# Special Situations Report — Weekly Call (readable notes)

- **Title:** Weak non-farm payrolls (−23k vs +80k, −103k of prior revisions) take the September hike odds from 60% to 40% and trigger covering 25% of the short book; CPI (Aug 12) and PPI (Aug 13) as the week's events; the long adds — T1 Energy (TE) at $4.70, Uber (UBER) at $68.50, AppLovin (APP) post-earnings, Reddit (RDDT) at $137, TLT at the yield peak, Redwood (RWT) in the mid-$4s, Nokia (NOK) as an asymmetric AI option, CFTR preferreds above 9%, Datadog (DDOG) back after −17%, Alphabet (GOOGL) on the Jeff Dean $250B sell-off, Vistanc (VISN) at $10.35; the intraday Doximity (DOCS) short squeeze short covered for +20%; the Warner Bros. Discovery (WBD) arb update at 70% odds; the "so-far-so-fake" paper gains that turn 47.4% S&P Q2 EPS growth into 28%; Micron (MU) as 25% of index earnings growth; the joint US-Japan yen intervention and the carry-trade risk; Nano Dimension (NNDM) at a negative enterprise value; and the Chinese-AI-biotech licensing wave
- **Show:** Weekly SSR research call (premium subscriber recording — no public video)
- **Guest/host:** Jay Singh (founder, Special Situations Report; ex-Goldman Sachs)
- **Date:** 2026-AUG-09 (Sunday call; the deck/transcript title it "2026-08-09 - SSN")
- **Source URL:** the premium recording is distributed via the SSR Discord + email (no public YouTube video; no Discord permalink captured for this week)
- **Source PDFs (this folder):** `transcript.pdf` (the verbatim premium transcript), `summary.pdf` (the condensed report + actionable items + economic calendar + the stock-symbol list), `Agenda.pdf` (the full weekly strategy deck the call walks page by page: p3 the Virginia-governor/NextEra paragraph, p5 the weak-payrolls summary table, p6 the September-hike probability chart, p10 the Oman/Iran territorial-waters map, p11 the Delta Air Lines CEO stock sale, p12 the FactSet EPS-revision charts, p13 the one-time-gains tables for Alphabet and Amazon, p14 the 88% beat-rate chart and nominal GDP, p17 the 2025/2026/2027 S&P earnings lines vs the index, p19 the DigiTimes memory-sold-out piece, p20 the S&P's Nifty-50 → Four-Horsemen → FANG → Mag 7 evolution plus the hyperscaler free-cash-flow-to-equity chart, p22 T1 Energy and the Clearway contract, p23 the Vistanc share price, p25 the yen intervention, p28 Morgan Stanley's $1.2T cloud-capex forecast, p29 the hedge-fund tech-buying chart, p35-38 the daily market recaps for Aug 3-6, p40 Goldman's compute-token forecast, p42 Eli Lilly, p43 Palantir, p44 the record 4M S&P call options, p45 Shopify, p47 the Chinese biotech licensing deals, p49 the Jeff Dean departure, p50 token prices, p52 the Uber risk section, p53 the Nvidia-employee note on Meta's excess compute, p57 the UBS note on AppLovin, p58 Datadog, p59 the LLM token-cost index, p60 Atlassian, p63 the Doximity intraday timestamp). Also saved from the ten research PDFs he uploaded to the SSR Q&A tab during the call: `BlackRock August View.pdf` (BlackRock's multi-asset August outlook), `GS Asset Management Weekly Pulse.pdf` (Goldman Sachs' institutional flow/positioning update), `MS on Prediction Markets.pdf` (Michael Mauboussin's Morgan Stanley piece on prediction markets, equities and decision-making — one of the two he tells subscribers to read) and `Standard Chartered Weekly Update.pdf` (Standard Chartered's weekly asset-allocation update). Referenced but not saved here: the Morgan Stanley 3Q asset-allocation guide, the Deutsche Bank geopolitics note, Apollo's study on AI's impact on the labor market (the other must-read), Apollo's yen-carry-trade chartbook, and the HSBC weekly global asset-allocation piece; plus the August asset-allocation Excel workbook he re-uploaded to the Discord allocation tab in the Q&A.
- **Note:** readable notes auto-extracted from the premium PDFs; fillers (um/uh/you know/false starts) removed, wording otherwise verbatim from the recording. Section headers carry the recording's `H:MM:SS` cue. No public video, so the per-name table has no deep-links (the Ref column is the section time as plain text). Auto-transcript garbles corrected to the right entity and listed here: **"SSN" = SSR** (Special Situations Report); **"Visin" / "Visteon" = Vistanc Networks (VISN)**, the ex-CommScope remain-co whose Ruckus sale to Belden and ~$5 special dividend were covered on the 2026-JUL-12 call; **"Scott Pesant" = Scott Bessent** (the transcript spells it both ways); **"Extera" = NextEra (NEE)**; **"Wi-Fi" = WYFI (White Fiber)**, the 70%-owned Bit Digital (BTBT) subsidiary; **"Max 7" = Mag 7**; **"Chair Walsh" = Chair Warsh**; **"Amman" = Oman** (in the Rubio quote); **"Cash App" = Block/XYZ** (the reporting company); **"Brooking Institute" = the Brookings Institution**; **"Jasper Cole" = Jesper Koll** (the Tokyo-based banker who publishes the Japan Optimist); **"Brian Krinsky" (BTIG) = Jonathan Krinsky**; **"MarketAxess", sold by Nano Dimension to Stratasys, = Markforged** (MarketAxess is the bond-trading platform, not a 3D-printing asset); **"Crowd" = CrowdStrike (CRWD)** and **"Sail" = SailPoint (SAIL)** in the security-software list; **"Jangsu Hengrui" = Jiangsu Hengrui**; **"Melco Crown" = Melco Resorts (MLCO)**; **"Qwen3.8 Max" = Alibaba's Qwen 3.8 Max**; **"TET" = TE**, the T1 Energy ticker used on every prior call (the report repeats "TET"); **"Mohammad Bagher Ghalibaf-Degir"** and **"Vanguard"** (the UK maritime-security firm) left as spoken; **"Russ Dumba"** (the Citizens analyst on Backblaze) left as spoken. Internal garbles/inconsistencies in the source, left as spoken and flagged here: the arb-spread list is read as **"NSCU, NP, and WBD, PSKY"** (PSKY = Paramount Skydance, the WBD acquirer; the first two are unintelligible); Micron's **"$25 contribution is nearly 25% of the total S&P 500 earnings for the quarter"** is then restated as "25% of the *growth*" — the report and the surrounding argument mean growth; the 1Q27-vs-1Q26 estimate comparison stumbles as **"94.58 billion … well above first quarter '26 estimates of 80 of 86.69"** (these are $/share, not billions); Nano Dimension's stock is given as **$1.60** in the balance-sheet walk and as **$1.85** in the verdict ("buying 2.20 of net cash for 1.85"); Musk's SpaceX compute targets are stated as **"10 gigawatts and 5 gigawatts by the end of 2027"** in the same sentence; the SPX momentum stat is **"up 6% over four trading days"** against BTIG's "surged 5% or more in four days"; Redwood is added "in the mid-40s" and immediately self-corrected to **mid-$4s**; Backblaze's EPS beat is read as **"8 cents from versus 10 estimate, versus 2 estimated, sorry"**; Grab's "adjusted EBITDA was at 54% year-over-year" omits the verb; Nano Dimension is described as **"based in Illinois"**; and the call consistently frames the September Fed risk as a **hike** (60% → 40%) while the report's Actionable Items line calls it a "rate cut/pause." The transcript footer's "Date of Transcript: 2024-05-24" and the "2026-01-26" filename are stale template artifacts — the real call date is **2026-08-09** (Sunday).

---

2026-08-09 - SSR - Macro: Weak Non-Farm Payrolls, Upcoming CPI and PPI, SHOP Earnings, T1 Energy, APP Earnings, TEAM Earnings, VISN Earnings, WBD Arb Update

## Introduction and Document Overview - 00:00

Okay, happy Sunday, everyone. So if you jump into the Q&A tab for our call, you can see there are 10 PDFs uploaded. Number one, you have the strategy piece we have for this week, which goes through the very weak labor data from last week, goes through CPI, PPI for next week, the economic calendar, our asset allocation positioning update, the Iran war update, and several earnings updates and trade updates.

Then we have Morgan Stanley 3Q asset allocation guide, the Morgan Stanley piece on prediction markets by Mauboussin, BlackRock August view update, Deutsche Bank on geopolitics, Apollo's impact on AI and the labor market from this past month, Apollo on the yen carry trade, Goldman Sachs weekly pulse update, Standard Chartered weekly asset allocation update, and the HSBC weekly update. There's quite a lot of reading material there and I hope it's helpful to you.

So let's quickly go through our strategy piece. It is now in the SSR Q&A tab. I'll give you guys maybe 30 seconds to open it before I start just so that everyone has a chance. If you go into the SSR Q&A tab, open up the PDF, the call won't match it exactly, but it's a nice guideline, because we go through material rather efficiently for you guys.

Okay. So, just give you, there are a couple other people that joined the call. For you guys, just open the strategy guide we have in the SSR Q&A tab, so everyone can follow along. Have a couple new subscribers as well.

I really would read the Apollo piece on the impact of AI in the labor market. That was quite a nice read.

And the Morgan Stanley piece on prediction markets, it's not for everyone, but if you read it, it is quite in-depth, and it doesn't just cover prediction markets, it covers equities and decision-making as well.

All right, so now hopefully as everyone has everything open, we can jump right in.

## Economic Calendar and Market Expectations - 03:17

So in terms of the economic calendar, we'll talk about last week's major news, which was the weak payrolls data, which is why we ended up covering a number of our short positions, about 25% of our short book. We had been adding longs for the past couple weeks during the geopolitical issues that the market was fearing. And I think we are in a decent position now.

This coming week, I think is extremely important just given the inflation fears we've had with oil prices. And that is the CPI and the PPI data coming out on August 12th with a focus on the core CPI and the PPI on August 13th. Other really important news is going to be retail sales on August 14th and then the U-Mich sentiment on the same day.

The market currently expects core CPI to come in at about 2.5%, so lower than the 2.6% for July. And on PPI, the market expects that core energy number — sorry, ex-food and energy number, the core number — to fall from 4.7 to 4.1%. The market is viewing inflation as not being a major problem. So we will see very soon how that plays out.

