| Ticker | Name | Research | View | What Singh said | Ref |
|---|---|---|---|---|---|
| AC | Air Canada | QT · SA · STK · FA | Positive | The week's first trade idea — an odd-lot Dutch tender arbitrage, sized for small accounts and explicitly insulated from the tape. "If you're Canadian and you're not subject to Canadian withholding tax… it closed on Friday at 27.53 and this Dutch tender is supposed to take place on September 24th, and it's going to be between 29 and 34" (the deck: a C$800M substantial issuer bid at C$29.00-33.00 for up to 27.59M shares, ~9.8%). "The minimum you would make is 5%… you'll likely get closer to the 5 than the 20% type return." The edge is the provision, not the price: "there's going to be a priority allocation without proration for holders of less than 99 shares. So if you own 99 shares across 10 accounts… held separately at a separate broker… you can basically own $27,000 of this… and if you were to make the most you could… you'd make about 4,600. So this is mostly for small traders, but it is something to take a look at." Funding is cash already received — "liquidity proceeds from a recent sale of their Aeroplan loyalty platform" (the deck says a minority-stake monetization). | 09:12 (deck p. 4) |
| EVC | Entravision Communications | QT · SA · STK · FA | Positive | The SEP-07 value special situation, now with a published model (the Excel is in this folder; the transcript garbles the ticker as "EVBC" and the report compounds it into "Everyday Health" — it is Entravision). "One of our analysts has modeled them for you… we calculated our own cost of equity using the company's betas, the risk-free rate… we used the CAPM… using the new 10-year Treasury as a risk-free rate. We use a risk premium of 4% and we put in the company's cost of debt. We then modeled out the free cash flows from revenue to EBITDA to EBIT. We adjusted for capex to get to free cash flow to the firm… net present value and then the terminal value, and then we subtracted out the net debt to get a fair price. Now, these are not perfect calculations, but this implies roughly 59% upside on EVC." The thesis is unchanged from the prior calls — "its fast-growing pivot into advertising away from its core business. It's been growing like a weed, and it trades relatively cheap to its cash flow growth" — with the caveat printed on the deck page: both EVC and PRTH "are small caps and market sensitive, so they will fall with a higher beta to large cap equities in a true risk off market." | 16:37 (deck p. 5) |
| PRTH | Priority Technology Holdings | QT · SA · STK · FA | Positive | The controlling-shareholder lowball, re-underwritten with a DCF — and already half-harvested. "We bought this in the low fives, we sold half of it above six. It rallied because of two small hedge funds getting involved. We can add back to it if it sells off on Monday back in the fives." The model: "it generates about 200 million of unlevered free cash flow… about 100 million of free cash flow [after debt]… if you were to value this 200 million… growing at single digits, that would probably be worth about 2.6 billion dollars, even penalizing this company with a 2% small cap premium in its weighted average cost of capital. So that gives us about $21 a share." He discounts his own number: "that doesn't take into account AI risk… it does have a software component. So it should probably trade at a discount to this $21, but it probably should not trade at 5.94… it basically trades at a quarter of what its cash flow say it should be trading at." The catalysts: the CEO (who "owns like 60%") paying "nine to 10 dollars a share just to get you guys off my back," or a full sale — because a take-private needs a majority of the minority: "there's 4% activists of the 40%, they need to convince another basically 16 plus percent… and they can reject the CEO's offer." The named flaw: the CEO's roll-ups (a $15M acquisition last month) — "he can make these acquisitions because he does control the shares and the majority of the minority doesn't apply… it only applies when he's bidding on the business. So it is not a perfect trade." Realised: "we sold half at a decent profit… over a 15% profit, and we will likely buy back if it sells off again." | 16:37 (deck p. 6) |
| RWT | Redwood Trust | QT · SA · STK · FA | Positive | Adding into a mechanical flush — common and baby bonds. "We added to Redwood equity and baby bonds on a stock overreaction to the convert issue last week. They issued a 150 million convert… to refi a 2027 bond… at a 35% premium, which is close to their book value… and they offered a 7% coupon, which is much lower than their other bonds, which had over 9%." The mechanism: "convertible arbitrage funds… had to short equity to hedge this 150 million convert. Assuming an initial delta between 50 and 65%, hedge fund arbitragers had to short between 75 million and 100 million worth of common stock… with a company with about a $500 million market cap, dumping 75 million of short selling into the open market in a single trading session creates an immediate downward pressure. So the stock was down like 22%." Plus a 2% book hit from selling $190M of legacy bridge loans and a concurrent $20M buyback. His fills (deck p. 12): "Added 5 bps RWT at 3.41-3.45; 20% yield, 45% discount to June book · Added another 10 bps RWT at 3.21 · Got my average down below $3.60 · Added 5 bps RWTQ and RWTS baby bonds at 11%+ yields" — bonds that "will likely be repaid at par just like it just repaid its 27 bonds at par." UBS concurs: the sell-off reflects "convertible arbitrage hedging, dilution concerns, and uncertainty… rather than a materially deteriorating credit profile." Sizing discipline stated: "you can't have too big of a position here because interest rates continue to be a risk… if it sells off further, we will continue to add… this company survived COVID… still one of the least levered," with borrowers at "around 750" FICO — "not a default risk… a mark-to-market risk." | 16:37 (deck p. 12) |
