# Special Situations Report — Weekly Call (readable notes)

- **Title:** Bessent loses control of the 10-year — three failed attempts in three weeks to talk the bond market down while US debt crosses $40 trillion, the 30-year hits 5.29% and the dollar falls with it; gold's largest weekly gain in a year (~$4,700/oz, GLD 370 → 420), silver ~$69 and copper ~$6.60 with Hudbay (HBM) held; a three-sigma move in crypto; Jeffrey Lacker's "Revisiting the Treasury-Fed Accord" testimony against Warsh and Bessent and the bills-only doctrine; the big-tech bond paradox (Google's 100-year bond, Meta and Oracle at 6-8%) crowding out Treasuries; Nvidia's 15%+ AI-server price hikes on DRAM/NAND inflation and a 75% gross margin called a peak; open-weight models (Zhipu AI's OX Alpha / GLM-6) closing on the frontier while Anthropic's Fable 5 plateaus at 11% of its own corporate spend into an October IPO; Citigroup (C) as the large-cap value rate beneficiary — $165 price target, buy on pullbacks at 9.5× forward and under 1.2× tangible book; a starter position in Priority Technology (PRTH) at $5.52 against the founder's $6.00-6.15 lowball take-private, with Buckley Capital at $15-20 and Steamboat opposing; Moderna (MRNA) shorted into the 175% cancer-vaccine squeeze and covered for a high-single-digit gain while the BioNTech (BNTX) short stays on; and the WBD/Paramount-Skydance spread compressing from ~20% to 8.8% as Cinema United flips to support the merger
- **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-23 (Sunday call; the deck/transcript title it "2026-08-23 - SSN")
- **Source PDFs (this folder):** `transcript.pdf` (the verbatim premium transcript, its section headers carrying the recording's `MM:SS` cues) and `report.pdf` (the condensed report + actionable items + the economic/earnings calendar + the stock-symbol list, with deep-dive write-ups on PRTH and Citigroup). No agenda deck was distributed with this week's call, though the recording walks a slide deck page by page (pages 7-44 are referenced throughout).
- **Note:** readable notes extracted from the premium PDFs; fillers (um/uh/you know/false starts/stutters) removed, wording otherwise verbatim from the recording. Section headers carry the recording's `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 — the transcript and the report title both garble it); **"Bessent" / "Essent" / "Besant" = Scott Bessent** (Treasury Secretary); **"Worsh" = Kevin Warsh** (Fed Chair nominee); **"P-Sky" = PSKY (Paramount Skydance)**; **"Triple C" = CCC Intelligent Solutions (CCCS)**; **"JBSNV" / "JBC" = JBS N.V.**, the 82% holder bidding for Pilgrim's Pride (PPC); **"Nuvea" = Nebius Group (NBIS)** — the Vineland, New Jersey data centre tied to a $17.4B Microsoft cloud deal and the $4.5B convertible offering are Nebius's (the report's symbol list mis-maps it to "NEO, Nuvea (Private / Specialty Debt)"); **"Merker" = Mercor**, the AI data startup Nvidia discussed investing in at a ~$20B valuation; **"Zipu AI" = Zhipu AI**, the Chinese lab whose zodiac-codenamed stealth models (Pony Alpha → GLM-5, OX Alpha → presumably GLM-6) are benchmarked against Anthropic's Mythos class; **"Hyperliquidator hype" = Hyperliquid (HYPE)**; **"CBAC" = TBAC**, the Treasury Borrowing Advisory Committee; **"Jean Hatzius" = Jan Hatzius** (Goldman's chief economist); **"Intisumeran" = intismeran autogene**, the Moderna/Merck personalised mRNA cancer vaccine; **"Thomas Priori" = Thomas Priore**, PRTH's founder, chairman and CEO; **"Buckley Capital Advisers" / "Buckley Capital Partners"** are the same activist (Buckley Capital); **"Bonta" = Rob Bonta** (California Attorney General); **"Jameson Greer" = Jamieson Greer** (US Trade Representative); **"Zervos" = David Zervos** (Jefferies' chief market strategist, called "head of macro" on the call); **"Brooking Institute" = the Brookings Institution**; **"Smith cap squeezes" = small-cap squeezes**; **"Iren" = IREN (Iris Energy)**; **"Yardeni" = Ed Yardeni**. Internal errors and inconsistencies in the source, left as spoken and flagged here: **"The two Korean companies, Micron and Samsung"** is wrong — Micron is American; the two Korean memory makers are Samsung and SK Hynix (the sentence before names all three correctly); the **defense request is given as "$500 billion"** and then itemised as "almost 600 billion of Patriot interceptors, 400 billion of THAAD, 200 billion of Navy missile defense… when you add all this up, it's almost a trillion dollars", which does not reconcile; **PRTH's market cap is given as "$500 million"** and two sentences later as **"its 460 market cap"**, and the earnings that preceded the bid are called **"Q3"** for a company on a December year-end; **"$10 would be a 40% premium… more than, sorry, based on the 561 close, $10 would be a 78% premium"** is a live self-correction; **Bitcoin is read as "briefly touching 78,000… then rallying to 77,000 later in the week"** on Wednesday and **"crossed 72k"** on Thursday; the **seventh-largest crypto short-covering/liquidation event** is dated both Thursday and Friday; **"a record of 63.4%"** of S&P 500 stocks yielding more than the 10-year contradicts the report's "dropped to 63.4%", and its exclusion is given as "the COVID-19 crash back in July of 2016"; **"ADP said private employers only added 9.5 thousand jobs per week"** mixes the monthly print with a weekly unit; **"China's… new loans falling by 50 billion"** omits the currency (yuan); **"Google issued a 100-year bond with a 6-year yield"** should read a ~6% yield; **"Companies who build servers under contract for large data center operators like Microsoft, Alphabet, and Google, and Oracle"** double-counts Alphabet and Google; the report's symbol list mis-maps **"ANAB"** to Anthropic (ANAB is AnaptysBio; Anthropic is private) and lists **JNJ and TSM**, which are not mentioned in the recording. The transcript footer's "Date of Transcript: 2024-05-24" and its "ssn_transcript_v6 - 2026-01-26" filename are stale template artifacts — the real call date is **2026-08-23** (Sunday).