## Recent Earnings and Trading Updates - 04:52

We had some very positive earnings this week. We had Team where we had a position, the stock was up 32%. We also had Shopify earnings beat, that stock did very well. And we made a really good intraday trade by shorting Doximity after a short squeeze pre-market. It was up at 37. We shorted it in the educational Discord, fell to 31, 32. So very easy 20% made. On the negative side, Vistanc reported, we barely had a position in that one. We just started to do work on it because they're going to pay a $5 special dividend, which they confirmed on their earnings. Because of an AI memory shortage, that stock fell like 12 to like 11. Intraday it fell to like 10:30. We added a little bit around 10:35. We're not going to make that one a big position because we're still worried about memory prices. And we added to Google after their chief scientist, Jeff Dean, announced his departure because it resulted in a 250 billion dollar market cap — that's billion with a B — loss. And it's just one data scientist. So we added to our Google after that sell-off. And then we added to T1 Energy which rallied a lot. We added to that at 470. So quite a lot to discuss on this call.

## Upcoming Earnings Schedule - 06:22

So if you go to the following page, we do own some Warner Brothers. And we'll talk about the arb section, but before we do that, we're at the end of earnings season and there's still a lot of companies that are going to report that are relevant to us next week. So let's discuss those. Berkshire Hathaway reported over the weekend and they bought back shares. So that's very important. We'll touch on that a bit during this call even though the PDF doesn't focus too much on that. Barrick Gold, where we own a position, ticker B, is reporting pre-market on Monday. We also have monday.com reporting on Monday. Perion, P-E-R-I, and then Embraer, the aerospace company, essentially the mini Boeing or Airbus, are reporting pre-market, EMBJ. After the close, we have AST SpaceMobile. And Rocket Lab, okay, so two of the space companies. Those companies have not done as well as the SpaceX IPO sold off. For those of you guys who don't recall, we were long SATS. We made 20% on that. We started to short SpaceX against it going into the unlocks, and we sold covered calls. And effectively, we made a little bit of money on it, on the way we set it up, and then we just took off the entire position. And it was good that we did because SpaceX has gone down about 50% since. And it's dragged down a lot of the space stocks with it.

So when will be the time to re-enter? A lot of the float is still unlocking, so we haven't added back SATS yet. But we're obviously monitoring that.

Now, after AST SpaceMobile, we also have Plug Power, Hims & Hers, which has been a super volatile name because of their essentially unregulated selling of GLP-1 drugs and their partnership with Novo Nordisk and now their lawsuit. So that's been a very volatile name. We also have GoPro, USA Rare Earths, which still doesn't generate any earnings, but that's going to be reporting and it's of critical national security importance to the U.S. Then we have Newegg. On Tuesday, we actually also own Sea. So on Monday, the name we own, Barrick is reporting. On Tuesday, the name we own, SE, or Sea Limited, is reporting. Great company. We published a model on it. We'll see how they do. We also have Cardinal Health reporting, Rackspace Technology. A lot of mid-caps reporting this week. After the close, we have Super Micro, which has had some accounting issues and an investigation. CoreWeave is reporting. They just did two deals last week. Lumentum is reporting, LITE. So there are a number of the AI companies and related companies reporting next week. Cava's reporting, fast casual dining next week. You also have Franco-Nevada reporting, which I think is going to be a very interesting one.

We also have on Wednesday, we have Nebius, which we did well buying during the AI sell-off and it rallied, but there was an insider at Nebius, we included in this PDF, who was talking about how Meta selling excess compute could put pressure on some of these guys, Nebius and CoreWeave. So it gave back some of the earnings. Also, Michael Burry, who lost money on this huge Palantir short squeeze, has targeted Nebius as a beta short, an AI beta short to hedge his portfolio. We'll talk about that. Then we have Brinker International pre-market on Wednesday. We have Amcor, WYFI. Okay, WYFI where we have a small position through BTBT, more of a speculative AI position. After the close, we have Cisco reporting, Cash App, you have Annovis, Allogene, Harmonic. On Thursday, we have Melco, MLCO. Melco Crown, we have Applied Industrial Technologies. We have JD.com, so Chinese tech starting to report, which is going to be very important to monitor as I believe that China is overdue for a stimulus announcement.

So that should be quite interesting.

After JD.com, you have Tapestry, Applied Materials after the close, the AI name. We have HeartBeam, PetMeds, Workhorse, The Metals Company, some irrelevant small names. And then Friday, nothing material. So, say the most important days this week are the CPI and PPI days. And for earnings, it's really Monday, Tuesday, and Wednesday, and maybe Thursday morning with JD. Then in the afternoon and Friday, you can take those off and enjoy your weekend.

## Special Situations and Merger Arbitrage Update - 12:11

In terms of the special situations update, in terms of large merger arb deals, Warner Brothers was topical on August 6th, following a slew of headlines including one, UK CMA approval, which the stock didn't rally much on, but it is a positive. Two, you had the UK cultural secretary's decision not to intervene in the merger on public interest grounds, which was also a positive, because one of the big risks is that UK is just more left-leaning and they tend to object more when it comes to deals. So that was a big risk which is no longer a risk.

Then you had Q2 earnings which is okay. I wouldn't say it was great. I thought it was a little bit underwhelming. And then four, a statement from Regal Cinemas CEO backing the merger, which is very positive because it's one of the biggest movie theaters in supporting the deal. Regal's CEO highlighted Paramount's merger commitments around theatrical releases, as well as the uncertainty created for studios and other stakeholders while the deal is in limbo. Support from Regal, the US's second-largest movie chain, follows a recent op-ed from the CEO of AMC, which is the largest movie chain in support of the transaction. According to some estimates, the two chains account for about 40% of U.S. theater viewership and approaching 50% for blockbuster films.

Overseas, the UK CMA cleared the transaction in phase one in terms of the movie studio market. The regulator highlighted competition from Universal, Disney, and Sony, as well as from a range of smaller studios. Lastly, on the second quarter earnings, Warner Brothers' sales missed consensus by about 5%, but EBITDA was in line on cutting costs, so their profitability was in line. At first glance on the top line, it looks like most of the miss was the studio segment, with slight weakness in networks and in line in the streaming segment. You can see that NSCU, NP, and WBD, PSKY are still the widest large merger arb spreads that are out there. And we think that the downside for Warner Brothers versus the upside and the probability of close, which we think is 70%, still makes that spread attractive.

## Activist Investing and Other Corporate News - 15:30

In other news, Virginia Governor Abigail Spanberger announced that she will be intervening in the proposed merger of Dominion and NextEra, which is a little bit concerning. APGE announced it has received HSR approval, so no antitrust risk there. BHF reported earnings and is still seeking approval from the insurance regulatory authorities in New York, Massachusetts, and Delaware. And Starboard revealed a stake in Shake Shack, which sent Shake Shack up 12% on that activist position.

There's some details on the Regal Cinema letter below. Virginia governor to intervene in the NextEra deal. There's a paragraph, second paragraph on page three, for anyone in the deal, you can read that. So Shake Shack rose 12% after the head of Starboard Value said that the activist investment firm has built a large stake in the burger brand. Starboard CEO Jeff Smith basically said on Wednesday on Bloomberg TV that his stake is worth several hundred million dollars. So I personally didn't think that Shake Shack was that cheap, but the stock had sold off quite materially this year.

We'll see if they publish a presentation, but usually when you get a big rally like that, I wait for it to calm down, then I buy the shares when they're kind of unchanged on the pre-announcement level. And then after that, I just try to buy the shares ahead of the activist actually doing something like changing the board, forcing a buyback, forcing them to sell assets, shoring up liquidity, addressing the capital structure. You kind of wait for the shares to go back to their unaffected price, which is the price before the press release announcement or the interview announcement.

And if you can get the shares there, you have a much better risk reward. DigitalBridge requested FERC to authorize its transaction no later than August 12th. At this time, the DigitalBridge public utilities could receive all remaining regulatory authorizations by August 12th. Therefore, DigitalBridge public utilities respectfully request that the commission authorize the pending transaction. So they're still waiting on that deal to close. EA announced the completion of their acquisition by PIF, Silver Lake Private Equity and Affinity Partners. So as you guys know, Electronic Arts is one of the biggest video game companies in the world, and they were acquired and that deal closed in the middle of last week, by private equity. I think that they got a relatively steal, because they didn't pay a big premium. They bought it at a 53 billion dollar market cap. The company had a net cash position. It was forecasted to do about 3 billion of EBITDA over the next two years. On a free cash flow yield, it was quite skinny, but if they add debt to it, it was I think only like a 10-15% premium on where the shares were trading last year.

Okay. So let's talk about some of the updates we made to the portfolio. So recent incremental adds, we added TE, T1 Energy at around 470, closed near six bucks at the end of the week, so that was a great trade. We'll see how it continues to play out. They signed a very interesting deal intra-week, which sent the shares up 15% and they continue to rally. Uber, we added around 68. We just think that they can really grow EBITDA from 10 to 15 billion, which is consensus. The stock is actually quite cheap, trades at 12 times forward. App, APP, after its earnings miss, we added to that. We think that over the next two years, if this was truly a hiccup like the management was saying, and on-app advertising is less disruptible by AI, that the stock could double over two years from the mid-300 level we added. Reddit, we added at 137, it's already at 160. That was a few days ago. We added to TLT and we actually timed the peak in the 10-year almost perfectly for our TLT add. We added more to Redwood after it went into the mid-40s and then it closed kind of in the high — sorry, mid-4s — and we first added in the 3s, it rallied to five, above five. We added now in the mid-4s and it ended in the high-4s at the end of the week. We think it's still a compelling name trading at a discount to book, 15% dividend yield. We added to Nokia as an interesting AI option play, just like Intel was trading very, very cheaply and then it kind of pivoted to AI and CPUs and silicon and then it rallied. We think that Nokia is kind of a sleeper name where it does have a new AI story and it's trading relatively cheap to that potential TAM. We added to CFTR, the preferreds, which is over a 9% yield. We owned Datadog, it doubled in the spring. We sold it and we bought it back after earnings because it sold off 17% on positive earnings. Cybersecurity is, I think, less affected by AI, which was the thesis that worked. Google, we added this week after the key engineer loss took it down 5%, like a 250 billion dollar market cap loss for one engineer, seemed kind of silly. The stock was a little bit extended, so we didn't add a lot, but we added to it on that engineer loss. And then we did that short-term trade in DOCS, Doximity. There was a big hedge fund blow-up and a massive short squeeze. It was up almost like 240% pre-market. It was still up 80% pre-market when we shorted it and then it went all the way down to being up like only 50%. So we made over 20% on it intraday. And we also covered 25% of our hedges on the weak jobs data, which we'll talk about that came out this week.