| RSVR | Reservoir Media | QT · SA · STK · FA | Positive | A new merger-arb long bought after the spread re-opened. "We added RSVR after a PE offer at 10.50 a share after the shares weakened below $9 a share… this stock was trading over 11 because of this offer… people thought there was going to be a topping bid above 10.50. The shares then fell all the way to nine. We bought them in the low nines, and we think that this 10.50 share at minimum should go through." The bidders: "the 10.50 cash bid from a majority insider, Wesbild / Richmond Hill, and competing interest from an activist fund that also wants to buy called Irenic Capital" (the deck: Irenic in the $10-11 range) — "Wesbild controls about 44% of the equity and Irenic owns 9.2%. So combined, these guys own 53%," with "no financing condition." The business underneath: "624 million market cap, a billion enterprise value, 450 million of debt, but 70 million of EBITDA… 40 million of free cash flow after paying interest, which is about an 8% free cash flow yield… growing between 10 and 18% a year," a music publisher at a 65% gross margin whose catalogs earn "inflation-hedge recurring revenue… low correlation to broader economic cycles." The bracket: "before the company received these bids, the unaffected price was around 7.50… worst-case scenario, I think it trades down below eight. Best-case scenario, it trades above 10.50 because there's a bidding war… we do think there's a high probability that this bid will go through." Asked when it closes: "there is no close date because it's not clear which bidder is going to win yet." | 16:37 (deck p. 13) · 1:51:20 |
| IMPP | Imperial Petroleum | QT · SA · STK · FA | Positive | A held tanker position, restated on record freight. "Tanker rates hit an all-time high at 800k per day. We still own IMPP. We think the stock could be 50 to 100% up from here, even if tanker rates go down from these local peaks. Tanker rates were up 43% just in the last week." The driver is ton-miles, not barrels: the Houthis "blocking and targeting Saudi affiliated vessels, forcing mainstream non-shadow tankers to… bypass the Red Sea altogether via the Cape of Good Hope, which is why tanker rates are so high, 800,000 a day… because… they have to go all the way around Africa." (The deck page header: "Tanker Rates at All Time Highs 800k per Day and Oil Refining Capacity at All Time Lows — Still Own IMPP.") | 1:41:28 (deck p. 47) |
| IJR | iShares Core S&P Small-Cap ETF | QT · SA · STK | Positive | The long leg of the small-cap pair, chosen on index construction. "By comparison, less than 7% of the S&P 500 is unprofitable… The S&P Small Cap 600 has an explicit positive earnings screen keeping its unprofitable share around 20%. So if you want small cap exposure, you should buy the S&P Small Cap 600 index over the IWM and short the IWM." The report names the vehicle: "Structurally favor the S&P SmallCap 600 over the Russell 2000 (via long IJR / short IWM)." (The call names the index, not the ETF; IJR is the report's expression.) | 1:11:00 (deck p. 26) |
| UBER | Uber Technologies | QT · SA · STK · FA | Positive | Two open-market insider buys on a name he already calls cheap. "Uber president, COO purchased 70,000 shares for 5.3 million… we've talked about Uber being cheap." And then the chief executive: "Uber's Dara Khosrowshahi also bought 141,000 shares for 10 million at an average price of 71. This marks Dara's first open market purchase of Uber stock since May 2022, when he bought 200,000 shares at 26.73, and the shares have almost tripled since he did that. So I think the market was quite excited about the two big insider purchases of Uber." (He misspeaks the COO's buy as "one of the biggest insider sales"; the deck confirms both are purchases.) | 1:41:28 |
| PANW | Palo Alto Networks | QT · SA · STK · FA | Positive | A house preference, confirmed by an outside list and a conference. On the Morgan Stanley non-AI picks: "Palo Alto, which we also like, is on the list." From the Goldman TMT conference: "cybersecurity again stood out as one of the most direct AI beneficiaries… Palo Alto's CEO also said that the AI security industry has not even been built yet. The market cap of the two [with CrowdStrike] is circa 490 billion versus AI companies in the trillions." The argument for why AI does not eat the category: "a multi-modal AI strategy is essential as different models, including Mythos, OpenAI and others, uncover different vulnerabilities and blind spots… while AI is highly effective for mainstream security tasks, it struggles with edge cases and contextual understanding, which is why you still need these cybersecurity companies." | 16:37 (deck p. 14) · 1:36:47 |
| KNTK | Kinetik Holdings | QT · SA · STK · FA | Neutral | A new special situation, explicitly on the work list rather than in the book. "Bloomberg reported that Kinetik, which is KNTK, is going to sell itself. So that's on the watch list. We haven't added it yet. We don't know enough about the situation." The ownership map he pulled live: "Blackstone owns about 14.8% of it and BlackRock owns 11%… Goldman Sachs owns 10%… collectively about 34%… Jamie Welch… the CEO… owns about 4%," with Class A at 54.75 (the deck adds that Blackstone also holds 70% of the Class C, "close to half of the shareholder voting power," and that ET, TRGP and PSX have been mentioned as suitors). The open question is the earnings base: "the company trades at about 13 times forward EBITDA, about 25 times trailing EBITDA… their EBITDA… more than doubled from the first quarter to the second quarter [100M to 250M]. So I need to understand what drove that before I can determine… the takeout potential" — because it owns "a large network of gas gathering and processing infrastructure in the Permian" and "takes a percentage of the price exposure." Next step named: "I do have a call with Jefferies and one of the Goldman analysts next week to figure out what would be an appropriate takeout bid." | 09:12 (deck p. 3) |
