at around 7:50pm on September 8. Nothing left to underwrite — the spread is the last few days of carry.	06:08
WBD	Warner Bros. Discovery	QT · SA · STK · FA	Neutral	The flagship arb has now compressed out of the top spot — which is itself the update. "The largest spread right now, given the compression in Warner Brothers-Paramount, which would have been long, is NSC UNP." Two weeks ago WBD was the widest spread in the book at 20% narrowing to 7%; it has now tightened enough that the rail deal — where the deck flags "fundamental headwinds and growing republican opposition to the deal" — carries the widest remaining spread on the screen. The related item still open on the calendar is the WSJ's report of "industry skepticism around Paramount's commitment to distribute 30 films per year."	06:08
NVDA	NVIDIA	QT · SA · STK · FA	Neutral	Cheap on the guide, uncomfortable on the buyer. The fundamental line is constructive: "Nvidia is guiding to 70% growth. AVGO is guiding to 100% growth. Both are supply constrained and saying real demand is higher, but the market is not giving them credit," and it "has agreed to acquire Hugging Face for 12.93 billion, giving it one of the largest open AI development platforms in the world." The positioning line is the warning: with earnings season over and systematic exposure rebuilt, retail is the incremental buyer — "retail investors bought 250 million of Nvidia shares on Wednesday, their third largest daily purchase since mid-May. The retail buying streak has gone to 15 consecutive trading sessions with investors purchasing more than 2.5 billion in Nvidia stock over this period, which is bananas." Sitting under both is the JPMorgan/Citadel advice to reduce risk for three weeks.	1:18:32
TEM	Tempus AI	QT · SA · STK · FA	Neutral	Flagged twice in one call from opposite directions, and explicitly left unresolved. On the sell-side: "Cantor Fitzgerald initiated Tempus, TEM, overweight with an $80 price target. I have to look at that." On the insider tape from his own new Form 4 tracker, Tempus appears among the week's largest sales — "the biggest sales were Lumentum, CoreWeave CEO selling 27 million, Tempus AI, Freshpet, and TeraWulf." No view taken; the name goes on the work list, which is exactly the state the insider screener is designed to surface ("we're working on prioritizing which ones are good and which ones are bad. That screener is supposed to be out in October").	50:10 · 1:18:32
Anthropic	Anthropic (private)	—	Neutral	The private mark that the whole AI credit chain is discounted off, and it is going the wrong way. "Anthropic's valuation has actually come down — estimates from about 1.4 trillion to about 900 billion. They've delayed the IPO again to mid-October. This is as Chinese models gain share." It matters in three places on this call: it is one of the four named customers inside Broadcom's 20-gigawatt, through-2028 commitment; it is the counterparty on a reported $35 billion cloud deal with Lambda, an Nvidia-backed cloud provider; and it is the source of the unrealised marks JPMorgan Asset Management flags as "an unusually high share of S&P 500 earnings due to unrealized valuation gains" — though on his own recomputation, "when you take away the other income or valuation gains, S&P 500 earnings is still very strong at above 30%." (Product news on the same page: Anthropic "launched Claude Fable 5.1 and Mythos 5.1, with Fable 5.1 positioned as the most advanced model yet for coding.")	1:15:21 · 1:18:32
LYB	LyondellBasell Industries	QT · SA · STK · FA	Negative	An explicit exit call on a prior long, asked and answered directly in Q&A. "LyondellBasell — we were long a while back. On the chemical front, we were effectively timing a cyclical bottom. Since then, chemical names rallied quite a bit. And because of rising energy prices, I am not really that positive on the sector anymore. So what I would do for LYB is just trim or sell. I'm just not as positive on chems as I was." The reasoning is the war, not the company: "I think the industrial economy and just with the energy price issue, I'm not as bullish. I actually think that we could see an escalation of this Iran war before it gets better after the midterms. There's no immediate catalyst for peace. So names that are sensitive to oil prices and gas prices, I'm just not as excited about." The cyclical-bottom trade worked; the input-cost regime that would carry it further has reversed.	1:31:39
CAT	Caterpillar	QT · SA · STK · FA	Negative	A short candidate supported by the single cleanest insider signal there is — the CEO halving his own stake. "Caterpillar CEO Creed sold 48% of his CAT stock for 46 million" (the deck's agenda line prices the same sale at $26 million). "For people looking at shorting CAT stock as a high valuation cyclical name, the fact that the CEO is selling half his shares might mean the company might have peaked in the short term." The macro fits the same direction on this call: private construction outside data centres is declining, mortgage rates could go above 7%, and BofA flags rising commodity costs against limited pricing power.	1:15:52

The mandate for the week: buy events, not beta — 00:00

The framing sentence of the entire call: "we are off to an interesting start for the month of September with CPI and PPI coming this week, with volatility in the bond markets, with escalating tensions in the Middle East. But despite that, what we are trying to do is add more event-driven names to the book that are less volatile when it comes to overall markets."
The two candidates, introduced in one breath: "EVC, which is an economically sensitive name but trades at a very large discount to what we think is its fair value. HIMX, which is another very cheap name to play the AI tailwind — and frankly, you can hedge with any of the AI ETFs to take out the beta from that name."
The rest of the agenda, pre-announced: "the top merger arbitrage names and spinoffs… S&P is going to spin off Capital IQ… we'll talk about PayPal and the Stripe bid, the rumor that came and went."

The calendar is the whole week: PPI Wednesday, CPI Friday, and a war premium inside both — 01:30

PPI is expected to jump from 0% in July to 0.4% in August "because of the resurgence in the war"; ex food and energy from 0.2% to 0.3%; headline y/y from 4.7% to 5.2%. But the clean core is expected to fall: "PPI ex food, energy and trade is actually supposed to fall. So a lot of this is just due to the war."
CPI — "which is even more important" — from 0.1% to 0.4% m/m, with core flat at 0.2% because it's ex-energy and headline y/y actually falling from 2.5% to 2.4% "because of the real estate drag. As you know, CPI has a much bigger real estate weight than PCE, and obviously than PPI."
Bookends: NFIB Tuesday (99.8 → 99.3), UMich Friday with one-year inflation expectations jumping 4% → 4.2% and 5-10 year flat at 3.3%. Earnings are a non-event: "a very, very, very light week," with only Oracle and Adobe Thursday after the close.

Jefferies initiates nuclear — previewed on the call, published next weekend — 01:30

Flagged off the back of Cembalest's Eye on the Market, which covers "the nuclear SMR renaissance" alongside the war tracker, S&P estimate evolution, the Treasury intervention, hyperscaler debt issuance and healthcare "pricing in a lot of bad news."
"Jefferies initiated on nuclear this week with a couple recommendations we'll share with you this coming weekend… just to give you a little bit of a preview, they initiated on BWXT, CCO, DML in Canada, EU, MIR, NXC, SOLS, and KAP. They're effectively long those types of assets, like Cameco, Kazatomprom, and some of the SMR names." (No house stance given — the write-up is promised, not delivered.)