On page five, let's talk about weak payrolls. So, like I've been saying for the last several years, bad news on the economic front tends to be good news in terms of the market, unless it's terrible, or until unemployment goes above 5%. But right now, unemployment is only around 4.1%. So any bad news means that the Fed is less likely to hike. And so this past week, we had negative 23,000 jobs versus an expectation of positive 80,000 jobs. Hourly earnings were only up 0.1% versus 0.3%, which means no major wage inflation risk. You can see that summary of the data on page five. And then you can see the consensus number in the table below. Also, the prior revisions were very negative.

So you can actually see that the Biden admin, people were arguing was doing the same thing, they would boast about good job numbers and then revise them down. This administration's not much different in the last three months. So if you look at what happened, they revised jobs down very dramatically. May was revised down 66,000 from 129,000 all the way down to only positive 63,000. So 66,000 revision, and in June revised down by 37,000. Together, those two months were revised by almost 100,000 jobs. So that is very suspect. So when you combine May and June, the prior two months' jobs numbers were revised down by 103,000, which also makes it less likely that we see a Fed hike in September. In fact, if you look at the following page, the probability of a Fed hike in September fell from as high as 60% on August 1st all the way down to 40% by the end of last week, which is one reason why the markets were a little bit buoyant on Friday.

## US-Japan Economic Relations and the Yen Carry Trade - 22:47

The U.S. has been selling euros to buy yen, which surprised Europe, but was obviously well-received in Japan. Japan was losing control of the yen, or there was a perception of that, due to inflation risk and their large debt load. And the U.S. went and supported them because a carry trade, which we'll read about in one of the attachments I put in the Q&A, SSR tab, the Apollo chart book on the yen carry trade, it's very important when it comes to risks. So we already had the carry trade blow out, a short-term carry trade, but there's still a medium-term carry trade, which is essentially investors borrowing in yen, as the yen has been weak because of 200% debt to GDP and low interest rate differentials. So borrowing in cheap yen due to low interest rate and, for example, buying the Mexican peso has been a very popular carry trade and no one wants those carry trades to unwind because it usually means risk off for all risk assets when those unwind.

Obviously, Bessent worked for Soros and Druckenmiller, and he's well aware of how carry trades can negatively impact the market, especially when investors are so levered as they are right now. So that intervention was quite something to look at. If you look at the yen outlook from carry trade to fiscal worries, the Apollo piece, basically, the USD/JPY had been driven by interest rate differentials from 1995 to April 2025. So this was like a 30-year carry trade. And then when the trade wars began in April of 2025, volatility increased, headline risk increased, it was more difficult to manage carry trade. So the overall carry trade has been shrinking since then, but it's still a major risk in the market. And now USD/JPY is driven more by fiscal worries than interest rate differentials.

According to Apollo, USD/JPY stopped following interest rates after liberation day. And what you saw after that was the U.S. 10-year minus the Japanese 10-year versus the USD/JPY started to move in opposite directions, which is interesting. They also have some really cool charts about rising activism in Japanese equities on page 6 and 7, corporate profits for all industries have been rising in Japan, medium-sized companies have been doing much better than the past. Inflation is still high over there, which is why they had to hike rates and inflation expectations have been rising. Japan's holdings of U.S. Treasuries have been soft, but not falling as fast as China, that's on page 19. So there are some very interesting charts in here that I think you should read, and bond flows in and out of Japan, starting on page 29. Very interesting chart book, as always.

## Iran War and Geopolitical Update - 25:54

Now in terms of the Iran war update, there's been a lot even over this weekend, which we'll update you on. So, our view is that Trump — one of the reasons why we were a little bit more bullish on the market was based on what we were reading and speaking with experts on the subject, because I certainly am not an expert on Iran. We concluded earlier in the week that Trump had no choice but to back off the war, not just due to midterms, but also to avoid a hike in September because of the pressure some dissenters had put on him. So, that was our view and I think it's played out that way. I think Trump isn't really putting pressure on them even though Iran says they're not interested in making a deal so far. We think that there's a big risk that after the midterm elections, we just go nuts on Iran and I don't know how that will play out. But I think in the near term because Iran's economy is not doing well, they're hoping to put economic pressure on them again. They're waiting for some diplomatic progress, but we'll see how that plays out.

On August 4th, Qatar said that language had been drafted around an Iran-US temporary ceasefire. The Qatar foreign ministry spokesperson had said that language for a possible US-Iran interim deal had been drafted. This was then denied. What is this, another fake ceasefire in progress? So the ministry spokesperson, Majed Al Ansari, didn't provide any timeline for a potential deal, which is why it felt like a weak discussion. He described that the focus is on a short-term resolution that would actually restart the US-Iran talks and it wasn't a peace treaty on its own. He admitted that nothing in the books was actually written when it came to timing or putting anything on the calendar of the talks. This was enough though for regional media to say that there was progress being made and the market started to rally and oil fell. So Brent crude fell below 80 for the first time since mid-July. The day prior, President Trump continued to address Tehran in threatening language. He said before reporters in the Oval Office that I want to give them every last chance before decapitation. Very tough to do what we have planned, still planned. We'll see what happens, but it is a very, very tough thing to do. So Trump was kind of giving mixed messages. In the end, he said, I think I'm very proud of the fact that I will give the people a chance.

Just on the heels of the above draft deal headlines, in a carefully timed CNBC appearance, Washington brought out its heavy hitter to re-anchor the administration's narrative, seeking to assure markets that a diplomatic deal with Tehran was not only real but imminent, but we didn't see a deal all week. They said that we may have an Iran deal tomorrow to open Hormuz, according to Bessent. Nothing really happened there. And then if you fast forward, a vessel had been struck in the Strait of Hormuz the following day near Oman, according to the UK maritime security firm, Vanguard. One crew member was missing.

Looming large over all of this is the fact that the Iranians have still denied that there are informal talks with Washington, which Trump the day prior described as a ruse, venting his frustration while insisting that the Iranians have been engaging. The Hormuz blueprint, as it basically stands — Reuters said that Iran would gain full control over vessels entering the strait under a temporary plan being discussed in Oman. But that Oman plan did not include the US, which we later found out. Bessent has a clip on Twitter you can find where he says, we've seen Trump last week threaten what would have been the largest military campaign since World War II against the Iranians. And now because of that, we're in talks. Again, there have been no talks. So you can see the timeline, starting March 21st, Trump threatened to hit and obliterate Iranian power plants unless Hormuz reopened within 48 hours. Then on March 23rd, delayed strikes for five days, citing productive conversations. On April 7th, he warned a whole civilization will die tonight. Hours later, a two-week ceasefire was announced. On April 21st, he said to be bombing if talks failed. Later that day, there was an extension of the ceasefire. On May 17th, he warned the clock is ticking. On May 18th, he then paused planned strikes after a request from Qatar, Saudi Arabia, and the UAE. On June 11th, he said that the US would hit Iran very hard tonight. Hours later, he cancelled the operation, saying the deal is near. And on August 1st, he said the US was locked and loaded, later called off the attack to pursue a deal with Iran, which still hasn't happened.

On August 6th, oil was up 3.5% as Iran was seeking to bypass the US and set a temporary 60-day deal with Oman for obviously their Asian customers. And then Trump on the same day was saying that oil prices were falling. And then today, Iran says it has rejected US talks as Trump displays patience for this Hormuz deal. So Iran and Oman are close to a pact, basically to establish a shipping route through the strait, but a final deal has not been reached. Iran basically has said they haven't been speaking with the US, citing repeated violations of the short-term interim peace agreement, and instead is exchanging messages through intermediaries. The US is taking a patient approach with President Donald Trump saying that the US can wait for Iran's economic suffering to soften its stance as Iran faces huge inflationary shocks and a lack of money coming in. Now, as long as the American violation of the memorandum of understanding, which I don't know if they — I mean, Iran didn't even abide by that, so I don't know what the situation is. But they're saying the US didn't abide by it, so until the US makes amends for its recent bombings, they're not going to talk. That didn't seem to bother Trump who on Sunday told Axios in a short phone call that the US is now low-keying it. We're only semi-negotiating with them, he said. You can only lie for so long. We are just watching Iran with its huge inflation and the fact that they have no money. Sunday's remarks follow weeks of Trump threatening massive strikes on Iran only to pull back saying he wanted to give negotiations a chance. Iranian year-over-year inflation has reached 77%, so there is truth to that comment, according to Iran's Central Bank. Tehran on Saturday upped the ante by renewing a hefty list of demands on Washington as conditions for a full reopening of Hormuz, indicating any immediate respite for oil and gas supplies may be limited. The Strait of Hormuz, through which one-fifth of the global oil and gas is shipped, has become a key factor of the war that the US and Israel started on February 28th. Trump has demanded the resumption of open navigation for months. And now with the deal being elusive, Iranian-backed Houthi rebels claimed that they struck a Saudi Aramco Jizan oil refinery in the kingdom. Saudi authorities earlier reported a fire on the facility but said that it was quickly extinguished and no injuries had been reported.