| WBD | Warner Bros. Discovery | QT · SA · STK · FA | Neutral | The one large-cap arb held in size, now mostly collected. "With respect to large-cap merger arbitrage and special situations, the only one of these that we own in size is this Warner Brothers–Paramount, which has compressed from over 20%. … Most of our merger arb exposure is in much smaller names." No new catalyst is discussed on the call; the position is carried as the largest remaining large-cap spread exposure while the book's new money goes to small-cap situations. | 09:12 |
| NSC | Norfolk Southern | QT · SA · STK · FA | Neutral | A tracker-sized position in the rail merger. "We own a little bit of this NSC/UNP as a tracker, but not a lot." Last week he called NSC/UNP the widest large-cap spread on the screen; this week it is confirmed as a small monitoring stake rather than a conviction position — consistent with the 2027 regulatory calendar that stretches its annualised return. | 09:12 |
| ROKU | Roku | QT · SA · STK · FA | Neutral | A pass on spread width, not on deal risk. "There are some antitrust issues around Roku and Fox A, but that spread is just not wide enough to be involved in." The news: "media reports last week about the DOJ preparing to expand its antitrust probe into Fox's proposed $22 billion takeover of Roku. Investors were slightly surprised but not shocked, given there's only a 70/30 odds of clearing… the base case is that this is going to close." The spread "widened out to a 6.5% IRR to a Feb close" (the deck: ~$4 to ~$5 a share, using a 3/1/27 close), with "a lot of big hedge funds putting the spread on." His verdict: "That spread is not wide enough for us to make any money on because I'm not going to borrow money to put on a silly spread like that." | 09:12 (deck p. 2) |
| KO | Coca-Cola | QT · SA · STK · FA | Neutral | The one Morgan Stanley pick he works through himself — and declines. "You have Coca-Cola here, which MS thinks has 13% upside despite the GLP-1 fears. … KO closed at 82.29… has a dividend yield of about 2.4%, but it's a growing dividend and it has free cash flow generation of about 12 to 14 billion a year… if you take 13 billion on 379, that's a three and a half percent free cash yield. So it's not glaringly cheap, but it has sold off pretty dramatically from local highs… in the low to mid 90s. Hasn't sold off as much as I would have liked, but it has pulled back more than the S&P." In Q&A he ranks it above McDonald's — "Coca-Cola is still doing well because Diet Coke sales are still doing well" — and then: "I'm not going to be buying McDonald's or Coca-Cola, but I do think that they are more defensive names." | 16:37 (deck p. 14) · 1:52:10 |
| MCD | McDonald's | QT · SA · STK · FA | Neutral | Answered in Q&A as a defensive name trading below the Street's bear case — and still not a buy. "It sold off because of GLP concerns… with interest rates. It was at 340 back on February 27th… it's now down to 252… probably 26%, and it pays a dividend of about 3%… it has been growing its dividend forever." The desperation tell: Spicy McNuggets back, a SpongeBob/One Piece Happy Meal on September 15th, a $2 in-app breakfast sandwich — "it looks like these guys are desperate to bring sales back," as are Starbucks, Pret, Potbelly and the rest. The Wells Fargo frame: "15 times EBITDA versus a historical 20 times… their upside case is 390, their base case is 300… McDonald's is already trading below their downside case, about 4% below that… about 19% upside" to base. The verdict: "I'm not an expert in McDonald's… I would be wary of some of the GLP-1 issues. I think a sixth of Americans are on GLP-1s… I'm not that excited personally." | 1:52:10 |
| LEN | Lennar | QT · SA · STK · FA | Neutral | The single earnings print he will read this week, as a housing gauge rather than a position. "The only thing that I'm going to be reading next week on the earnings front is Lennar, because I want to see how their sales are doing with higher mortgage rates." The setup from his calendar: housing starts at 1.3M "still relatively low because of mortgage rates being at 7%," a bounce only because July fell 12%, and "building permits are down from 1.43 million to 1.40 million. So the following month, you're probably going to see housing data fall back down again." (The report's calendar flags the focus: cancellation rates, mortgage-rate buydowns and gross margins.) | 06:36 · 09:12 |
| ORCL | Oracle | QT · SA · STK · FA | Neutral | A blow-out backlog financed by the shareholders — and a founder sale that was announced and withdrawn inside the call. "Oracle delivered more than 300,000 GPUs to its AI cloud customers since the end of Q4… RPO surged by 209 billion to 664 billion, but at the same time, Larry Ellison announced a new 10b5-1 plan to sell 7.5 billion of stock… Oracle also guided FY revenue to at least 90 billion and adjusted EPS to $8 a share while disclosing 28.5 billion of capex, negative 5 billion of free cash flow and a $20 billion ATM equity program, which means that it's going to be issuing stock. It's probably why Larry Ellison is trying to front run that and sell his own stock. First time in basically ever. So that's not going to be good when the market opens." Then, from a subscriber in Q&A: "Oracle chairman Larry Ellison just canceled the planned 7.5 billion stock sale without selling a single share, especially I think given this AI slowdown. I don't think he wanted to look like he was panicking." The overhang is removed; the ATM and negative free cash flow are not. | 1:41:28 · 1:51:49 |