Special situations: VRNS, SPGI's Capital IQ spin, ATAI's close, and the antitrust calendar — 06:08

VRNS: "+10% intraday upon a takeout bid from Thoma Bravo… no deal price was actually mentioned, but the reporting does seem to suggest a deal is likely." From the unaffected 41 there is a path to $50, but at 46 "there's probably not much upside from here."
SPGI: the Capital IQ spin "is obviously positive because Moody's trades at a higher multiple than SPGI," with FactSet (15× P/E, 12× EBITDA) the comp and SPGI at 22× forward versus MSCI 26× and MCO 27×.
The calendar: ATAI closes Wednesday; antitrust expirations at LCII (Friday), ROKU (9/8), LNTH (9/17), TECH (9/18), KVUE (9/29) after South African approval; Elliott has a sizable Deutsche Telekom stake pressing it to drop a T-Mobile merger for buybacks; and with WBD compressed, "the largest spread right now… is NSC UNP."

LPSN: the two-day layup, alerted on the Discord — 06:08

"There's also a microcap spread which we talked about in the Discord, which we thought was an easy layup, which was LPSN. As you recall, we bought it in the 280s. It rallied up to 316 and it looks like the deal has now closed. So that was a very quick, almost like a two-day trade and now I think the shares were halted, but it looks like the deal is done."
Worth noting where it lived: not on the weekly call, but in the real-time alerts channel — which is the part of the product that carries the shortest-fuse trades.

PayPal: a rumour, a range, and a named bidder map — 06:08

The setup: a fresh Betaville report of a possible new bid, after a 40% rally from the summer unaffected price and then a sell-off when the Stripe talks reportedly failed. "We bought at 51, we sold at 55, and now they're back at 54.96."
The bounded case: the board "rejected the initial offer of 60 to 50 a share, which was 53 billion, as inadequate, signaling a price target closer to 70. We think that without a bid, shares should trade around 53, so there's not a lot of downside versus upside if there's a new bid. However, there is no proof of any new bid."
Every candidate comes with its own blocker: PE (KKR/Blackstone/Silver Lake) — rates make a $50bn+ LBO hard; big tech (Alphabet/Amazon) — "spending so much money on AI that this is likely not going to be a deal this year," though 400 million active accounts would be powerful; and the networks/banks (V/MA/JPM) — "this would be a great acquisition… but we don't see any evidence of that." If nothing comes: restructuring, layoffs, buybacks, or carving out Venmo or Braintree.

The Goldman macro summary: the trend is intact, the next three weeks are not — 13:07

Memorial Day to Labor Day: S&P cap-weighted +4%, equal-weight +7%. But "the TMT momentum pair just printed its worst two months on record" and TMT momentum is −50% from the June highs.
"Going into the September FOMC, the market's now pricing in a 52% probability, at least on Kalshi, of a hike, which we think should be less given the midterm election cycle — but because of inflation, the kinetic warfare, energy prices skyrocketing, especially European gas, there might be a reason to hike." And "September tends to be the single worst month of the year for US equities and midterm elections are particularly iffy… J.P. Morgan and Citadel have said reduce risk."
Two positioning facts that shape the rest of the call: private credit "is still seeing redemptions, but the second quarter actually saw 99 billion of fund raising for the strategy, which is quite alarming"; and software has outperformed the Nasdaq for five straight weeks on "the counter-trend rally and short covering… the best streak in the post-COVID era. Today it's only 12% from the highs. Probably time to take some risk off in software."

The war moves to the oil fleet: three tankers for two warships — 16:31 · 21:31

"Iran fired ballistic missiles toward two US Navy warships… the Pentagon responded by permanently disabling two Iranian oil tankers and completely destroying a third" — the MT Downey off Kharg Island, the Stark 1 near Jask, and the unladen Kylo destroyed in the Gulf of Oman — vessels CENTCOM calls "part of a multi-billion dollar shadow network that funds the IRGC."
The doctrinal shift: "moving beyond direct military infrastructure such as radar air defense sites, the US strategy is now targeting the Iranian oil economy," with commander Brad Cooper's stated intent to impose "a direct three-for-two economic cost while signaling a willingness to systematically destroy Iran's remaining oil fleet if attacks on US vessels continue."
The tape: WTI 92.64, Brent 97.31 — "almost at 100" — the 10-year at 4.8% and the 30-year at 5.25%. "That's not setting up well for the following week."

A dated stalemate: nothing changes until October 27 and November 3 — 16:31

The two-sided read: the WSJ says "the US is winning the war for control of Hormuz… the blockade has placed a complete chokehold on Iran's oil industry," while the NYT reports "Iran has grown more confident in its ability to confront Washington, plans to extend the conflict for months, and could pursue a significant escalation."
"It's hard to see the situation changing much in the Middle East, at least after the two major political events — Israeli elections October 27th and US midterm elections November 3rd. Until then, the best case scenario will be the kinetic activity stays fairly benign… but the risk is after midterms, escalation could be quite severe."
Why it is stuck: "the US lacking the ability because of severe munition and interceptor shortages, and the will due to the approaching midterms and extreme reluctance to deploy troops, to bring the war to conclusion, while Tehran is as hardline and stubborn as ever. So you're going to see these small attacks until midterms and then likely things could escalate after that." One more restraint noted: "if one of the navy warships was actually struck, the US would have done a lot more than hit three Iranian tankers."

Trump threatens the Fed with tariffs — and it would force the opposite — 24:42

"Trump threatens now to stop trade with half of our trading partners in the world if the Fed doesn't cut interest rates. In a bizarre tweet on September 4th, Trump threatened the Fed, tried to blackmail the Fed, saying that we're going to charge new tariffs. Why would the Fed care? In fact, if we do charge new tariffs, the Fed would actually have to hike, so it's very counter-intuitive."
The payrolls that moved hike odds were weaker than the headline: "of the 162,000 jobs, like 70% were leisure, healthcare, and government" — the deck breaks it out as leisure/hospitality 62,000 (38%), government 35,000 (22%), healthcare/social assistance 28,000 (17.5%) — "so not as strong as the numbers seemed, which is why the market didn't sell off more."

The bank split on September 16: UBS and Barclays hawkish, Morgan Stanley on hold, Singh below all of them — 24:42 · 42:28 · 42:56

Kalshi: 52% hike vs 48% no change. "We think that probability should be lower for a hike, but given where the war is, anything is possible."
UBS (Jonathan Pringle) wants 25bp on September 16 to 3.75-4% and again December 9 to 4-4.25%, switching from unchanged because of Warsh's Jackson Hole speech and the payrolls print. Barclays sees two hikes. Morgan Stanley expects a hold with core inflation falling to 3.1%. Singh: "the most important data point before the 16th meeting is August CPI. I'm not going to draw a conclusion until then, but I'm less hawkish than UBS obviously."
On Warsh's motives: "I feel like Warsh is just playing this game because there are 12 voting members and he doesn't want to look like he's clashing against them. Bessent is effectively doing his job by buying bonds in the market, but it's still not a guarantee that Warsh is going to hike."
The path matters more than the level: "if there's just one hike, the S&P should do reasonably well after a couple weeks of uncertainty. But if there's more than one hike, well, then things can get a lot worse in risk assets."