J.D. Vance on Saturday suggested that there had been recent progress in talks, but he said we remain skeptical. So if he's skeptical, then I don't know who actually believes there is progress. The question, he said, is whether Iran will be able to give the things that are necessary for us to be happy, for us to feel like we've gotten what we need out of this particular engagement, according to Fox News. While Washington has repeatedly suggested it's part of the negotiations over management of the Hormuz strait, which Iran disputes. So some intermediary countries like Oman are trying to re-establish the ground for negotiations. So is Pakistan, by the way. Oman, which borders the strait across from Iran, said that the talks with Tehran were progressing in a positive and constructive atmosphere in a post on X or Twitter, authorities called for a halt on actions in the strait to give space for diplomatic efforts. Iran demands that the US lift its naval blockade on Iranian ports, withdraw military forces from around Iran, remove sanctions, release its frozen assets, and pay compensation for war damage, according to Mohammad Bagher Ghalibaf-Degir, a hardliner who heads the Supreme National Security Council. Traffic through the Hormuz Strait has been largely blocked since the start of the war, but some ships have continued to ferry cargoes as part of a shadow program involving US military help. So oil prices were volatile last week, swinging between gains and losses with Brent eventually settling back above $83 a barrel as traders weighed prospects of a deal. You can see where Oman is and Oman's territorial waters and where Iran is on page 10. And then on page 11, this is an interesting sell by Delta Air Lines' CEO who exercised stock options and sold 19 million of his stock holdings after a 30% run-up in three months. And it could be a strategic sell in his mind given that the market is anticipating an Iran war peace, and Delta Air Lines stock has rallied because of it. He doesn't seem to believe that it's a good risk reward at this moment.

The Trump administration is also drafting a ban on U.S. imports of new models of Chinese data center components, according to Reuters. The proposed FCC rule expected this year aims to prevent espionage, malware, and supply chain risks. The move could hit China's Zhongji Innolight, which benefits U.S. rivals such as Coherent and Lumentum, which is interesting.

## S&P 500 Second Quarter Earnings Review - 36:05

So second quarter earnings, now that we're almost over with earnings season, has been extraordinary by all measures. The pace of growth on a year-over-year basis has been off the charts and has wildly surprised analyst estimates going into the quarter. The second quarter headline EPS for the S&P 500 is tracking at 47.4% versus a 23% expected at the start of reporting season. The two charts below show how much earnings estimates have jumped as we've progressed through the quarter and how positive revisions to EPS estimates have only been larger one time since 2006. You can see that graph from FactSet on page 12 and then quarterly revision paths. And you can see how the second quarter of 2026 has just seen an explosion in earnings growth. A lot of it has been Mag 7 names which have seen lower free cash flow, and which has resulted in higher earnings as they're also borrowing more money to buy more GPUs and selling compute. So earnings are up, but the ROI isn't clear.

## The Role of One-Time Gains in Earnings Growth - 39:30

Now on page 13, as I've seen below, the 47.4% EPS growth is the fastest pace of growth since the second quarter of '21, when earnings were rebounding from the COVID reopening trade. Before we start off, all the caveats that I want to give with this explosive earnings growth, we have to acknowledge that there's a healthy breadth to how earnings are beating expectations.

The next chart shows how the current beat rate of 88%, which means that 88% of companies are beating sell-side consensus estimates, is at a record high and well above the long-term average. Further, a robust level of beats is happening across sectors, which is important to flag as we talk about the ultra-dominance of Mag 7 and tech within the absolute earnings growth levels. You can see this graph on page 14. We must note that the source of this broad-based strength in earnings is thanks in part to strong nominal economic growth in the quarter as well. The chart below shows the acceleration in nominal GDP in the second quarter of '26 to 6.5% year over year, or growing at an 8.1% annual rate. Though this nominal GDP acceleration in the black line was driven by higher inflation, not higher real growth, which is the blue line, the translation into earnings is that earnings are a nominal calculation. So earnings growth benefits from rising prices and inflation.

So the summary thus far is the second quarter of 2026 has delivered incredibly strong headline earnings growth with robust breadth across sectors supported by a resilient U.S. economy. However, we now have to introduce our caveats.

A lot of the earnings growth was due to one-time gains. Not all of it, but a decent percentage. The first and second quarters of 2026 have benefited from unusually large non-operating unrealized gains on the investments at both Alphabet and Amazon, as they invested in Anthropic, for example. These gains are so big that if you remove the impact from these paper profits, the year-over-year earnings growth for the second quarter falls from 47% to only 28%. But 28% is still a massively impressive growth rate, given that we are now emerging from a downturn.

It is helpful for us to take some time to understand what these gains are in order to better anticipate how earnings could progress in the coming quarters. Buckle up for some basic financial accounting. Now, both Alphabet and Amazon have large investments in companies that have seen significant valuation increases in the first half of 2026. Both Alphabet and Amazon have investments in Anthropic, which had its valuation vault from only 183 billion at the end of 2025 to 380 billion in the first quarter of 2026, and now 965 billion, almost 1 trillion in the second quarter of 2026, as it has raised incremental rounds of capital. Alphabet has an investment in SpaceX, which had its valuation launched from 800 billion at the end of 2025 to 1.25 trillion in 1Q26, to an IPO valuation of 1.77 trillion at its IPO in the second quarter, and it's since down from that.

Without getting too technical, these changes in valuation for Anthropic and SpaceX get reflected as non-operating unrealized gains on the income statement, because the companies didn't actually sell any shares. And these gains have been huge. So you can see in the tables below, some key statistics from the first quarter of 2026 and the second quarter of 2026 for Alphabet and Amazon that illustrate these gargantuan gains. As an example, Alphabet reported $9.11 of earnings in the second quarter of 2026, up nearly 300% year over year, but only $2.88 of those earnings came from its operations. The rest came from gains in these share prices. So like six of the nine came from gains in the shares that they own. Now, their earnings were still up 30%, but the rest of Alphabet's earnings, 6.23, came from gains on investments, primarily Anthropic and SpaceX. You can see a similar dynamic with Amazon in the table below.

We are already starting to see how fleeting these paper gains can be as we see the valuation of SpaceX plunge in public trading. SpaceX is now down over 40% since the end of June. So if this valuation were to persist through the rest of the quarter, Alphabet would have to recover a mark-to-market loss on this position.

Further, in order for more gains to be recognized on an investment in Anthropic, the AI company would have to raise incremental capital either through private markets or an IPO. If Anthropic does not raise capital in the third quarter, for example, then neither Alphabet nor Amazon will recognize gains for the current quarter. As we look to 2027, we do not think it is likely that we will see the magnitude of gains repeat, which means that the reported GAAP EPS for Alphabet and Amazon have the potential to decline in the first quarter and second quarter of 2027.

Given how much these gains have boosted total S&P 500 earnings — remember the second quarter of 2026 earnings would be 20 percentage points lower without the gains — these gains are likely to act as a Pierce the Veil-esque dangerous dead weight for 2027 growth statistics. Interestingly, top-down Wall Street analysts do not seem too concerned for now that a large chunk of the first quarter and second quarter earnings are unlikely to repeat in 2027. The chart below of technical spaghetti shows, for example, how first quarter '27 estimates of 94.58 billion are well above first quarter '26 estimates of 80 of 86.69, that include these significant gains estimated to be $6 of EPS in the first quarter of 2026. When you strip out these gains, the implied growth rate for first quarter next year jumps from 17% to 27%.

As we are learning in 2026, never say never when it comes to supernormal earnings growth, but this estimate sure does set a high bar to kick off next year. And it's likely that earnings do miss next year when it comes to Big Tech. The high bar for 1Q27 raises the question as to whether or not there's much upside left to the 2027 full year earnings estimate, which now sits at $405 a share or 13% growth on top of 2026's stellar 30% growth already. As we near 2027, the equity market is likely to become increasingly sensitive to changes in these forward earnings estimates, meaning downward pressure on 2027 earnings could usher in more volatility, while upward revisions could support continued price gains.

## Memory Sector's Impact and Circular Financing - 45:25

You can see that S&P earnings for 2026, 2027 on page 17, and the stock market in white. So 2027 earnings are in the purple line, 2026 earnings are in the blue line, and 2025 earnings are in the red line. No discussion about the prospects for 2027 earnings growth over 2026's stellar numbers would be complete without acknowledging the massive contribution that memory stock earnings are making to overall S&P 500 earnings this year. The chart from FactSet below shows how Micron not only is the largest contributor to the S&P 500 earnings in the second quarter, but its $25 contribution is nearly 25% of the total S&P 500 earnings for the quarter. So Micron is 25% of the growth of the entire S&P 500 for the second quarter. Given the industry expects a continued shortage of memory supply compared to ravenous AI demand, Micron is expected to double earnings again in 2027, but this 100% growth is a meaningful deceleration from 2026's expected 850% growth, which is not a typo. The continuation of this memory and tech infrastructure cycle is highly dependent on the hyperscalers' willingness and ability to spend aggressively on physical AI infrastructure and memory.

Thanks to negative free cash flow, meaning hyperscalers in aggregate are set to spend more on capex than their operating cash flows in 2027, hyperscalers' ability to spend is partially dependent on debt and equity markets that are helping to fund the spend, which is why CDS for companies like Oracle is blowing out, which is the credit risk the market is pricing for them raising debt. For now, capital markets appear wide open for funding hyperscaler investments, but at an increasingly higher price.

It is not just the external financing that is important to consider. It is the circular financing that must be appreciated. So consider these statistics in light of last week's news that Nvidia is considering a 250 billion dollar backstop of OpenAI's infrastructure spending, which they're spending mostly on Nvidia GPUs. OpenAI makes up 40% of Microsoft's cloud backlog and currently accounts for 70% of Microsoft's AI revenue. While Microsoft is 22% of Nvidia's revenue and Nvidia is 17% of Micron's revenue. How circular is that? All of this to say that a key real operating — not like those so-far-so-fake paper gains — reason why the S&P 500 earnings have been able to deliver such powerful growth in 2026 has been the ultra-powerful memory trade. Memory remains a highly cyclical, highly competitive industry, so a deceleration in AI capex, of which we're seeing no evidence at this time, would result in an amplified downside in these earnings, which would push the Nasdaq a lot lower.

So this circular conclusion, overall, even if we remove the one-time gains for Amazon and Google for the first quarter and second quarter of 2027 earnings, the underlying operating earnings growth rate seems highly impressive. There has been breadth to these earnings helped by nominal GDP, but the dominant driver of the large upside to estimates has been the so-far-so-fake paper gains from Amazon and Alphabet, helped by the memory super cycle. These are all important features to appreciate as we think about the prospects for 2027 growth, on which the market will grow increasingly focused as this year enters its final months. The so-far-so-fake one-time gains plus memory earnings surge in 2026 do create a notable tough comparison for 2027. I think after midterm elections in November, the market will start being worried about 2027 earnings. So we have another few months of a range-bound market.