| NVDA | NVIDIA | QT · SA · STK · FA | Neutral | Fundamentals untouched, sentiment exposed — and the name at the end of the funding chain he is worried about. The flywheel: "if they aren't supporting their cloud backlogs, how are they going to increase capex? … If they don't increase capex, who's going to buy Nvidia GPUs so that it grows earnings 30% every year and who's going to be buying all this memory? Now, it's not going to matter in the next quarter or two." The balance: "for the actual AI trade, Nvidia, memory, power, optical… nothing here changes the near-term capex number. No hyperscaler guide has moved… for now, this is sentiment risk, not fundamental risk." The near-term tape worry: "semiconductors could sell off and energy might be the only sector that rallies tomorrow because of this AI slowdown." From Goldman's TMT conference, Nvidia's own message: "the bottleneck has basically moved from chips to power. Demand is broadening rather than narrowing." | 16:37 (deck pp. 10-11) · 1:33:45 · 1:36:47 |
| Anthropic | Anthropic (private) | — | Neutral | The weekend's catalyst, and a valuation he now treats as a ceiling. "Anthropic CEO Dario Amodei publicly called for a formal industry-wide slowdown of the pace of frontier AI model development… the trigger was a security incident involving an autonomous agent swarm… the OpenAI–Hugging Face incident," with Anthropic committing to external auditors like METR and Altman and Musk backing the call. The IPO read: "we think that he's going to delay his IPO till October at the earliest, next year is the most likely," against investors "pushing hard towards an October 26th IPO reported at a close to two trillion valuation… apparently this is a dangerous enough moment that going public is ill-advised, unless, of course, you happen to be a company trying to sell two trillion of equity into it." The letter itself "is not a growth warning… he does not touch the numbers at all… but the fact that he felt the need to explicitly reassure investors… tells you he knows exactly how the market is going to hear it." His mark: "I doubt Anthropic is going to be worth more than 1.7 trillion anytime soon." And the systemic link: "the real risk is that Anthropic or OpenAI run into funding issues and slow down cloud spending, which in turn lowers the backlog growth of Microsoft, Google and Amazon, who then… cut capex… but we won't know that for a while." | 16:37 (deck pp. 10-11) |
| ATAI | ATAI Life Sciences | QT · SA · STK · FA | Neutral | The arb's terminal event, noted in passing. "ATAI, the deal is expected to close on September 11th, so that likely closed." The deck adds the NASDAQ trader alert: halted after the after-hours session on September 10, closing before the open on September 11, and suspended effective September 14 — two days later than the September 9 date given on the prior call. | 09:12 (deck p. 3) |
| GME | GameStop | QT · SA · STK · FA | Neutral | A large insider buy he declines to read as a signal. "GameStop's CEO bought 1 million shares at an average price of 20.38. Don't know what he's doing." (The deck: Ryan Cohen's purchase is worth ~$20.4M, taking his beneficial ownership to 43.1M shares, 8.5%.) Contrast the same page's Uber entry, where a first CEO purchase since 2022 is read as information — the difference is whether the buyer's intent for the balance sheet is legible. | 1:41:28 |
| VLY | Valley National Bancorp | QT · SA · STK · FA | Negative | The new short — the interest-rate hedge for a Fed he now expects to hike. "If they hike, that's going to be very bad for commercial real estate, which is why we were adding a bank short that's the most exposed to commercial real estate in the United States." The transmission: "commercial real estate is not valued on the back end of the curve like mortgage rates for residential are. Commercial real estate is benchmarked off SOFR, which is basically the same as the Fed funds rate." The name: "VLY… is a short target primarily due to its structural exposure to commercial real estate debt… elevated liability costs, very minimal NIM… close to its 52-week highs despite being one of the most exposed banks." The numbers: "a total loan book of about 51 billion… commercial real estate exposure of 30 billion, so almost 60%… non-owner-occupied commercial real estate is about 329% of total risk-based capital" (down from 474%) — and the quality problem: "it sold a lot of its good properties and kept a lot of the bad properties." Plus "non-interest income… accounts for only 13% of its total revenue" and an allowance for credit losses that "has historically trailed" peers (deck: ~1.19% vs 1.4-1.5%). The chart: down 50% to almost $6 by April 2023, now "around 13.80," the highest since 2009. Target: "we think this bank could easily be down 20-30% on an interest rate scare." Why this short and not a crowded one: "the short interest isn't too high, it's around 9%… this is a short that I can put on and not be as scared about… if the short interest was like 15, 20%, I would be more worried." Risks named: CET1 up to 11% and S&P's positive outlook — "but I don't think S&P ever thought that we would be hiking rates again." | 04:19 · 16:37 (deck pp. 7-9) |
| IWM | iShares Russell 2000 ETF | QT · SA · STK | Negative | The short leg of the small-cap pair, and the index most exposed to a hike. "The Russell 2000 is the most sensitive to interest rates with 40 to 45% of the index being unprofitable… which is why the Russell is so weak last week… pre-2010, the non-profitability was only 15 to 20%." Three mechanisms: "30 to 45% of the debt held by Russell 2000 companies carries floating rate interest exposure compared to less than 10% for the S&P 500"; "a 2026-2027 maturity wall"; and "40% of… Russell 2000 constituents have interest coverage ratios below 1.5, which means their EBITDA cannot cover their interest payments." The instruction: "you should buy the S&P Small Cap 600 index over the IWM and short the IWM." (At the close of the call the Russell was the least weak index — "only down 16 bips because… bond yields are actually back down" on the growth scare — the pair's reminder that it is a rates trade.) | 1:11:00 (deck pp. 25-26) |