The 40-year bond bull market is over — and "return to normal" is the wrong frame — 30:26

The argument against the consensus comfort: yields "are back to levels we haven't seen since the early 2000s, but we can't call those 2000 level yields normal given 2000's yields were well below the yields of the 1990s and certainly well below the 1980s. So what is normal?" Each half-decade over 45 years "delivered lower yields than the prior period — until now."
The call: the 10-year fell "from an apex of 15% in 1981 to a nadir of 0.5% in 2020… that 2020 low marked the bottom and the end of the 40-year bond bull market," made official by "the breakout of the prior downtrend in 2022. The more important comparison is that recently we have experienced an uptrend in yields versus a downtrend, and that will continue — or yields will at least continue to stay relatively high."
Why the trend mattered: from 1981 to 2020, using seven-year rolling windows, "yields were lower at the end of the seven years than at the beginning 90% of the time," so every refinancing lowered the cost of capital — which "allowed for many mega trends: the proliferation of financial engineering, a huge boom in private equity, a general rise in multiples for public stocks, a persistent expansion of corporate profitability as interest expense fell… and an era of minimal consequence for governments to keep increasing public debt, which is one of the reasons our debt is 40 trillion today."

The consequences are already visible: $1.2T of debt service, a PE distribution drought, and cash with nowhere to go but AI bonds — 30:26

"The cost to service public debt in the US has exploded higher to an annualized 1.2 trillion, with interest expense as a percentage of GDP rising to levels not seen since the 1990s."
"There's been a distribution drought for the past four years where private equity firms are not able to return capital to their investors at the same pace they were five years ago… and this is also why private credit is struggling," with global PE deal count "falling dramatically."
The subtlest point of the call: "this is one of the reasons why the public bond market is so tight. There's roughly 3 to 400 billion of interest that's spewed off these deals, but there are no new LBOs, and there's no new supply. So that cash flow is being reinvested — and it's one of the reasons why this cash flow can be reinvested in all these new AI bonds, because there's nowhere else to put it."
The corporate side is only now turning: net interest expense actually fell as the Fed hiked in 2022 because borrowers had termed out and cash balances paid out ("we call that the boomer stimulus") — "we expect these aggregate borrowing costs to begin to creep higher as debt from the 2010s and early 2020s needs to be refinanced at higher rates and corporates go on a record borrowing spree to fund surging capex. We've seen the borrowing costs for companies like Oracle go up, and for the Mag 7 to start to go up as well."

Japan breaks the anchor: a record modern 30-year JGB and a carry-trade unwind — 37:11

"The spike in Japan's 30-year yield or JGB above 4.18%, alongside the 10-year approaching 3% for the first time since 1996, marked a structural break in global fixed income" — effectively an all-time record for the modern 30-year JGB, since the MoF only began regular issuance in 1999.
Drivers: the global debt and energy sell-off, BoJ tightening on local inflation, and record fiscal expansion under new PM Sanae Takaichi's semiconductor/AI industrial agenda — against national debt "exceeding 200% of GDP, double that of the US. As long as the Bank of Japan is not buying bonds, who else is buying bonds at 200% of GDP?"
The transmission: "as ultra-long domestic JGBs now offer a real yield of 4.2% on 30 years, the structural math favoring the carry trade collapses" — capital repatriation, a steepening long end, "a massive rally in Japanese banks," and Japan selling a record $88 billion of foreign securities in August, much of it Treasuries, to support the yen.

Norway, France, and the ceiling on how far this can go — 37:11 · 43:18 · 43:54

Norway's $2.3 trillion sovereign wealth fund is "cutting all government bond holdings… they could sell about 80 billion of US Treasuries, which is about 10% of our TGA." Alongside China and Poland "moving their relative reserves from Treasuries into higher yielding assets and gold."
France's 30-year hit 4.9274%, the highest since September 2008: a 5% budget deficit ("the US is well above 6%, but because France isn't growing, it's a bigger issue for them"), political gridlock, and a stagflationary heat-wave and wildfire shock costing above €15 billion, about half a percent of GDP.
The ceiling: "bonds in the US are effectively reasonable here. If you look at the 10-year yield versus three-year average nominal GDP… we don't expect yields to go much higher from here without inflation going higher from here." And the scale on which to watch it: net interest is "almost 19%, 20% of US tax revenue. Once you get above 30%, I think people will start panicking — that will be in the next few years. And once you get to 50%, you have to really depreciate the currency dramatically. That's when you start to see revolutions." Truflation, "the closest real-time inflation indicator we have," is at the highest point of the year.

Citadel and JPMorgan: three weeks of caution, with retail as the only bid — 44:53

"While they remain positive or constructive on US equities longer term, in the short term they are quite concerned over the next several weeks." The mechanics: "given earnings season is over, retail is the incremental buyer, systematic exposure has been rebuilt — so quants and momentum funds may not be buying any further. Upside catalysts are becoming less obvious. There's more AI risk as time goes by."
Dated: "Scott Rubner of Citadel is more constructive by mid-October. JP Morgan is also cautious for three weeks." The retail evidence arrives later on the call: $2.5 billion of NVDA bought over 15 consecutive sessions.

Credit: CCCs break out, and the honest sizing of it — 45:46

"Triple-C US corporate credit spreads have been breaking out for the last four months… around 13% plus yields [JPMorgan index], BofA's a little higher, about 15%."
Then the counterweight, unusually candid for a bearish datapoint he raised himself: "remember that's only 10% of high yield. So if high yield's 1.5 trillion, this is 150 billion. And if you see defaults even at 10% of 150 billion, you're only going to get 15 billion, which is nothing… high yield defaults meant a lot more in 2008 versus now, because the equity market's just so much bigger than the high yield debt market."
The consumer datapoint that does worry him: "60-day delinquency on US subprime auto loans has climbed to roughly 5.2%, the highest level on record and more than double where it was four years ago. So that's concerning on the consumer."

Private credit gates are back — across seven managers at once — 1:14:07

"Blackstone disclosed that its flagship $77.2 billion credit fund, BCRED, enforced a 5% quarterly limit again after receiving 4 billion of withdrawal requests, or 10%. Cliffwater also put in a 5% cap after it got another 16% of redemptions, 1.8 billion. Blue Owl capped its redemptions at 5% after it got 20% redemptions, and BlackRock and Morgan Stanley both triggered 5% quarterly tender caps after withdrawal requests hit 9.3% and 10.9%. Ares and Apollo both have also had to apply liquidity gates."
The conclusion: "so the private credit risk has not gone" — and it connects back to the bond-bull-market section, where the four-year PE distribution drought is named as the underlying cause of the illiquidity.