But right now, I'm worried about 2027 prospects. And for now, 2026 and 2027 estimates continue to climb higher, giving no signal that there's any concern about over-earning, unsustainable growth, or tough comparisons this year and next, even though the numbers do seem that way. The resilience of the U.S. economy is a supportive factor for this optimism. Arguably, it's the continuation of the external and circular financing of the AI infrastructure boom that will matter far more over time. You can see the S&P 500 year-over-year EPS growth — so this is including Amazon and Google paper gains. You see Q1 at between 25 and 30%, now 47% for Q2. Even when you strip out the gains, it's 28%. So this Q2 earnings growth is just unreal. Now, Micron, Samsung, and SK Hynix have sold out all of their 2027 memory chip capacity, which was an incremental positive over the past week. The market didn't really put a lot of credence on it. SanDisk, Samsung, Micron's current annual NAND capacity has been fully sold out through next year. Customers are now being allocated only 60 to 70% of the volumes they initially requested. Industry insiders pointed out that 2027 will enter the most severe moment of memory shortage.

Allocation amounts are mostly decided, but the final shipment pricing will be determined closer to delivery. According to DigiTimes, it's hard to see those memory oversupply claims in early to mid 2027 when they're all sold out of capacity already. So you can read this piece on page 19 from the original source. And on page 20, you can kind of see how the S&P 500 has evolved. The Nifty 50 names, Coca-Cola, IBM, Xerox, McDonald's, Disney, Johnson & Johnson, and others were the biggest rallying names back in the '60s and '70s. Then you had the Four Horsemen, IBM, Digital Equipment, HP, and Intel in the '80s. And the '90s, you had the Four Horsemen of Tech, which is Microsoft, Intel, Cisco, and Dell. In the early 2000s, it was Big Tech, Microsoft, Apple, Google, Amazon. And then that continued because in 2013, it was called FANG, Facebook, Amazon, Netflix, Google. Then 2017 was FAANG with two A's, Facebook, Apple, Amazon, Netflix, Google. This was after Carl Icahn went activist in Apple and Apple started returning billions and billions of dollars in share buybacks and the stock did quite well. And in 2021, it was MAMA, Meta, Apple, Microsoft, Amazon, and Alphabet. And now since 2023, it's been the Mag 7, Apple, Microsoft, Nvidia, Amazon, Alphabet, Meta, and Tesla. I would say that AVGO is a much better company to have there than Tesla, but it is what it is. There's only one hyperscaler left with positive free cash flow to equity right now and that's Microsoft, in the light blue in the chart on page 20. Everyone else is kind of in cash burn mode.

According to Morgan Stanley, which is contrary to Michael Burry's view, Microsoft and security companies — those two are very obvious, so Microsoft and digital security companies like Datadog — are winners in this AI trade, and Neoclouds also, they think, seem to win in every case because lease rates for GPUs continue to go higher. So as long as that's true and Meta doesn't flood the market, which is the biggest risk, they have effectively three scenarios: one, closed models win, two, hybrids win, and three, open AI models win. So nothing to do with OpenAI, but these cheap Chinese models that people customize. And so across all three categories, Microsoft tends to win on the cloud. On the model providers, you have a lot of competition coming from MiniMax, Knowledge Atlas, Baidu, and Tencent, so there might be increased competition there. On the security software side, you have PANW, CrowdStrike, Fortinet, Okta, and SailPoint as well. SAP, Now, and Shop. Shop has done well. We own that one as well this past week. And then on edge devices, you have Dell, HPQ, and Apple. But the infrastructure guys should benefit regardless of whether it's closed models or open models that win. So, just something to consider. And hyperscaler debt issuance by quarter started to be really nuts. I think a lot of these companies have raised over 100 billion of debt to fund these AI data centers and that excludes all their off-balance sheet debt, which is non-recourse.

Over the past week, something that was surprising was that Amazon's Jeff Bezos sold another 4 billion of Amazon shares. It was about 15 million shares with an aggregate market value of 4.07 billion. So this was listed as of August 3rd, but the statement came out after, so nobody got to trade on it, obviously. On page 22, we can talk about adding to T1 Energy. So this is a solar equipment manufacturer which announced a contract to supply independent power to Clearway Energy Group for 641 megawatts of solar modules. So T1 emphasized commitment to increasing the domestic content of its modules, noting that domestic modules with domestic cells are highly valued given rising uncertainty around trade and tariff policy. Obviously, before, a lot of the polysilicon came from China and they're trying to change that now.

So, T1 Energy expects to offer modules with greater than 60% domestic content in 2027. Solar panel manufacturers are awaiting the pending Section 232 decision, which would determine whether the President can impose tariffs or import quotas on polysilicon, a fundamental raw material to produce solar cells. So customers such as Clearway are seeking high domestic content modules for building critical momentum for a made-in-America supply chain according to T1's CEO, Dan Barcelo. So that's actually a positive thing for the company. You can see the stock kind of bounced off the sub-four level back to 5.85 in the first week of August below.

## VISN Update and Special Dividend - 56:11

Then on VISN where we only have a very small position, we added five bips again at 10.35. They reported Q2 results last week. Effectively, net sales were 319.6 million, down 1.4%. That wasn't really the bad part. They also announced a $5 kind of special dividend. That's going to be paid out later this year. But the issue was margins. So GAAP operating performance swung to an operating loss of 8.9 million compared to 7.8 million in 2Q25.

They are going to be returning a $5 dividend to investors. So the Ruckus sale was completed, which is positive. On July 1st, 2026, Vistanc finalized the sale of Ruckus for 1.85 billion in cash. The special cash distribution follows an earlier $10 payout from a few months ago. The management has now announced a new $5 per share special cash dividend to be paid in August 2026. And so there's been 3.4 billion of total return from this company. Management expects to have earned about $15 a share, or 3.4 billion in total to equity holders from the different asset sales they've done. And it now has a debt-free balance sheet. Proceeds were utilized to repay all long-term debt and redeem all Series A preferred shares, leaving the company with zero in outstanding debt or prefs and over 2.5 billion of book equity as of this mid 2026.

In terms of the full year outlook reduction, the reason why the stock fell was that management lowered the full year 2026 adjusted EBITDA guidance for the core Aurora segment by 25 million, from 225 to 250 to 200 million to 225, specifically because they had a memory shortage. So they weren't able to put down as many boxes to upgrade that coaxial cable signal. So that was a big issue. Margins are being squeezed by sharp memory chip price inflation, supply availability constraints, difficult prior year licensing comparisons.

And temporary stranded costs. So index rebalancing. There's also some near term trading that has impacted Vistanc's removal from the Russell indices following the shrinkage of its market cap position. So it could be technical in nature. You can see the share price on page 23. So we still have a decent average cost there.

## Japanese Yen Intervention and Market Implications - 58:58

The Chinese real estate market has fallen to its lowest prices in at least 20 years. Japan's Ministry of Finance makes a habit of intervening to support the yen. So this is interesting. Over the past four years, let me just recap this. Over the past four years, Japan's Ministry of Finance has made it a habit of intervening to support the yen. It hasn't helped. Last week's episode felt different. It produced a big turnaround and this time the ministry had help from a significant quarter, the US Treasury Secretary Scott Bessent. With the currency near its weakest in four decades, the intervention led to the strongest level since May, since buying the yen in September of 2022, the first such operation since 1998. The ministry has spent at least a quarter trillion dollars without stemming the tide, without actually strengthening the yen. The conservative figure does include last week, pending the official amount.

This matters beyond Japanese shores because the yen carry trade, the strategy of borrowing in yen and parking in a higher yielding currency like the Mexican peso, has been an ultra reliable source of profits. Over the past five years, it has strongly beaten the S&P 500 total return. A big yen appreciation might prompt the carry trade to unwind, as it did to chaotic effect two years ago. Last week's move has already jolted the carry trade out of what has been a startlingly steady upward trend.

Now, the upward trend here is actually a weakening, so I don't know what they're talking about here. But carrying on regardless, you can see the Mexican peso / Japanese yen carry trade total return. So that's in blue, so it's outpaced the S&P. So these actions are growing more frequent precisely because they are ever less impactful. So those betting against the yen know that the ministry's resources are finite and that Japan is in a poor fiscal position. Four years ago, the market reportedly viewed around 145 yen per dollar as Tokyo's line in the sand. And before the latest intervention, we were well above 160.

So Washington's uncharacteristic support, which seems to have included rate checks and selling euros to buy yen, is a potential big deal and sends a powerful signal. Mizuho Bank's Masayuki Nakajima argues that this reflects Washington's growing concern over the risk of further yen depreciation, or that rising yen volatility in Japanese bonds could spill over into US Treasuries. Now with that, very quickly, I'm just going to grab some water and I'll be right back. We're on page 25.

## US Involvement and Bank of Japan Policy - 1:04:22

Okay, back in action. So US involvement in the yen carry trade remains problematic. Robin Brooks of the Brookings Institution, a long-term skeptic of Japanese efforts to prop up the yen, basically says US participation raises more questions than answers, especially the very odd news that the US sold euros to buy yen. This kind of twist, in his opinion, undercuts the efficacy of US participation because it invariably will have markets wondering why the US didn't just fund the yen instead of buying out dollars. FX intervention is a confidence game. The last thing you want is to give markets any kind of reason to ask questions.

So what about the Bank of Japan? In June, it raised its policy rate to 1%, the highest since 1995. But it stayed on hold last week and one more hike is expected this year. Expectations have stayed relatively constant, even as the market has priced in increases for other central banks following the oil supply shock. The yen weakening has been almost exactly in line with the widening gap between the BOJ and the Fed. So this is contrary to what Apollo says, but Prime Minister Sanae Takaichi, who received a massive electoral mandate six months ago, has an adverse stance on policy tightening as she goes for growth. So this is putting pressure on the yen.

The timing and magnitude of the ministry's action with an estimated 52.8 billion spent on July 31st marking the single largest intervention on record, suggests a new level of resolve. What the Bank of Japan does next is critical. Not long ago, the oil shock — Japan is an energy importer — combined with a weak yen to suggest renewed inflationary pressure was a big issue. But the numbers remain ambiguous, with last year's big spike possibly attributable to a bad rice harvest, which resulted in higher inflation. The case for hawkishness isn't clear from the latest data. Even after the oil shock, headline inflation remains 2% in Japan. Governor Kazuo Ueda failed to provide meaningful forward guidance — and we're almost done with this — when the market was desperate for it. Jesper Koll, the long-term Tokyo-based investment banker who publishes the Japan Optimist newsletter, was scathing.