| SKM | SK Telecom | QT · SA · STK · FA | Negative | A full exit, in profit, because the optionality has been paid for. "We're exiting SKM. You know, we've done really well in this name. We're just worried about this AI slowdown and panic after Dario's letter being published." The original trade: "a high asymmetry deep value proxy trade, a dividend-paying Korean telecom holding 3.7 million shares of Anthropic, giving massive private exposure relative to its $15 billion market cap." Why now: "the arbitrage now is closed. At its lows of 19 to 20… SKM was fundamentally dirt cheap, providing Anthropic optionality for near free. Now it's rallied to 38 to 40 range. The stock was already pricing in a neutral core telecom at 11 billion for the core business, plus Anthropic at a private mark of 1.7 trillion… so it's fully priced in." The deck's framing of the flipped asymmetry: holding at peak "means you no longer hold a pure 'free option.' You hold a Korean telecom trading near fair value, carrying single-stock Korean market discount risks… and downside exposure to any private valuation markdown." Execution: "we're selling in the pre-market near highs as Anthropic is fully valued… even if they open a little bit lower, I'm not worried about it" — and the door left open: "if you need to buy it cheaper, you can always buy it cheaper" (deck title: "Can Re-enter Later"). | 16:37 (deck p. 10) |
| UWMC | UWM Holdings | QT · SA · STK · FA | Negative | Named as the management counter-example inside the Redwood thesis. "It's not like the UWMC where you have trash management that, you know, did an equity rights offering and diluted the living crap out of investors after taking, you know, billions of dollars in dividends for themselves. This is a completely different type of management team." A governance judgement rather than a trade — the reason one mortgage lender's sell-off is a buy and another's is not. | 16:37 (deck p. 12) |
"View" is the house's stance in this call (Positive / Neutral / Negative), not a price rating. Referenced only (kept in the talking points, not tickerized here): the Morgan Stanley "mostly non-AI" global picks he sends as a screening list (MS Non-AI Top Picks.xlsx) and reads a flavour of — ABN AMRO (~10% upside), ADES Holding (Saudi offshore drilling, ~40%), Abu Dhabi Commercial Bank (~20%), Accton Technology (Taiwan networking, ~86%; the transcript's "Actin"), Alibaba (BABA) (~60%), Alpha Bank, CATL, Rolls-Royce ("interesting"), Warner Music Group (WMG) ("could eventually be sold"), Northrop Grumman (NOC) and Lloyds Banking (flagged as a value trap) — MS's calls, not his, except PANW and KO which he addresses directly; the Jefferies nuclear initiation read out from the deck — Buys on BWXT, Cameco (CCO/CCJ), Denison (DML), enCore (EU), Mirion (MIR), NexGen (NXE — resolving last week's "NXC"), SOLS and Kazatomprom (KAP), Holds on Global Atomic (GLO), Centrus (LEU) and Uranium Energy (UEC) — a sell-side view he reports ("I don't know how he can forecast that") without taking a position; Fox (FOXA) as Roku's acquirer and Paramount Skydance (PSKY) as WBD's; the deck's deal page — Centerspace (CSR) agreeing an all-stock merger into Independence Realty (IRT) at 3.8 shares, Bio-Techne (TECH), GBTG, BioLife (BLFS) and LXFR regulatory steps, and ET, TRGP, PSX as rumoured KNTK suitors; the Goldman TMT conference cast — CrowdStrike (CRWD), Adyen, Shopify (SHOP), Capgemini, Lam Research (LRCX) ("fundamentally sold out"), STMicro (STM), T-Mobile (TMUS) and SpaceX's $1.11B/month hosting deal ("I'll let you read") — and PayPal's tariff comment ("the PayPal issue seems more PayPal specific"); the news recap — Palantir (PLTR) with Nebius (NBIS) and Nvidia, IBIT inflows, Meta (META)'s Muse agent, IREN's ERCOT Batch Zero inclusion, Rigetti (RGTI) and D-Wave (QBTS) government funding, Apple (AAPL)'s $2,000 iPhone Duo, Google's $13bn Finland build, Amazon's first sterling bonds, Microsoft's 12 GW → 38 GW capacity plan, Adobe (ADBE)'s slight beat and Robinhood (HOOD)'s August metrics; Trip.com / TripAdvisor on the earnings calendar; Blackstone's real-estate head quitting and Palmer Square (private, $27bn CLO platform) exploring a sale; and AT&T as the source of the wireless-services CPI spike. Private and unlisted names: OpenAI, xAI, METR, Hugging Face, Stripe (Link), Wesbild / Richmond Hill and Irenic Capital (the RSVR bidders), and the two unnamed PRTH activists. Unresolved: "Oxtay" in the special-situations list, and the exact Redwood baby-bond symbols (the deck's "RWTQ and RWTS"; the report adds "RWTN"), kept inside the RWT row.
A jargon-free summary of the thesis behind each name — what it actually is and why he holds that view. (Plain-language companion to the table above; renders on each ticker's consolidated page.)
Air Canada has offered to buy back about C$800 million of its own shares, using cash from its Aeroplan loyalty programme. It is doing so by "Dutch auction": shareholders say what price between C$29 and C$33 (Singh says 34) they are willing to sell at, and the company picks the lowest single price that lets it buy the number of shares it wants. Everyone whose offer is accepted gets that same price. The shares closed at C$27.53 — below the bottom of the range — so anyone who buys now and tenders should receive at least C$29.
The catch in most buybacks is "proration": if more shares are offered than the company wants, everyone only gets part of their shares bought, and the rest are left with you at whatever the market price is. This offer has a special rule for small holders. If you own 99 shares or fewer and tender all of them, you are exempt from that scaling-back and get fully bought. Because the rule applies per account, someone with ten separate brokerage accounts can put roughly C$27,000 to work this way.
The return is modest — about 5% at the floor, more if the auction clears higher — but it is available within about two weeks and has almost nothing to do with where the stock market goes. That is why Singh calls it "mostly for small traders." One important limit: it only makes sense for accounts that are not hit by Canadian withholding tax on the buyback.