Commodities: a super-cycle, copper at a new high, diesel cracks at a record $106 — 46:32 · 1:16:51 · 1:18:32

"BofA says rising commodity costs and limited pricing power are increasing margin pressure on staples companies. You're starting to see a super cycle in commodities — oil, diesel, agricultural prices, even copper hitting a local high today."
Copper's driver list: "US tariff anticipation, regional stock price squeeze, higher energy prices, the LME inventory drain, falling output, declining ore grades, data center hype growth because of the massive copper wire required, grid resiliency requirements, and just macro portfolio rotations."
Diesel is the sharpest: "US diesel refining margins have surged to a record $106 a barrel, topping $100 for the first time in history and surpassing the roughly $85 peak seen during the 2022 energy crisis, which is very good for refiners right now… reflecting a severe shortage of global diesel supply amid the Iran war, with US diesel inventories down 103 million barrels, the lowest on record for this time of year. Retail diesel prices have climbed to $5.78 a gallon, just four cents below the all-time high of June 2022."

Housing and the rate-sensitive consumer — 46:32 · 47:04

"Mortgage rates are closely tied to the 10-year treasury yield, so private construction outside of data centers is also declining… data center construction is really the only thing accelerating higher."
"Mortgage rates could easily go above 7% according to Mark Fleming at First American. This would put further pressure on the US housing market. Americans also buy lots of stuff to outfit new homes like furniture and appliances, so this should negatively impact Restoration Hardware, Home Depot, Lowe's, etc." (Read against the WY thesis, this is the reason he calls himself "early.")

Chinese open-weight models and the collapse in token prices — 47:28

"According to OpenRouter, Chinese open-source models are roughly 70% of the market versus US LLMs." SiliconData's usage-weighted token-expenditure index "was less than a dollar per million tokens for the first time ever… versus 2.04 at the end of May. So in three months, token costs are down 50% due to US companies weighing these Chinese models."
The caveat he supplies himself: "OpenRouter over-indexes developers who actively switch between models and over-indexes smaller developers. So that could be one of the reasons the index has fallen so sharply." The knock-on is the Anthropic mark: the IPO slips to mid-October and the valuation estimate drops $1.4T → $900B "as Chinese models gain share."

The insider tape, and a Form 4 screener shipping in October — 49:30 · 50:10

"We started to track in our tracker Form 4s and we're working on prioritizing which ones are good and which ones are bad. So that screener is supposed to be out in October."
Top buys: Insperity (NSP) chairman/CEO $3M "as a buy the dip trade"; Lumen's Kate Johnson $500k; FTAI's Wesley Edens $300k "where the stock is down quite a bit"; an Okta director "flagged as an unusual 10B5-1 buy signal"; Kratos' DeMarco $250k; Intel's CFO $250k "as Intel has sold off." Plus Vistra's CEO at $270k.
Biggest sales: Lumentum, CoreWeave's CEO ($27M), Tempus AI, Freshpet and TeraWulf — and, separately, Caterpillar's CEO selling 48% of his stock.

Defense: seven-year THAAD and PAC-3 frameworks, and a backlog that doubled in a quarter — 52:39

"You could see how aggressively we are spending on Terminal High Altitude Area Defense missiles, THAAD, and Patriot advanced capability interceptors, the PAC-3 MSEs. There's a $35 billion contract… in Patriot missiles, 58 billion for Lockheed Martin, and General Dynamics also has a seven-year framework deal."
The number that quantifies it: "just for Lockheed Martin alone, the missile and fire control backlog has doubled in the last one quarter." It also ties back to the war analysis — the US is constrained by "severe munition and interceptor shortages," and "there've been… a $30 billion contract just signed last week" to fix that.

Entravision: the political-ad windfall hiding inside an ad-tech re-rating — 53:55

The mechanical catalyst: EVC "is quite an interesting name going into midterm elections because its advertising spend should explode given its key Hispanic demographic in specific states that are the most important in the US when it comes to ad spending during the midterms."
The re-rating candidate: ATS revenue +230% to $182M and operating profit +673% to $40M in Q2, "unrivaled access to the US Hispanic demographic… the fastest growing demographic and voting bloc," $83M of cash and marketable securities, and a model that blends recurring digital with retransmission-consent fees and cyclical political windfalls.
The reason it is available: management guided ATS revenue down sequentially for Q3 on "normal customer spending timing" and "this pause in EBITDA growth resulted in a big sell off in the stock," taking consensus to $18-25M/quarter. His mark: $100-115M of 2026 adjusted EBITDA, 7.5-8.5× forward against a 12-12.4× trailing — a discount to ad-tech at 12-15× and a premium to legacy broadcasters at 4-6×. Risks: the TelevisaUnivision contract expires at year end, a $3.3M legacy-media loss, and 85% growth in cloud costs and sales commissions.

Weyerhaeuser: a call option on housing, priced at a 30-40% discount to dirt — 1:01:07

"Unlike traditional equity REITs, which collect recurring office or apartment rents, Weyerhaeuser is a timberland REIT, meaning its earnings function as a call option on housing activity, single family construction, and commodity lumber pricing" — 10.5 million acres plus mills making lumber, OSB and engineered wood.
The operating leverage: "a $10 MBF change in lumber prices adds about 50 million in annual EBITDA"; a $10/MSF OSB move adds $30M; and "single family home construction requires roughly three times more lumber per unit than multi-family apartments."
The two structural kickers: it "trades at a 30-40% discount to its private land NAV," and land-solutions EBITDA guidance is raised to $450 million from solar, wind, carbon capture and higher-and-better-use land sales. The payout: a 3.5% base dividend plus a variable one targeting 75-80% of FCF — "you could see massive special cash payouts in a housing bull market." The caveat: "we're early on Weyerhaeuser."

Himax: 40% of automotive display drivers, and the sensor inside every Meta Ray-Ban — 1:06:46

What it does, in his own plain terms: "display drivers act basically as the brains for screens, translating signals from a computer or processor into instructions that tell individual pixels to illuminate and display colors" — sold as a tier-two fabless supplier to panel makers like AUO and BOE, then into Apple, Dell, Samsung and the automakers.
The share: "Himax is the global market share leader in automotive display driver integrated circuits… Himax owns 40% of the global market share for automotive displays." The optionality: WiseEye always-on edge vision, LCoS micro-displays and wafer-level optics — "all of the Meta Ray-Bans use Himax."
The mispricing: Q2 revenue $227M (+14% q/q) with 33.1% gross margin exceeding guidance, snapping four quarters of declining growth; Q3 guided +7-11% to $243-250M (+23.8% y/y at the midpoint) with margin to 34% and flat inventories. "Its last quarter beat and its stock is down simply because of the AI sell off… the market doesn't seem to be rewarding this strong beat." Forward P/E ~30, "but the TAM in the smart glasses is so high… and auto has bottomed."

Midterms: the market now expects a sweep; he does not — 1:13:49

"Markets now expect the Democrats to sweep both the House and the Senate. I still think it will be only the House. Less than one year ago, Democrats only had a 20% chance of sweeping the midterms. It's unbelievable what's happened so quickly."
The same date drives the war forecast, the political-ad thesis inside EVC, and his read on why the Fed is politically constrained — a single calendar item load-bearing across three separate sections of the call.