The fact that you, Governor Ueda, tell us with great confidence that Japan's inflation re-acceleration above 2% in the second half of Japan's fiscal year, got undermined immediately by its lack of action. So why exactly are you not hiking if you're so confident? Could it be that Japan's financial system is so weak that an accelerated rate hike would trigger a possible banking crisis? Koll, an optimist, and Brooks, a skeptic, agree that even this level of intervention won't work on its own. For Brooks, the yen is falling because Japan's high public debt prevents the country from allowing yields to rise freely. With bonds kept capped, the yen becomes the main symptom of a debt crisis that's getting papered over. Koll thinks that Takaichi's aggressive pursuit of growth will probably work, but agrees that the likely path of the yen is weaker. He suspects that the Bank of Japan is holding off because of worries about the secondary banking system, but that the new fiscal policy is almost guaranteed, ultimately, to be inflationary. Thus risks are still asymmetrically tilted to an even weaker yen, so I wouldn't be buying the yen here.

## Citadel on Retail Excess Unwind and Market Outlook - 1:08:01

Citadel's Rubner has been talking about the retail excess unwind of leverage, levered ETFs, margin. He's been talking about how the fact that it's mostly been unwound and that's one of the reasons why the market's bounced last week. Retail investors reduced risk, leverage normalized, market concentration declined, and many of the market's largest technical headwinds have begun to fade.

As a result, he said, we believe investors can increasingly spend less time focused on positioning and more time focused on fundamentals. Position sizes and risk budgets are likely to rebuild gradually, supporting a lower volatility grind rather than the type of V-shaped recovery that we saw earlier this year. The technical reset that we have been waiting for has largely occurred, and July didn't change the structural bull market, it just reset it, according to him. The S&P 500 information technology forward P/E ratio fell from as high as 31 all the way down to 20. At the same time, that's also because free cash flow fell for big tech, but that's another conversation.

## Morgan Stanley on Cloud Spending Surge - 1:09:14

Morgan Stanley, on page 28, actually sees cloud spending surging to 1.2 trillion for 2027, which is up 30% year-over-year and 170 billion above pre-Q2 estimates. The firm said that all four major US hyperscalers remain capacity constrained as AI demand continues to outpace supply. Alphabet or Google, Amazon, Meta all raised 2026 capex guidance, while Microsoft maintained its spending outlook and didn't increase it, which is why Microsoft's stock rallied so much, 13% in a day. Morgan Stanley believes that the consensus is still too conservative, forecasting cloud capex could reach even 1.4 trillion, driven by strong AI adoption, rising cloud demand, and multi-year customer commitments and improving returns on infrastructure investment.

## Cost of Intelligence and LLM Token Costs - 1:10:05

The cost of intelligence is plunging every day according to them, which is true. Despite relentless demand for agentic compute, thanks to China. Token costs are back to where they were at the start of 2026, crushing Goldman's optimistic margin outlook.

So Silicon Valley Data publishes their LLM token cost index. It's quite helpful to see. This is actually positive for AI demand. Qwen 3.8 Max by Alibaba has also reshaped the cost performance Pareto frontier and front-end code arena, with a pricing of about $2 per input M-token and $6 per output M-token, which is a tenth of what Claude costs.

## Hedge Fund Tech Buying and Dan Niles on Earnings - 1:10:58

On page 29, you can see the hedge funds bought tech stocks last week at the fastest pace since December 2022. And then Dan Niles on Mag 7 earnings, quick summary. The forced asset sale by Situational Awareness he thinks helped drive a sharp rally since July 30th as pressure on the short side flipped into short covering.

On July 29th, he wrote that from a technical standpoint, he felt that forced liquidations and margin calls, not just at Situational Awareness, but in two other hedge funds that run with leverage, was leading to a technical bottom. And his view was that we could have seen at least a short-term bottom at the end of July. And on July 30th, Morgan Stanley's TMT, Tech Media Telecom's momentum index rebounded a record 19% on Thursday and then added another 1% on that Friday. This followed a decline of 52% from June 22nd to July 29th, which was one of the worst declines we've seen. It's now down 44% from June 22nd. And this is why he focuses on avoiding speed bumps as he warned in his June 20th post. No one would have predicted that the market would have been down 50% in tech momentum since then. Looking at Mag 7 results, stocks reacting to earnings results mostly came down to two factors. Number one, did estimates go up for the third quarter if capex went up? And two, did they report results before or after the Situational Awareness forced sale? Microsoft's results strengthened the recent view that Copilot could be a winner in the enterprise for AI long-term. He wrote in his earnings preview earlier that Microsoft operates natively within the Microsoft 365 ecosystem where enterprise work is already happening. The 450 million Microsoft 365 paid seats, but only 30 million of them use Copilot right now. It's only about less than 10%. Microsoft guided above consensus for the third calendar quarter while capex remained unchanged. So the market viewed that as positive. Azure also saw growth improve sequentially from 39% year-over-year to 43% year-over-year with guidance of 45% year-over-year for the third quarter. That really buoyed the market.

Helped by this Situational Awareness forced sale, the stock saw its fifth highest one-day percentage stock move in history at positive 16% on that Thursday. Meta, unfortunately, had both revenues and operating income go down for Q3 outlook, while revising up capex and OPEX. They also did not announce any definitive plans around a public cloud offering or API for their foundational models to monetize the spend. The stock declined 8% in reaction to Thursday, which would likely have been worse if not for the Situational Awareness forced sale news. Amazon, while guiding both revenues and operating income below consensus for the third quarter and increasing capex, had AWS revenue growth accelerate from 28% to 37%, which was a massive 900 basis point improvement, the highest growth rate since COVID.

AWS normalized operating margins expanded 1% as well, so the stock actually rallied 15% despite revenues and operating income below consensus, because of how well AWS did, following a 4% rally on Thursday as investors continue to re-gross in the AI names. This brings him to Google, which he thinks remains a long-term winner in consumer AI with a complete AI stack. Google like Amazon did guide capex higher while implied revenues and operating income declined for the third quarter. But Google Cloud platform performance crushed AWS performance. GCP saw revenue accelerate from 63% in the first quarter to 82% in the second quarter, while operating margins expanded 3% sequentially instead of just 1% for Amazon. But Google, unfortunately, reported a week prior to the Situational Awareness forced sale and saw their stock decline 7% the next day. And then Google rallied after that, but then also saw their stock fall after one engineer left, down 5%. So that's one of the reasons why we've been adding to Google. Apple was the anti-AI trade leading up to the results and the stock hit an all-time high intraday on Wednesday. But the stock declined as investors re-grossed AI names on the SA forced sale. And it fell 7% on Friday in reaction to revenue and gross margin guidance that was below consensus on a lack of memory supply, because Apple requires memory for its iPad, its computers, and its phones.

## Market Drawdown and Company Specifics - 1:16:02

As a big picture, the severe drawdown in the AI favorites from June 22nd to July 29th was good for the market. It reminded investors not to take too much risk and to be vigilant on leverage. Long-term bond yields also hit 20-year highs, and the unresolved Iran war are factors that he continues to monitor as we monitor. He thinks that out of the megacap earnings the past two weeks, his favorites are Google, Amazon, and Microsoft.

Outside of this, we did see a probe in the life insurers. We've been warning on life insurers now for the last couple of quarters. And FT has been reporting that there's a formal investigation into Egan-Jones that we've been talking about — they've been rating thousands of companies with only 20 analysts and getting paid to do these ratings, so it's a conflict of interest.

Another name which is a potential special situation, which we're not fully sold on yet, is a company called Nano Dimension. Nano Dimension is ticker NNDM. Actually trades at a negative enterprise value of negative 80 million. So Nano Dimension's current operating business does generate a loss, but if you look at Bloomberg or in the financials, the loss is overstated.

So, what is Nano Dimension? It's a 3D printing company, focuses on R&D of 3D printed electronics that includes a printer for multi-layer printed circuit boards and the development of nanotechnology based inks. So it's an interesting company based in Illinois. But it looks like the company's burning 75 million a year — that's no longer the case. They've actually cut their burn by canceling some leases and selling two assets. So as of the second quarter of 2026, Nano Dimension's balance sheet exhibits a classic net-net deep value characteristic. It has cash of about 433 million.

It has total debt of zero. Has a market cap of 340 million with the stock trading at around 1.60. Its basic cash per share of around 2.10 to 2.25 a share. So an implied enterprise value of negative 80 million. Now the market currently values Nano Dimension's core operating business at a negative value, implying that the operating business will destroy more capital than the cash currently sitting in the bank. But why is this starting to change?

Historically, NNDM was unappealing. It was a cash burn story. However, under shareholder pressure and activist pressure, management initiated a three-phase strategic plan to stop the cash burn and to monetize assets. So what they did was they sold Markforged to Stratasys for about 42.5 million in cash earlier this year. They also sold the additively manufactured electronics division and Fabrica product lines for up to about 12 and a half million cash. So in total they generated about 55 million in cash from that. And they started to cut their employee count and they started to do fixed overhead reductions. So they terminated its corporate headquarters lease, which cut 38 million in future lease commitments, and they unlocked about 25 million in cumulative net cash savings, even after the lease break fee.

They also — that cut helped them cut their burn; combined strategic actions are expected to lower cash flow burn by about 25 million. So that's positive. And then they have a phase three strategic review where management suspended 2026 guidance and is actively evaluating strategic alternatives including potentially selling the company, a full sale of remaining units, capital return, or a merger. If management returns cash to shareholders via aggressive share buybacks, a special dividend, or a complete liquidation or sale, investors are now buying below cash value.

Why the discount still exists? So a negative EV stock is only cheap if the cash isn't burned before it reaches shareholders. Ongoing operating losses despite cost cuts — Nano Dimension still reported adjusted EBITDA loss of 9.6 million in the second quarter. Every quarter without profitability slowly erodes net cash. So there's also reinvestment risk. Historically, management used its massive cash hoard, which it raised during the 2020-2021 equity boom, to pursue expensive acquisitions rather than return capital to shareholders. The risk remains that the remaining cash could be deployed into deals rather than being dividended to shareholders.