Entravision is a Spanish-language TV and radio broadcaster in the US that has turned most of its business into digital advertising technology — the software that places online ads automatically. Singh introduced it two weeks ago as a company the market still prices like a shrinking broadcaster even though its ad-tech arm is growing very fast.
This week his team published a valuation model. The method is a standard one: estimate how much cash the business will throw off in future years, then work out what that future cash is worth today. To do that you need a "discount rate" — the return an investor should demand for the risk. They built it from the 10-year Treasury yield (the safe return), plus a 4% premium for owning stocks, adjusted for how much more the company's shares swing than the market. Then they subtracted the company's debt to get a value per share. The answer: roughly 59% above today's price.
He is careful to say the model is not precise, and the deck adds a warning that matters right now: this is a small company whose shares tend to fall harder than the big indexes in a genuine sell-off. The value case is long-term; the ride in the meantime can be rough.
Priority Technology processes card payments and business-to-business payments. Its founder and chief executive owns around 60% of the company and has offered to buy the rest from public shareholders for about $6 a share — roughly five times the cash the business produces in a year, which is very cheap.
Singh's team built a cash-flow model that values the business at about $21 a share, even after adding an extra penalty for being a small company. He does not expect $21. The business has some software exposure that artificial intelligence could disrupt, and it carries a lot of debt. But the gap between $21 and $5.94 is so large that he thinks the chief executive will have to raise his offer — perhaps to $9-10 — or the company will end up being sold to someone else.
The reason he has leverage is a governance rule. When a controlling owner buys out the minority, the deal normally needs approval from a majority of the shareholders who are not the controlling owner. Two activist funds own about 4%; they need about another 16% of the independent shares to vote no, and the bid fails. The weakness in the thesis is that this protection does not apply to the chief executive's other decisions, such as buying small companies with the firm's cash, which makes the future cash flow harder to predict. Singh bought in the low $5s, sold half above $6 for a gain of more than 15%, and plans to buy back if the price drops into the $5s again.
Redwood Trust is a mortgage company structured as a real estate investment trust. It needed to refinance a bond due in 2027, so it sold $150 million of convertible bonds — bonds that pay interest but can be swapped for shares if the share price rises enough. They paid 7%, cheaper than Redwood's existing 9%-plus debt.
The trouble was who buys convertibles. Hedge funds that specialise in them usually "hedge" by selling the company's shares short at the same moment — betting against the stock to offset the share-price exposure hidden inside the bond. For a $150 million convertible, that meant selling roughly $75-100 million of shares, in a company worth only about $500 million, in a single day. There were nowhere near enough natural buyers, so the stock fell about 22%. At the same time Redwood sold $190 million of its weakest loans at a small loss, which spooked people further.
Singh's view is that the fall was plumbing, not a change in the business — UBS said the same. He bought shares at $3.41-3.45 and again at $3.21, bringing his average cost under $3.60, and bought Redwood's exchange-traded "baby bonds" (small-denomination bonds that trade like shares) yielding about 11%, which he expects to be repaid in full just as the company repaid its last bonds. He likes the management, and its borrowers have strong credit scores (around 750), so the risk is not defaults. The risk is interest rates: rising rates push down the value of mortgage assets, which is why he keeps the position small and adds only in steps.
Reservoir Media owns music catalogues — the rights to songs and recordings — and earns royalties whenever they are streamed, broadcast or licensed. That kind of income is steady, grows with inflation, and has little to do with the economy, which is why investors who want predictable cash flow like buying it.
Earlier this year its largest shareholder, which already owns about 44%, offered $10.50 a share in cash to buy the rest. A second investor, the activist fund Irenic Capital with 9.2%, has also shown interest in buying. The shares rose above $11 on hopes of a bidding war, then drifted all the way down to about $9 as nothing happened. Singh bought in the low $9s.
The arithmetic is the attraction. If the $10.50 deal goes through, he makes roughly 15%. If a bidding war breaks out, more. If both bidders walk away, he estimates the shares fall to somewhere under $8 — the pre-bid price was about $7.50, but the business has grown since. The company earns about $70 million a year before interest and tax, produces $40 million of free cash (about an 8% yield on its market value), and the bids come without financing conditions. The one open question is timing: there is no closing date yet, because nobody knows which buyer will win.
Imperial Petroleum is a small company that owns oil and fuel tankers and rents them out. Its income depends on the daily rate charterers pay for a ship, and that rate is set by how much ship capacity is tied up at sea.
Right now a great deal is tied up. With the Strait of Hormuz largely closed and Houthi forces attacking Saudi-linked vessels in the Red Sea, many tankers are sailing all the way around Africa instead of through the Suez route. The same cargo takes far longer to deliver, so each ship is occupied for longer, and fewer are available. Rates jumped 43% in one week to a record of about $800,000 a day.
Singh still owns the stock and thinks it can rise 50-100% even if rates fall back from this peak — the point being that a small shipowner earning record rates can pile up cash very quickly relative to its size. The risk is the reverse: a reopening of the shipping lanes would bring rates down just as quickly.
There are two popular ways to buy "small US companies" through one fund: the Russell 2000 (the IWM fund) and the S&P SmallCap 600 (the IJR fund). They sound interchangeable. They are not, because of how each decides who gets in.