Index rebalance, Caterpillar's CEO, and Chevron in the Orinoco — 1:15:01 · 1:15:52

S&P 500 adds: Bloom Energy ("which is why Bloom Energy was up so much after hours on Friday"), Illumina, and one name the source garbles. Deletions: The Trade Desk, Builders FirstSource and Molson Coors — "as beer drinking is down" — all three to the SmallCap 600, which is much less followed.
"Caterpillar CEO Creed sold 48% of his CAT stock. For people looking at shorting CAT as a high valuation cyclical name, the fact that the CEO is selling half his shares might mean the company might have peaked in the short term."
"Chevron is finalizing a deal that will significantly expand its Venezuela footprint, adding two giant oil fields in the Orinoco belt. It has a first mover advantage, and we think Chevron will continue to be the biggest player in Venezuela going forward."

Power is the constraint, not chips — 1:18:32

From Musk at the G20: "AI growth is now being constrained by power. I think that power will be the biggest opportunity over the next few years. There's a potential 15 gigawatt shortfall just by next year alone, with AI chip production growing 40, 50% annually, while power outside China is only growing at 10 to 20%."
Corroborating items in the same recap: Trump pushing back on municipalities blocking data centres; data-centre construction the only accelerating category of building; and Broadcom's guide anchored on enabling "over 20 gigawatts of compute capacity."

Gold, silver and the debasement trade, asked directly — 1:33:12

"Gold is inversely correlated with real rates… gold is slightly weak, down about 18 bips, but silver is higher by about 50 bips."
The two-horizon answer: "in the near term, if we see a hike, you'll see some temporary pressure before the uptrend resumes. So if you want to trim some gold, or sell covered calls on your gold, it would make sense. But I would use this as a buying opportunity for adding to gold going into 2027."
The structural case, sized: "the debasement trade will still be the single biggest trade of coming years, driven by giant out-of-control deficits all around the developed world… we have 18 trillion of OECD debt issuance this year with 4 trillion net new issuance. Two-thirds of that is the US alone." Sentiment confirmation from India: "Modi is trying to curb gold buying… but that I think will just push people to want to buy gold even more as the central bank is buying gold," with inbound shipments +32% in the first four months of the financial year.

LyondellBasell: the cyclical-bottom trade is closed — 1:31:39

The exit, given in plain terms: "we were timing a cyclical bottom. Since then, chemical names rallied quite a bit. And because of rising energy prices, I am not really that positive on the sector anymore. So what I would do for LYB is just trim or sell."

The generalisation matters more than the name: "I actually think that we could see an escalation of this Iran war before it gets better after the midterms. There's no immediate catalyst for peace. So names that are sensitive to oil prices and gas prices, I'm just not as excited about." A whole-sector de-rating driven by a geopolitical calendar rather than by company fundamentals.

Entravision started life as a Spanish-language television and radio broadcaster in the United States — the largest independent group of stations affiliated with TelevisaUnivision, the company behind Univision and UniMás. That old business is shrinking, and last quarter it lost $3.3 million.

Almost nobody has noticed that the company is now mostly something else. Its "advertising technology and services" arm sells the plumbing that places digital ads automatically — software that decides, in the fraction of a second before a web page loads, which advert you see and what the advertiser pays for it. That division grew revenue 230% last quarter to $182 million, and its operating profit rose 673% to $40 million. The company holds $83 million of cash and securities and carries no meaningful debt problem.

The reason you can buy it cheaply is a scheduling quirk. Management told investors that the ad-tech division's revenue would be lower in the third quarter than the second — not because business is worse, but because big clients happen to spend in lumps. Analysts cut their quarterly forecasts to $18-25 million and the shares fell hard. Singh's estimate for the full year is $100-115 million of adjusted earnings before interest, tax, depreciation and amortisation, which puts the stock at roughly seven and a half to eight and a half times that figure. Pure digital advertising companies trade above twelve to fifteen times; dying radio stations trade at four to six. Entravision is being priced much closer to the radio stations than to the software company it is becoming.

And then there is the free option. Every two years, American political campaigns spend enormous sums on television advertising, and the money is concentrated in the handful of states that decide elections. Entravision owns the Spanish-language stations in exactly those states, aimed at the fastest-growing group of voters in the country. That advertising is booked at very high margins and arrives in the second half of an election year — which is now. The main thing that could spoil it: the contract that makes Entravision a TelevisaUnivision affiliate expires at the end of this year.

Himax is a Taiwanese chip designer. It does not own factories — it designs the silicon and pays TSMC to manufacture it, a model called "fabless." Its chips are display drivers: the small processors that sit behind a screen and translate a computer's instructions into individual pixels lighting up in the right colour. They are in televisions, laptops, Kindles, phones — and, above all, in cars.

Cars are where the position is genuinely strong. As dashboards have replaced dials and needles with curved screens and head-up displays, Himax has ended up supplying roughly 40% of the world's automotive display driver chips — the largest share of anyone. That business went through four straight quarters of shrinking revenue as the car industry slumped, and has now turned: last quarter revenue rose 14% from the prior quarter to $227 million, profit margins beat guidance at 33.1%, and management guided the current quarter to $243-250 million, which would be growth of nearly 24% against a year ago, with margins improving again.

The share price fell anyway, because the whole artificial-intelligence complex sold off and Himax got sold with it. That, in Singh's telling, is the entire opportunity: "its last quarter beat and its stock is down simply because of the AI sell-off." He also names the way to hold it without holding the market's mood — buy Himax and short an AI-themed exchange-traded fund against it. If AI stocks fall as a group you make money on the short and lose it on Himax, leaving you exposed only to whether Himax itself does better than its peers. That is what "hedging out the beta" means.

The part that could turn a cheap car-parts business into something more is smart glasses. Himax makes an ultra-low-power vision sensor called WiseEye that lets a device watch its surroundings continuously without flattening the battery, plus the tiny projectors and optics that put an image in front of your eye. Every pair of Meta Ray-Ban glasses uses Himax. If those glasses become a mass-market product, a chip supplier with a design win in all of them is worth a great deal more than thirty times next year's earnings.

Weyerhaeuser owns 10.5 million acres of American forest and the mills that turn the trees into lumber, plywood-substitutes and engineered beams. It is structured as a real estate investment trust, but it is nothing like an office or apartment landlord collecting monthly rent. Its profits rise and fall with how many houses America is building, which makes owning the shares closer to owning a bet on a housing recovery than owning a property portfolio — Singh calls it "a call option on housing activity."

The reason a small change in the housing market produces a large change in profits is that the company both grows the wood and mills it. A $10 move per thousand board feet in the lumber price adds about $50 million of annual earnings; a similar move in oriented strand board adds about $30 million. And detached houses use roughly three times more wood per home than apartment blocks — so if mortgage rates fall and builders return to single-family homes specifically, the effect compounds.

Meanwhile the land itself is worth more than the stock market says. When timberland changes hands privately, buyers pay considerably more per acre than Weyerhaeuser's share price implies — a gap of 30-40%. So you are buying millions of acres of prime forest at a discount to what a private buyer would pay for it, and being paid to wait: the company pays a 3.5% base dividend and tops it up with a variable one designed to hand back 75-80% of its free cash flow, which means the payout would rise sharply if earnings inflected. It is also monetising the land in ways that do not depend on lumber at all — leasing acreage for solar and wind farms, carbon storage, and selling parcels for development, a business it has just guided to $450 million of earnings this year.