So the verdict is that NNDM is a classic event-driven asset play. At current levels, you're effectively buying 2.20 of net cash for 1.85. This is not a massive discount yet, but it's quite interesting. Starting to get interesting, and you're getting the operating business for free. While this is cheap on paper, it only becomes a compelling trade if the phase three results in concrete capital returns or a total sale of the business after a board reevaluation.

## Caterpillar and Backblaze Earnings Surprises - 1:21:27

Now Caterpillar reported earnings last week and it really surprised the market. Caterpillar surged after the company crushed Wall Street's second quarter expectations and raised its sales outlook. Really, I think hurting Burry, I think Burry was short this name. So like Palantir was up 40% very recently and Caterpillar surged. So not a good week for Burry.

Now, the company's earnings crushed the second quarter expectations, easing concerns that demand for its power generation equipment used in data centers was beginning to cool. The company reported stronger sales across its businesses, posted a record equipment backlog, and increased its forecast for 2026 sales and revenue growth to the mid-to-high teens. Sales rose to about 20.5 billion for the period, exceeding the 19 billion average estimates compiled by Bloomberg, with the company's power and energy unit and construction industries businesses starting to see growth. Shares were up about 13% in early trading on Tuesday after the company reported these results. And we can jump very quickly to the following page on Backblaze. So Backblaze shares jumped about 50% pre-market trading after the backup software products provider reported earnings that beat the average analyst estimate and raised its fiscal year 2026 guidance. It was also up on a new five-year deal with CoreWeave, which means that its earnings could be even higher. Citizens Bank analyst Russ Dumba has this stock as an outperform, reflecting broad-based strength across the businesses and direct sales execution. Second quarter results were also strong at 8 cents EPS versus 10 estimate — versus 2 estimated, sorry. Revenue of 42.7 million versus an estimate of 39 million. B2B cloud storage revenue of 26.6 million versus estimates of 23.9 million. EBITDA of 12.8 million versus estimates of 8.9 million. So just beat across the board. The 335 million strategic agreement with CoreWeave, we still have to investigate what that 335 million is, resulted in it rallying further.

## BTIG on SPX Rally and Daily Market Recaps - 1:23:48

BTIG on the SPX rally, they're very incredulous. The SPX was up 6% over four trading days as of Tuesday night. But you know what, the market continued to rally despite BTIG's Jonathan Krinsky saying that the S&P 500 has only surged 5% or more in four days three other times. One in 1999, one in 2000, and one in November of 2020 during COVID.

So very quickly, on page 35, there's a daily recap of what happened in the market each day. On August 3rd for example, Palantir reported Q2 revenue of 1.94 billion up 93% year over year. They also reported strong margins and adjusted free cash flow reached 1.22 billion, which is a 63% margin. It posted a rule of 40 score of 155%. So very, very strong. And the stock was up 40% after that.

CEO Alex Karp is talking about how demand for sovereign AI has been unleashed. The stock market had one of its best days in 2026 on August 3rd. Semis, software, hyperscalers all participated together. Throughout the year, one sector would go up in tech while others would go down. But with oil falling 7% on the 3rd, and the market seeing the deleveraging event in the past, it was a big rally. Microsoft had the largest single day increase in its market cap of almost half a trillion, 450 billion, with NVIDIA having a record 441 billion back in April 2025. Trump said that the US will hold off on attacking Iran and other Middle Eastern countries. That was also on August 3rd. The semiconductor index saw intraday moves of 2% in all 22 trading days in July, matching the extreme volatility seen in 2020. Before that, a stretch like this had not occurred since the 2008 financial crisis. The SOX fell about 21% in July, its worst month since October 2008, wiping out 2.2 trillion in market cap.

The move pushed 60-day realized volatility to an amazing 66%, the highest since the pandemic peak of about 85% in 2020, and 75% in 2008. Jeff Bezos filed a Form 144 on August 3rd to sell 15 million Amazon shares. We already covered that, so we'll move on. Hedge funds are moving back into tech stocks, we covered that as well. Tesla sales in Spain plunged 80% in July to just 131 vehicles, so that was very surprising. Grab reported record 2Q26 results, raised its full-year guidance and announced an additional 750 million of share repurchases. So Grab is quite an interesting company. I think we've talked about it in the past. We have no big position in it. But if you look, Grab Holdings is one of the biggest tech growth stories. It's effectively like an Uber/delivery company based in Singapore. It's a relatively small market cap. It's down about 50% since its 12-month peak. But the business is an 11 billion company with basically 4 billion of net cash, does about 3.7 billion of revenue and it should do 700 million of EBITDA this year. So it is cheapish and this large share buyback program is about 5 to 7% of its total shares. It also, revenue grew 22% for the company, so it's also growing relatively quickly. Its on-demand GMV rose 21%. Profit just for the quarter reached 235 million, which was much higher than expected. Adjusted EBITDA was up 54% year-over-year. Margins were healthy and it's done cumulative buybacks of 1.7 billion since 2024. So it could be an interesting one to take a look at. Leveraged and inverse ETFs tied just to SK Hynix are basically on 5.5 billion in assets already, so the Koreans and Americans haven't learned from leverage yet.

Bernstein sees the global server market reaching 1 trillion by 2028, which is frankly insane. They think the global server shipments and GPU AI server shipments should be at 15 to 22% CAGR through 2028. The firm expects GPU-equivalent server shipments to grow 48% in 2026 with rack shipments reaching 61,000 this year and 88,000 in 2027. NVIDIA Rubin and AMD Helios racks are expected to start shipping in Q4 and that's one of the reasons why they're so bullish. And custom ASICs are expected to represent roughly 45% of AI shipments with Google TPU shipments expected to grow 90% in 2026, which is interesting. On August 4th, SpaceX missed expectations reporting 2Q revenue of just 7.8 billion, up 92% year-over-year, with EPS of -0.09 versus 0.024 negative expected. EBITDA did jump 191% from 3.5 billion. The issue is that SpaceX needs to invest so much in terms of debt and it needs to raise a lot of debt to fund its mega AI data center. Starlink subscribers did double year-over-year to 12 million, backlog reached 47 billion and cash and marketable securities stood at about 100 billion. AMD basically scared the market with higher capex, reporting a strong 2Q26 with revenue up 50% year-over-year to 11.5 billion and adjusted EPS up 246% year-over-year to 1.66 and the market still didn't appreciate it.

Data center revenue more than doubled, rising 107% year-over-year to 6.2 billion while adjusted EBITDA reached 3.3 billion and free cash flow came in at 1.6 billion. Capex was 808 million as AMD continues to scale for AI demand and Lisa Su, the CEO, who's cousins with Jensen of NVIDIA, said the company delivered record revenue and profitability with accelerating EPYC demand, growing Instinct deployments and Helios beginning to ramp, which is their new GPU chipset. Palantir had its best day since April 2025, closing up about 30% before its post-market move. Rocket Lab was awarded a 397 million dollar US Space Force contract to build, launch, and operate Flatellite spacecraft for the space-based air moving target indicator program, which is quite interesting. Anthropic reportedly signed a $10 billion compute deal with NVIDIA-backed Volta Infrastructure, securing capacity from a Norway data center operating in partnership with Bitdeer. Secretary of State Rubio said Iran and Oman are making progress. Barclays says S&P 500 earnings are coming well ahead of expectations with over 85% of companies beating, which we talked about. The Trump administration is reportedly drafting a ban on US imports of new Chinese data center components, which we talked about. The Wall Street Journal reported today that hedge fund Situational Awareness has built a $45 billion AI focused powerhouse before scrambling to raise cash after aggressive bets soured in July. We talked about this one as well. He had to sell his entire liquid portfolio to Citadel. He's left with basically his Anthropic investment, which was up 80%, and he made a new investment last week, which was surprising that people gave him capital for that. Elon Musk said that SpaceX has decided to build exclusively on NVIDIA on the fourth, which also supported the market rally. He's called Vera the best architecture and the best AI computer. Musk says that SpaceX deeply values its partnership with NVIDIA and expects to surpass 2 gigawatts of compute by year end with capacity growing to 10 gigawatts and 5 gigawatts by the end of 2027, which is bananas. Well, he'll have to raise so much debt in order to do that.

S&P 500 on the fourth closed at its highest level on record, adding roughly 1.2 trillion in market cap. On the fifth of August, SpaceX reported revenue of 7.8 billion. We talked about this. Let's jump to Rocket Lab. So Rocket Lab, we talked about as well. Let's jump to August 6th on page 38. So, SanDisk reported revenue of 8.97 billion beating estimates of 8.6 billion, up 372% year over year and 51% quarter over quarter. Adjusted EPS came in at 39.25 versus 34.45 expected. While adjusted gross margin reached a record around 84.6%. This was above the estimate of 81.5%. Data center revenue was up 103% quarter-over-quarter to about 3 billion. Edge revenue rose 48% quarter-over-quarter and 392% year-over-year to 5.4 billion while consumer revenue declined to about 556 million. Now for the first quarter of '27, SanDisk guided revenue to 10.3 to 10.8 billion, which was below the 11.1 billion estimate, which is why the stock sold off. It also gave an adjusted EPS of $44 to $46 a share, and the company also announced a new $14 billion buyback authorization, bringing the total remaining repurchase capacity to $15.5 billion.