The Russell simply takes the next 2,000 companies by size, profitable or not — and today 40-45% of them lose money. The S&P 600 has a rule that companies must have positive earnings to be added, so only about 20% of its members are unprofitable. When interest rates rise, the loss-making, heavily indebted companies get hurt first, because much of their borrowing is at floating rates that reset immediately.
So Singh's instruction is to hold the S&P 600 if you want small-company exposure, and bet against the Russell 2000 alongside it. The pair makes money if the unprofitable, debt-laden small caps do worse than the profitable ones — which is what he expects if the Fed raises rates.
Uber runs ride-hailing and food delivery. Singh has said on earlier calls that it has become cheap. This week two of its most senior executives bought shares with their own money on the open market: the chief operating officer about $5.3 million, and the chief executive, Dara Khosrowshahi, about $10 million at around $71 a share.
Executives receive shares as pay all the time, and they sell for all sorts of personal reasons. Buying on the open market is different — it is a deliberate decision to put more personal money into the company at today's price, by the people with the best view of its prospects. The last time Khosrowshahi did this, in May 2022, he paid $26.73 a share, and the stock has nearly tripled since. That is why Singh says the market was "quite excited" about the two purchases.
Palo Alto Networks sells cybersecurity — firewalls, cloud security and tools to detect attacks. Singh says the house likes it, and it also appears on Morgan Stanley's list of favoured stocks that are not directly artificial-intelligence plays.
The argument from this month's Goldman Sachs technology conference is that AI makes security more necessary, not less. AI writes code faster, and every AI agent a company deploys is another identity that can be hijacked. AI tools can find software vulnerabilities much faster than people, but different AI models spot different weaknesses and still struggle with unusual situations — so companies need a security platform that combines several models and human judgement. Palo Alto's chief executive put it as "the AI security industry has not even been built yet," and the two leading security companies together are worth about $490 billion against AI companies valued in the trillions.
Kinetik owns pipelines and processing plants that gather natural gas from wells in the Permian Basin of New Mexico and Texas and prepare it for sale. Blackstone is its largest shareholder, and Bloomberg reports the company is exploring a sale.
A company putting itself up for sale is exactly the kind of situation Singh trades — but he has not bought, and he explains why. The shares trade at about 13 times next year's expected earnings before interest, tax and depreciation, but about 25 times last year's, because earnings more than doubled in a single quarter. A buyer will pay based on what it believes normal earnings are. If the jump came from gas prices rising after the wars in Ukraine and Iran — and Kinetik does take a cut of commodity prices, not just a fixed fee — a buyer may not pay for it.
So the name goes on the watch list, with a named next step: calls with analysts at Jefferies and Goldman Sachs to work out what a realistic takeover price would be. Until he knows the price, he does not know whether there is a trade.
Fox has agreed to buy Roku, the TV-streaming platform, for about $22 billion. Once a takeover is agreed, the target's shares trade a little below the offer price; the gap is the reward for waiting and for taking the risk the deal is blocked. This week the Justice Department signalled a deeper antitrust review, and the gap widened.
Singh still declines. The gap now works out to about 6.5% a year if the deal closes next February. Plenty of big hedge funds are happy with that because they borrow money to multiply small returns. Singh does not use borrowed money, and 6.5% a year with a real chance of the deal being blocked is not enough compensation on its own: "I'm not going to borrow money to put on a silly spread like that."
Coca-Cola is the classic defensive stock: people keep buying drinks in a recession, and the company has raised its dividend for decades. Morgan Stanley thinks it has 13% upside even with the worry that weight-loss drugs will reduce how much sugary drink people consume.
Singh does the check himself. The company generates about $13 billion of free cash a year on a market value of about $379 billion — a 3.5% cash yield, plus a growing 2.4% dividend. That is solid, "not glaringly cheap." The shares have fallen from the low-to-mid $90s to about $82, but not as far as he would like before buying. He rates it a better defensive holding than McDonald's, because Diet Coke is selling well while fast-food visits are being hurt by the same drugs — but he is not buying either right now.
McDonald's shares have fallen about 26%, from $340 in February to $252, which is the bottom of their three-year range. Two things did it: rising interest rates, which make steady dividend stocks less attractive than bonds, and weight-loss drugs of the GLP-1 type, which Singh estimates one in six Americans now take and which reduce how often people eat out.
You can see the pressure in the menu: returning favourites, a SpongeBob Happy Meal promotion, a $2 breakfast sandwich in the app — and the same cheap deals across Starbucks, Pret, Potbelly and the rest of the industry. Wells Fargo recently valued the company at $300 in its central case and $260 in its gloomy case, using a lower multiple of earnings than McDonald's historically commanded because growth is slowing. The shares are already below that gloomy case.
So it looks cheap against the analysts' own numbers, and it pays about 3% while you wait. Singh still passes: he is "not an expert in McDonald's," is wary of the weight-loss-drug effect, and says plainly he is not buying it.
Oracle's cloud business is booming on AI demand: it has delivered more than 300,000 AI chips to customers in a quarter, and its signed-but-not-yet-delivered contracts jumped by $209 billion to $664 billion. The problem is paying for it. Oracle plans $28.5 billion of capital spending, expects free cash flow of about minus $5 billion, and set up a $20 billion programme to sell new shares into the market over time — which dilutes existing owners.
In the same week, founder Larry Ellison filed a plan to sell $7.5 billion of his own shares. Singh read that as the founder getting ahead of the company's own share sales, and expected the stock to react badly. Then, during the call, a subscriber pointed out that Ellison had cancelled the plan without selling a share — Singh thinks because he did not want to look like he was panicking during an AI-slowdown scare. The founder sale is off; the company's need to issue shares and its cash burn are not.