Singh is careful not to oversell it. "I'm not as bullish on this one because the stock will not likely move too much, but it looks like it's bottoming," and the trigger — mortgage rates falling — is not here yet; his own analysis on the same call says rates could go above 7%. He puts the inflection in 2027, and says plainly: "we're early on Weyerhaeuser."

Broadcom designs custom chips for the handful of companies building the world's artificial-intelligence data centres. Unlike Nvidia, which sells the same general-purpose graphics processor to everyone, Broadcom co-designs a bespoke chip for a single customer — which makes each relationship stickier and longer-dated.

The results were strong: about $30 billion of revenue against expectations of $29.25 billion, earnings of $3.32 a share against $3.21, and $13.7 billion of free cash flow, meaning 46 cents of every revenue dollar came out the other end as spendable cash. But the forecast is what matters. Management says its AI revenue will be $58 billion this financial year, roughly double to $115 billion next year, and double again to $230 billion by 2028 — underpinned by commitments to supply chips for more than 20 gigawatts of computing capacity for Google, Anthropic, Meta and OpenAI.

Singh's sum is simple. The chief executive is guiding to $30 of earnings per share by fiscal 2028. Put a fairly ordinary multiple of 20 on that and the stock roughly doubles — and that assumes growth slows, when the company is currently compounding at 50% a year. "This stock is basically pricing in that growth is peaking." He bought some, and was puzzled by the reaction: "I don't know why the market was not happier."

S&P Global is several businesses in one wrapper: it rates the creditworthiness of bonds, it owns the S&P 500 and the index-licensing business behind it, and it sells financial data through a platform called Capital IQ. The ratings and index arms are extraordinary businesses — near-monopolies with very high margins. The data business is good, but it competes with FactSet, Bloomberg and others, and the market pays less for it.

The company is now looking at spinning Capital IQ out as a separate listed business. Two things make that attractive. First, whatever is left behind is a purer version of the high-margin ratings-and-indices company, which the market should value more highly — as evidence, Moody's, which is essentially just a ratings agency, trades at 27 times earnings and S&P at 22. Second, the piece being separated gets its own price rather than being buried in a conglomerate: FactSet, the closest comparable, trades at about 15 times earnings and 12 times cash profits.

This is the classic special-situations trade Singh runs repeatedly: a conglomerate discount closing because a company chooses to break itself up. It follows the same company separating its Carfax and Mobility businesses only two months ago — a management team that has shown it will actually do this rather than merely discuss it.

LivePerson makes customer-service chat software. It agreed to be taken over, and once a takeover is agreed the shares stop trading on business prospects and start trading on one question: will the deal actually close, and when? The buyer's price is fixed, so the shares sit slightly below it — the gap being the market's price for the risk and the wait. Buying that gap is called merger arbitrage.

Here the gap was unusually wide for how safe the deal looked, because LivePerson is a tiny company. Large funds cannot buy a meaningful position in a stock trading near $3 without moving the price, so the spread stays wider than the risk deserves. That is precisely the kind of inefficiency an individual investor can take and an institution cannot — which is why Singh keeps a microcap sleeve at all.

The trade lasted days: bought in the $2.80s, out at $3.16 as the deal closed and the shares were halted — around 13% in what he calls "almost like a two-day trade." Note where it was communicated: not on this weekly call but in the real-time Discord alerts, because a spread this short-lived is gone by the time a weekly recording goes out.

PayPal processes online payments and owns Venmo. Earlier this year Stripe was reportedly interested in buying it; those talks fell apart, and now an M&A blog says another bid may be circulating. Singh has stopped treating it as a company to own and started treating it as a price range to trade — he bought at $51, sold at $55, and it is back at $54.96.

The reason a range works is that both ends of it are anchored to something real. The floor: without any takeover, he thinks the shares are worth about $53 on the business alone, so buying near $50 means little can go wrong. The ceiling: the board already rejected an offer in the $50-60 range, valuing the company at $53 billion, as too low, and signalled it wants something nearer $70. So the downside is a couple of dollars and the upside, if a bid appears, is many.

He is honest that no bid is confirmed, and he walks through who could plausibly make one and why each probably will not right now: private-equity firms would need to borrow $50 billion-plus at today's interest rates; Google and Amazon are spending everything they have on artificial intelligence; the card networks and big banks would find it a natural fit but show no sign of interest. What PayPal would bring any of them is 400 million active payment accounts. And if nobody bids, the company can still improve its own value — cutting costs, buying back shares, or selling Venmo or Braintree separately.

Varonis makes software that watches who inside a company is opening which files — a way to catch a breach or an insider leak. It has been looking for a buyer for two and a half months, and last week reports emerged that Proofpoint, a security company owned by the private-equity firm Thoma Bravo, is in advanced talks to acquire it. The shares jumped more than 10% in a day.

The reason Singh is lukewarm is arithmetic rather than doubt about the deal. Before any of this leaked, the stock was $41 — that is the "unaffected" price, the level that tells you what the business is worth without a buyer. It ran to $48 and settled near $46. If a bid arrives at $50, you make roughly four dollars from here. If the talks collapse, you fall back toward $41. That is a poor ratio, and it is a very different trade from the one available in March, when the stock was $20, or even at $30 when the rumours started.

The general lesson he is drawing, and the reason the name still appears on the call: the money in a takeover is made by owning the company while it is merely a candidate, not after the newspapers have named the buyer. "There's probably not much upside from here."

ATAI is a merger-arbitrage position that has run its course. The acquisition is scheduled to complete before the market opens on Wednesday 9 September, and the stock will stop trading after Tuesday's after-hours session, at around 7:50pm, so shareholders receive the deal consideration instead of a share price.

There is nothing left to analyse. Once a deal is days from closing, the remaining gap between the market price and the payout is just a few days' worth of interest — the return for holding cash that is about to be released. It appears on the call as a calendar item rather than an idea, which is exactly the right way to treat it.

Warner Bros. Discovery has been the flagship position in the merger-arbitrage book for months, bought when the gap between the market price and the agreed takeover price by Paramount Skydance was around 20%. That gap has been closing steadily as regulatory objections have been worked through, and this week it closed enough that Warner is no longer the widest spread on the screen.

The way Singh mentions it is itself the update: asked which deal now offers the most, he answers "the largest spread right now, given the compression in Warner Brothers-Paramount, which would have been long, is NSC UNP" — the contested Norfolk Southern / Union Pacific railroad merger. In arbitrage, a narrowing spread is a position doing its job: most of the return has already been collected and the remaining reward for holding it is small.

Still open in the background: the Wall Street Journal reports scepticism about Paramount's public promise to release 30 films a year, which is one of the commitments regulators and cinema chains have been weighing.

Chevron is one of the few Western oil companies that never fully left Venezuela, and it is now finalising a deal to add two very large fields in the Orinoco belt — one of the biggest heavy-oil deposits on earth — to what it already operates there.

Being first back matters more than usual here. Venezuela's oil industry needs partners with the equipment, the technical staff and the licences to operate, and those relationships take years to build. The company that is already on the ground when a country reopens tends to get the best acreage and the best terms, and everyone who arrives later negotiates against an incumbent. That is what Singh means by "first mover advantage," and why he expects Chevron to remain "the biggest player in Venezuela going forward."