President Trump warned that if Iran backs out again, they're going to get hit really hard. August 6th, that was just a bluff. China's gold-backed ETFs have seen 14 days of inflows through Monday. So Chinese gold ETFs, after many, many weeks of outflows after the spike in Jan, they're now reporting inflows back into gold ETFs, which is interesting, where we've been adding to gold names like AGI, which did very well at the end of last week. Also adding to Kinross. US job openings fell by 178,000 on the sixth in June to 7.36 million, the lowest level since March and the second straight monthly decline, which corresponds to the weak payroll data. Openings now are down almost a quarter million over the past two months. Declines have been led by private education, health services, leisure and hospitality, which was surprising. Wholesale trade down. So leisure and hospitality were down 86,000, wholesale trade was down 74,000, professional business services were down 71,000, which is quite sad. Situational Awareness actually reported it's back to investing with a $400 million private bet according to Bloomberg. So Shopify reported in the second quarter, 3.58 billion in revenue, beating estimates of 3.45 billion, up 34% year-over-year. GMV rose to about 115.6 billion versus 112.12 billion, while operating income increased by 68% year-over-year to 488 million, and MRR came at 221 million. So it was very impressive from Shopify. Raymond James is sticking with a strong buy rating on SpaceX, which I'm surprised by. Google's AI leadership is reportedly undergoing a major shuffle. Demis Hassabis is also leaving his role as CEO of Google DeepMind to become chairman of the unit, while also adding the title of chief scientist at Alphabet, continuing to help lead Isomorphic Labs. Google DeepMind CTO Koray Kavukcuoglu will become senior VP of the unit reporting to Sundar Pichai himself. Meantime, longtime Google AI leader Jeff Dean was leaving to start Discovery Loop, an independent public benefit corporation backed by Google, which resulted in that 5% sell-off. Zeta closed up 12% at a 52-week high as analysts raised price targets following the company's latest updates. KeyBank increased the target they have to 27 from 22. So the company had early traction for Athena, especially among customers with over 1 million in ARPU, which is a growth engine for them. CoreWeave signed a multi-year agreement with Solidigm, an SK Hynix subsidiary. They also did a couple other deals which we'll talk about. Goldman's compute forecast is on page 40. Goldman predicts compute token consumption to grow from 5 quadrillion now to 120 quadrillion in 2030. So that is absolutely bananas, right? So that is a 24x increase.

Hyperscaler revenue growth as a whole has accelerated in the last quarter, also because of just spending and growth of their balance sheets. Goldman's Peter Callahan basically said that a V-shaped recovery has happened. Nasdaq is now up 945 bips or nine and a half percent in just four sessions since last Thursday, punching back above its 50 DMA to the upside. Since then, the S&P made even more highs. The four-day move stacks up with how tech has traded out of or during other notable market events over the last 20 years like the GFC, COVID, the '22 hiking cycle, liberation day, etc., except that this happened a lot faster. So why the big move? Cleaner positioning and improved technicals, momentum factor bounce and/or reduced levered ETF footprints, tighter valuations with the Nasdaq PE at a 10% discount to the five-year average, stronger visibility and fundamentals and improved ROIC estimates for the Mag 7.

## Company Earnings and Market Analysis - 1:38:19

Eli Lilly also beat, I'm not going to go through that in detail on page 42. Palantir beat, we talked about that briefly. That's on page 43. There's an ADP miss. And then I want to just show you that on August 4th, there's 4 million S&P call options being bought, which is the highest ever in history of S&P call options being bought. That's on page 44. It just shows you how speculative the market is. Shopify beat, the details of that strong beat are on page 45. TD Cowen, Jefferies, they're all very very bullish on the name. The price targets aren't that aggressive. AI is making biotechs too productive in China, resulting in a spike in lab monkey prices. What most people don't know is that US pharmaceutical companies are actually buying a lot of their IP from China, and the US government is not happy about that. And that's because China is using AI to create all these early stage biotech IPs and it's resulted in them effectively pricing these lab monkeys at 26,000 each because there's just so much AI research to do on these new drugs. And because the US doesn't do this type of research, depending on AI and immediately testing on monkeys, there's a big productivity and speed disparity between Chinese early clinical trial design and US biopharma workflows. And as a result, China now dominates the early stage pipeline, roughly 34% of all global clinical stage drugs being produced there and 40% of the world's new candidates. Though, I think the US really needs to step up the FDA and screening process so that we don't have to rely on China for all these new early stage drugs. Western pharma companies are now buying all these Chinese assets instead of spending 1 billion plus developing drugs in-house over a decade and spending on highly paid in-house scientists. Western megacap companies are aggressively licensing pre-vetted AI-designed molecules from Chinese biotechs. So there are a lot of these deals that have been announced recently. So if you look at the middle of page 47, Bristol Myers Squibb entered a landmark R&D alliance with China's Jiangsu Hengrui Pharmaceuticals, covering up to 13 preclinical programs. The potential deal reaches up to 15.2 billion, which we're going to be paying this Chinese company, with 650 million upfront. Pfizer also inked a major licensing agreement including a 1.25 billion dollar upfront deal, potentially worth up to 6 billion, for this Chinese 3S Bio company on experimental cancer drugs. And Regeneron Pharmaceuticals paid 80 million upfront and a 2 billion potential deal with Hansoh Pharmaceutical to license an experimental obesity drug. AbbVie and Merck both have signed multi-billion dollar licensing agreements targeting Chinese developed antibody drug conjugates or ADCs and multi-specific antibodies where Chinese AI biology platforms have significantly shortened hit-to-lead optimization times. You can read the same thing about specialized and focused US biotechs and why congressional scrutiny is increasing, like over 30 to 45% of all US pharmaceutical licensing deals are now involving Chinese assets and US lawmakers have voiced bipartisan national security concerns about this. Margin debt relative to market caps spikes before every correction, every correction. We're right about this. Margin trading in Japan stocks is also the highest since 1990, so it's not just Korea. On page 49, Google stock dropped about 5% in minutes after the long-time chief scientist Jeff Dean announced his departure, but we actually bought the stock on that sell off and I think it was a good opportunity. You guys can read about why he left on page 49, and token prices are now down 40% from the May all-time high. We talked about this with Goldman on page 50.

## Uber Analysis and Other Market Moves - 1:42:38

We added to Uber shares because Uber was around 68.50 last week. It's obviously compelling as leader of a global duopoly and many valuation metrics make it look cheap, whether it's free cash flow yield. The company can do 15 billion of free cash flow by 2028 on a 139 billion dollar market cap with not a lot of debt. That is very impressive. That is effectively an 11% free cash flow yield for a growing business. It's still growing top line at 10 to 15% a year. On a trailing basis, the company is doing about 10 billion of annual levered free cash flow and is expected to generate 15 billion by the end of '28, which is just unreal. Going from a six and a half, 7% free cash flow yield to 11% on a two-year forward basis. The company was also recently upgraded to Triple B+ or investment grade and Uber authorized a $20 billion share buyback program and already repurchased 3 billion of stock in the first quarter. The ecosystem lock-in — Uber has passed 50 million global subscribers. So that itself is very valuable and members spend roughly three times more than non-members, driving down customer acquisition costs or CAC. And then high margin advertising, Uber's advertising arm runs a $2 billion annualized run rate, which is very impressive, and they're going to grow that advertising business just like Amazon has grown their advertising business. And then margin expansion, Uber saw operating margin reach 14.6% in recent quarters. Operating income grew 50% year-over-year as fixed overhead costs scaled lower than gross bookings. There's some insurance tailwinds as well around mobility. And then the risk is that the autonomous vehicle strategic moat isn't there. The primary bear thesis against Uber is that autonomous vehicles or robotaxis — Waymo, Tesla, Zoox — will eventually disintermediate human driver ride sharing, but I think that that's going to take many years. You guys can read about more of those risks on page 52, and then on page 53, there's an NVIDIA employee describing risk from Meta selling excess compute in the market. That is on page 53. Effectively it's one of the reasons why the stock sold off at the end of last week. They have about 400,000 GPUs they need to dump on the market, and they have no other way but to enter the game. So effectively because Meta's AI build has just not been going as well or is not growing as quickly, they have almost half a million GPUs that they could rent out to other people. We don't know if that number is true or they're going to find new use cases for it, but it's just a red flag that someone brought up in a call. AppLovin had a very, very modest miss. Like, I'm talking about less than 1% miss and the stock was down double digits. So we did buy that, even though there's both a bottom line and a top line, like, modest miss. But the more notable disappointment, I thought, in App was that their free cash flow missed by 27%. But that was driven by timing of international cash and tax payments. So after I understood that, I wasn't as worried about it. And Q3 revenue guidance was essentially in line with prior consensus, which means that while they weren't growing a lot, or as much as they were before, it looks like they're monetizing their clients better, which is very positive. So model timing they said was the primary culprit. AppLovin experienced a lighter than normal pace of AI model improvement during the second quarter, with a significant performance step-up occurring just after the quarter ended, of course. The consumer vertical was a very bright spot for the company. Compute costs pressured margins, which we all know, and free cash flow was timing driven. Still have an SEC inquiry there. But I do think it's a very interesting name to have in the portfolio, AppLovin, because I think it's less exposed to AI than your traditional search companies. And then UBS's note on App is on page 57. Alphabet seeking to raise 25 billion of debt, and Citadel equities being up 14% year to date because of taking over Leopold's book at the right time. And we had sold Datadog back in June, we're now adding it back after it's down 17% on a beat. That's on page 58. And then on page 59, you can look at the LLM cost index, which is interesting. And then TEAM, our position being up 32% pre-market and almost 40% on the day on page 60. You guys can just go through that beat. They're at 187 EPS versus 150, so this is not a small beat. And then Doximity, we did that intraday trade, it was up around 240%. We started to short it when it was up 80% when the market opened. And it effectively made over 20% in a couple hours on that trade. So we can see the timestamp on page 63.

## Q&A Section - 1:48:19

We're going to go into Q&A here, for uploading the call results. So Yojo is saying, let's look at SPCX and SATS again. Yeah, so we can revisit that. EJW's asking, can you attach the August allocation spreadsheet? It's hard to see the PDF. Yeah, the spreadsheet is already in the allocation tab. I'll just upload it again. Let's see. In here. So you have the Excel again. Let me see here.

So just download this and I'll upload it again in the Q&A tab. All right, so I uploaded it again.

All right guys, if that's the last question, we will continue to send you guys updates during the week and if you have more Q&A, we can often just answer that via a message in the Discord. But the recording will be up shortly. I hope you guys have a very successful trading week this week and hopefully the 10 attachments we shared on the carry trade, Goldman Sachs's weekly pulse, the Standard Chartered weekly update, HSBC update, DB on geopolitics — these are all very helpful reads. There's a lot of data for example in the Morgan Stanley piece on inflation, on small cap versus large cap, growth versus value, emerging markets divergence, Japan, Eurozone, US equities, Chair Warsh on reform, not hawkish signaling, the inflation trajectory, concerns on Iran and the Strait of Hormuz and the updated shipments through the Strait, on AI's impact on labor. There's quite a lot in the uploads and there's more in the Discord than obviously on email because the attachments are too big. So I hope you guys have a chance to go through that and benefit from it, and our recording will be up very shortly.