Nvidia makes the chips almost every AI system runs on. Nothing about its business changed this week — no customer cut spending, no data centre was delayed. What changed is the mood, after the heads of Anthropic and OpenAI both talked about slowing down the development of the most powerful models.
Singh separates the two. Fundamentally, he says, "this is sentiment risk, not fundamental risk," and the rule is to wait for real evidence: a delayed model, cancelled chip orders, or a company lowering its forecast. But he describes the chain that worries him. The AI labs depend on raising huge sums from investors. If Anthropic's stock-market listing slips and OpenAI cannot list at all, the labs may spend less on cloud computing from Microsoft, Amazon and Google. Those companies would then slow their own spending — and they are the ones buying Nvidia's chips and all the memory that goes with them. None of that shows up for a quarter or two, but a mere slowdown in growth can be enough to push the shares down, and he warned semiconductors could sell off the next morning.
Anthropic is a private AI company preparing to list its shares on the stock market, reportedly as soon as late October at a value near $2 trillion. This weekend its chief executive, Dario Amodei, called for the whole industry to slow the pace at which the most powerful AI models are developed, after an incident in which a group of autonomous AI agents behaved in dangerous, unintended ways. Anthropic said it would let outside auditors check its training and releases.
Singh's reading has two halves. On the business, the letter changes nothing: it contains no delayed launches, no spending cuts, no lower revenue targets. On investor psychology, it is awkward timing for a company about to ask the public for an enormous sum — especially as OpenAI's Sam Altman, whose company is not listing this year, called this "an ill-advised moment to go public." Singh now expects the listing to slip to October at the earliest and more likely next year, and doubts Anthropic will be valued above $1.7 trillion any time soon. The larger danger he names is financial: if the AI labs cannot keep raising money, their cloud spending slows, and that ripples back through the biggest technology companies.
Valley National is a regional bank based in New Jersey that lends heavily against commercial property — apartment buildings in the New York area, suburban shopping centres, offices and warehouses. About $30 billion of its $51 billion of loans are commercial real estate, which is more than three times the bank's capital cushion. Regulators treat 300% as a warning level.
Singh is betting its shares fall, and the reason is how commercial property loans are priced. Home mortgages follow long-term bond yields, but commercial property loans float with short-term rates that move almost one-for-one with the Federal Reserve's policy rate. If the Fed raises rates next week, property owners' interest bills rise straight away, more of them struggle to refinance, and banks like Valley have to write down loans. Valley is worse placed than big banks for three reasons: only 13% of its revenue comes from fees, so it has little income that is unaffected by lending; its reserve for bad loans has been smaller than its peers'; and when it cut its property exposure after 2023, it sold the good loans and kept the weaker ones.
Two practical points shape the trade. The shares have rallied to around $13.80, near a five-year high, so a lot of good news is already in the price. And only about 9% of the shares are already sold short, which means there is less risk of a "short squeeze" — a sudden rise as short-sellers rush to buy back. He thinks the stock could fall 20-30% on an interest-rate scare. The risk to the idea is that the bank has genuinely strengthened: its capital ratio is up to 11%, and S&P Global has a positive outlook on it — though Singh notes that outlook was set before anyone expected rates to rise again.
The Russell 2000 is the most widely followed index of small US companies, and IWM is the fund that tracks it. Singh wants to bet against it because of what is inside: 40-45% of the companies lose money, far more than the 15-20% before 2010, a result of years of near-zero interest rates that let unprofitable firms borrow and list easily.
Those companies are exposed to rate rises in three ways. Between 30% and 45% of their debt has floating rates that reset as soon as the Fed moves, versus under 10% for S&P 500 companies. A lot of their debt comes due in 2026-27 and must be refinanced at today's higher rates. And about 40% of them do not earn enough to cover their interest payments comfortably. If the Fed hikes, this index should be hit harder than larger, profitable companies — which is why Singh pairs the bet against it with a purchase of the S&P SmallCap 600, which screens out loss-makers.
SK Telecom is South Korea's largest mobile phone company. It also owns about 3.7 million shares of Anthropic, bought as an early investment. For ordinary investors who cannot buy shares in a private company, SK Telecom was a back door into Anthropic.
When Singh looked at it near $19-20 a share, the market was valuing SK Telecom as if it were just a phone company, with the Anthropic stake almost free. That is the kind of mispricing he looks for. The shares have since doubled to $38-40, and at that price, by his arithmetic, the market is paying about $11 billion for the phone business plus a value for the Anthropic stake that implies Anthropic itself is worth about $1.7 trillion. The free part is gone.
With Anthropic's chief executive now talking about slowing AI development just before a planned stock-market listing, Singh thinks the listing will be delayed and doubts Anthropic will be valued above that level soon. Owning SK Telecom at the top would mean holding an ordinary Korean phone company at a full price, plus the risk of any cut to Anthropic's valuation. So he sold everything in pre-market trading near the high — and notes that if the shares fall back, he can always buy them again more cheaply.
Captured from the premium Special Situations Report weekly call (2026-09-13); transcript, report, agenda and weekly-deck PDFs in this folder, plus the analyst DCF workbooks (EVC, PRTH), Morgan Stanley's non-AI top-picks list, and the third-party research circulated with the call (BlackRock, Deutsche Bank, Fidelity, Franklin Templeton, J.P. Morgan, Morgan Stanley) — readable notes in transcript.md. Stances are the house's framing in this call, not price ratings. Not investment advice. © Special Situations Report for source material.