The political structure around it is unusual and worth noting: under the deal the White House published, the US government took a 35% stake in the parent company at no cost to taxpayers, plus the right to buy a fifth of all production at cost and first refusal on the rest — an arrangement that ties American policy to keeping the fields producing, which indirectly protects the operator running them.

Vistra generates and sells electricity in the United States. It matters to this call for a reason that has nothing to do with utilities and everything to do with artificial intelligence: the constraint on building data centres is no longer chips, it is power. Singh puts a number on it from Elon Musk's remarks at the G20 — chip production is growing 40-50% a year while electricity supply outside China grows 10-20%, implying a shortfall of roughly 15 gigawatts as soon as next year.

The specific signal this week is an insider purchase. Chief executive James Burke bought 2,000 shares at $135, about $270,000 of stock, on the open market — his eighth-largest such purchase out of thirteen. Insider buying is a weaker signal than insider selling is when it is small or ceremonial, but an open-market purchase of this size by the person who sets the company's forecasts is real information: he is choosing to increase his own exposure at today's price.

It also sits inside the newest piece of SSR's tooling. Singh says the firm has started systematically tracking Form 4 filings — the disclosure executives must make when they trade their own company's stock — and is building a screener, due in October, to sort the meaningful purchases from the routine ones.

Micron is one of only three companies in the world that make the high-bandwidth memory chips that sit next to an AI processor and feed it data. Memory has historically been a brutal commodity business — everyone builds capacity at once, prices crash — and the argument here is that this cycle is behaving differently.

Two things happened this week. On supply, China's CXMT reportedly halted its small-scale attempt to make the same advanced memory, removing a potential fourth competitor. On pricing, Samsung has now locked roughly 70% of its production capacity through 2031 into long-term contracts, following SK Hynix and Micron, which did it first. When almost all the output is pre-sold years ahead, the leftover spot market becomes very tight — and it shows: the spot price of one advanced memory part is running four to five times the contract price.

Micron is meanwhile expanding hard, adding roughly 60,000 wafers a month of high-bandwidth memory capacity. The reason that does not break the price is that the demand for it is contracted rather than speculative. This is the "own the bottleneck" idea Singh returns to repeatedly: in a boom, the money accrues to whoever supplies the scarce input, not to whoever assembles the finished product.

Dell assembles and sells the servers that AI data centres are built from. Its quarter was, in Singh's words, "a massive, massive beat": $47 billion of revenue against the $44.9 billion expected, up 58% on a year earlier, with earnings of $7.00 a share against $4.90 forecast. Its rival HPE beat too, up 34%, which suggests the demand is real across the industry rather than one company taking share.

But he treats the earnings call as more valuable than the earnings. Because Dell buys components from everybody, its management is effectively a survey of the entire supply chain — and they described shortages across memory, flash storage, processors, hard drives, power-management chips, microcontrollers, the special substrate that chips are mounted on, the glass used in advanced packaging, and optical parts.

That list is bullish for anyone selling those components — which is why the same section of the call also supports the Micron argument — and it is a warning about how long it will take to build the capacity everyone has announced. When a shortage runs across nine different components simultaneously, the constraint is not one factory; it is the whole chain.

Baidu is China's dominant search engine and one of its main AI developers. The news here is plumbing rather than business: the Shanghai Stock Exchange has added Baidu to Stock Connect, the scheme that allows investors in mainland China to buy shares listed in Hong Kong.

This kind of change moves stocks for reasons entirely unrelated to profits. A large pool of money that was previously forbidden from buying a share becomes able to buy it, and some of that money is in funds that must hold whatever is in their benchmark. Baidu's own finance chief estimates it brings an additional 10-15% of fund flows into the stock. Singh files it as "an interesting one to watch" rather than a conviction position — the flow is real, but it is a one-time repricing, not a change in the business.

The wider context is more interesting for the company. On the same call he notes that Chinese open-weight AI models now account for roughly 70% of usage on OpenRouter, a service that routes requests between models, and that the average price of AI processing has halved in three months to under a dollar per million tokens. That is bad news for the American labs charging premium prices, and structurally better news for a Chinese developer that gives its models away and monetises the surrounding services.

Take-Two owns Rockstar Games, the studio behind Grand Theft Auto. The next instalment is scheduled for release on 19 November, and the shares fell 7% as leaked material from the game continued to circulate online.

Singh's view is that the market confused an embarrassment with a delay. Leaks are unpleasant for a studio, but they do not change a release date that has been publicly committed to, and they do not stop the millions of people who have already decided to buy the game. "We think that GTA will still be released over the next couple months and it was a market over reaction."

This is the smallest and simplest form of event-driven investing: separate news that changes the cash a company will collect from news that merely changes how people feel about it, and buy the second kind when it is sold as though it were the first.

Nvidia makes the processors nearly all AI systems are trained on. On the business itself Singh is positive: the company is guiding to 70% growth, says it cannot make enough chips to meet demand, and has agreed to buy Hugging Face — the main public repository of open AI models — for $12.93 billion. "Both are supply constrained and saying real demand is higher, but the market is not giving them credit."

His hesitation is about who is buying the shares, not about the company. Institutional investors have finished rebuilding their positions, the quantitative and momentum funds that were buying have finished buying, and with earnings season over there is no upcoming event to move the price. What is left is retail: individuals bought $250 million of Nvidia in one day and more than $2.5 billion across fifteen consecutive sessions — "which is bananas."

An asset whose only remaining incremental buyer is the general public is a fragile one, because that buyer has no obligation to keep buying and tends to sell together. That is why both JPMorgan and Citadel are advising clients to reduce risk for the next three weeks, and why a good company can still be a neutral position for a month.

Tempus applies AI to genomic and clinical data for cancer treatment. It appears twice on this call, pointing in opposite directions — which is why it earns a mention but not a view.

On one side, Cantor Fitzgerald began coverage with a positive rating and an $80 price target. On the other, Tempus shows up on Singh's own new tracker of insider transactions as one of the week's largest sellers by company insiders. Analysts publish opinions; executives sell shares with their own money and better information. When those two signals conflict, the honest answer is not to average them.

His response is exactly that: "I have to look at that." It is also a good illustration of what the Form 4 screener being built for October is meant to do — surface the conflict quickly, and then let the analyst decide whether a particular sale is meaningful or simply an executive diversifying.

Anthropic is a private AI laboratory, so it has no share price — but its estimated value has become load-bearing for the public market, because Amazon and Alphabet own stakes in it and book the paper gains through their profits. That is why JPMorgan Asset Management flags "an unusually high share of S&P 500 earnings due to unrealized valuation gains."

This week the mark went the wrong way. Estimates of Anthropic's value have fallen from roughly $1.4 trillion to about $900 billion, and the initial public offering has been pushed back again, to mid-October. Singh attributes it to competition: Chinese open-weight models are taking share, and the price of AI processing has halved in three months, which squeezes the revenue any lab can charge for the same work.

It matters in two other places on the same call. Anthropic is one of four named customers behind Broadcom's commitment to enable more than 20 gigawatts of computing capacity through 2028, and it has reportedly signed a $35 billion cloud contract with Lambda. A great deal of the AI infrastructure story is underwritten by a private company whose own valuation just fell by a third. The reassuring counterweight Singh supplies: strip the unrealised gains out entirely and S&P 500 earnings growth is "still very strong at above 30%."

Gold pays no interest, so it competes with bonds. When the yield on a bond rises faster than inflation — what economists call the real interest rate — holding gold costs you more, and the gold price tends to fall. That is exactly the environment right now, with the market pricing a coin-flip chance the Federal Reserve raises rates on 16 September. Gold slipped slightly on the day; silver rose.

Singh separates the next few months from the next few years, and gives different advice for each. Near term, if the Fed hikes, gold takes "temporary pressure," so this is a reasonable moment to trim a position or sell covered calls against it — that is, sell someone else the right to buy your gold at a higher price, collecting a fee now in exchange for capping your upside. Longer term he is emphatic: use the weakness to add "going into 2027."

The reason is arithmetic rather than sentiment. Developed governments will issue roughly $18 trillion of debt this year, $4 trillion of it genuinely new borrowing, two-thirds of that American. He does not think that debt will be defaulted on; he thinks it will be inflated away, which means every paper currency loses purchasing power together and the metal that cannot be printed gains it. He calls this the debasement trade, "the single biggest trade of coming years." The supporting anecdote is India, where the prime minister is trying to discourage gold imports — which Singh reads as a sign of how strong the demand is, not how weak.

LyondellBasell turns oil and natural gas into plastics and industrial chemicals. Singh owned it as a cyclical trade — buying a commodity business when the cycle is at its worst, on the expectation that the cycle turns. That worked: chemical shares rallied significantly from the bottom.

He is now closing the idea, and the reason is not the company. Oil and gas are the raw materials a chemicals plant buys, and the war in the Middle East has pushed their prices sharply higher while the finished plastics cannot be repriced fast enough to keep up. A cyclical recovery only pays if the margin between input and output widens; here it is being squeezed from the input side.

His instruction is direct: "what I would do for LYB is just trim or sell." And the point generalises beyond one stock — he expects the Iran conflict to escalate rather than resolve before the November midterms, with "no immediate catalyst for peace," so "names that are sensitive to oil prices and gas prices, I'm just not as excited about." That is a whole-sector de-rating driven by a political calendar rather than by anything a chemicals executive controls.

Caterpillar makes construction and mining equipment — the classic cyclical business, whose profits swing with how much building the world is doing. The shares have run to a high valuation for a company of that type, which is why some investors have been looking at it as a short: a bet the price falls.

The new information is an insider sale of unusual size. The chief executive sold 48% of his personal holding. Executives sell shares for all sorts of innocent reasons — taxes, a house, diversification — but selling roughly half of a stake is different in kind from trimming, and the person doing it has the clearest possible view of the company's order book. Singh's phrasing is appropriately measured: it "might mean the company might have peaked in the short term."

The rest of the call supports the same direction. Construction outside data centres is declining, mortgage rates may go above 7%, and Bank of America is warning that rising commodity costs are squeezing margins for companies without pricing power. A richly valued cyclical, in a decelerating cycle, whose chief executive is halving his stake, is a coherent short case rather than a single datapoint.

Jay Singh — Weekly SSR: CPI/PPI week with Kalshi at 52% for a hike, CENTCOM disables three Iranian tankers, the 40-year bond bull market declared over — and two new event-driven names built to survive all of it, Entravision (EVC) and Himax (HIMX)

1. Stocks & names mentioned

2. Talking points

The mandate for the week: buy events, not beta — 00:00

The calendar is the whole week: PPI Wednesday, CPI Friday, and a war premium inside both — 01:30

Jefferies initiates nuclear — previewed on the call, published next weekend — 01:30

Special situations: VRNS, SPGI's Capital IQ spin, ATAI's close, and the antitrust calendar — 06:08

LPSN: the two-day layup, alerted on the Discord — 06:08

PayPal: a rumour, a range, and a named bidder map — 06:08

The Goldman macro summary: the trend is intact, the next three weeks are not — 13:07

The war moves to the oil fleet: three tankers for two warships — 16:31 · 21:31

A dated stalemate: nothing changes until October 27 and November 3 — 16:31

Trump threatens the Fed with tariffs — and it would force the opposite — 24:42

The bank split on September 16: UBS and Barclays hawkish, Morgan Stanley on hold, Singh below all of them — 24:42 · 42:28 · 42:56

The 40-year bond bull market is over — and "return to normal" is the wrong frame — 30:26

The consequences are already visible: $1.2T of debt service, a PE distribution drought, and cash with nowhere to go but AI bonds — 30:26

Japan breaks the anchor: a record modern 30-year JGB and a carry-trade unwind — 37:11

Norway, France, and the ceiling on how far this can go — 37:11 · 43:18 · 43:54

Citadel and JPMorgan: three weeks of caution, with retail as the only bid — 44:53

Credit: CCCs break out, and the honest sizing of it — 45:46

Private credit gates are back — across seven managers at once — 1:14:07

Commodities: a super-cycle, copper at a new high, diesel cracks at a record $106 — 46:32 · 1:16:51 · 1:18:32

Housing and the rate-sensitive consumer — 46:32 · 47:04

Chinese open-weight models and the collapse in token prices — 47:28

The insider tape, and a Form 4 screener shipping in October — 49:30 · 50:10

Defense: seven-year THAAD and PAC-3 frameworks, and a backlog that doubled in a quarter — 52:39

Entravision: the political-ad windfall hiding inside an ad-tech re-rating — 53:55

Weyerhaeuser: a call option on housing, priced at a 30-40% discount to dirt — 1:01:07

Himax: 40% of automotive display drivers, and the sensor inside every Meta Ray-Ban — 1:06:46

Midterms: the market now expects a sweep; he does not — 1:13:49

Index rebalance, Caterpillar's CEO, and Chevron in the Orinoco — 1:15:01 · 1:15:52

Power is the constraint, not chips — 1:18:32

Gold, silver and the debasement trade, asked directly — 1:33:12

LyondellBasell: the cyclical-bottom trade is closed — 1:31:39

3. In plain English

EVC — Entravision Communications Positive

HIMX — Himax Technologies Positive

WY — Weyerhaeuser Positive

AVGO — Broadcom Positive

SPGI — S&P Global Positive

LPSN — LivePerson Positive

PYPL — PayPal Holdings Positive

VRNS — Varonis Systems Neutral

ATAI — ATAI Life Sciences Neutral

WBD — Warner Bros. Discovery Neutral

CVX — Chevron Positive

VST — Vistra Positive

MU — Micron Technology Positive

DELL — Dell Technologies Positive

BIDU — Baidu Positive

TTWO — Take-Two Interactive Positive

NVDA — NVIDIA Neutral

TEM — Tempus AI Neutral

Anthropic Neutral

GLD — SPDR Gold Shares Positive

LYB — LyondellBasell Industries Negative

CAT — Caterpillar Negative