---

2026-08-23 - SSR - Bessent Loses Control of 10Y, Gold Rallies, Crypto 3 Sigma Move, Citi Value, PRTH Special Situation, MRNA Biotech Squeeze, WBD Success in Spread Compression

## Introduction and Market Overview - 00:00

Alright, so happy Sunday everyone. Thanks for your time. We love to share our thoughts with you today on a few very important topics. As you saw last week, Bessent increased Treasury bond buybacks, effectively a quasi QE, that the market didn't take very lightly. There are three times during last week where Bessent tried to calm the bond market and despite that, the 10-year continued to climb. The 30-year reached 5.29. And given the 10 year is a benchmark for mortgage rates, this is not a very good thing for the consumer. Consumer spending has weakened according to Goldman. GDP estimates have slowed to one and a half percent for the second half of the year. The majority of GDP growth is AI related spending, which is another topic to discuss altogether.

But we'll go through the economic calendar. There's quite a big economic calendar this week. Gold rallied the most in a year this week. Crypto saw a three sigma positive move during the week. We think that banks will benefit from higher rates if we don't see this rally get under control. Citigroup is a large cap value name. We have $165 price target on it. We'd like to buy it on pullbacks. It should benefit from higher rates. PRTH is a special situation that we picked up around 550 during the week. Effectively the company has a majority shareholder and this shareholder is competing with activists to take over the company. He has given the company a lowball bid at like five times cash flow, at $6 a share. We think that he may have to raise that bid. So we'll talk about that name as well.

There's a biotech short squeeze last week. We shorted MRNA for a very quick profit. And we've kept a short on BioNTech because we think that the management suite has quit. They don't have nearly the success of MRNA when it comes to cancer vaccines and it just benefited from that small cap squeeze, which we'll talk about. And then on special situations, Warner Brother had a couple positive things happen with some movie theaters flipping to support the deal. And they also have posed a $7 million per day ticking fee, which goes from P-Sky to Warner Brothers, which only helps their downside case. And then LBRDK is closing as well.

## Economic Calendar Details - 02:40

On the economic calendar for this coming week, it is quite a heavy calendar. The most important economic events next week are the ADP employment, the FHFA housing price index, new home sales, consumer confidence, personal income and spending, and the most important being the PCE price index, which is supposed to drop to 3.6 from 3.7 for the month of July, right ahead of Jackson Hole. And obviously the core PCE, which is supposed to stay flat at 3.3%. Outside of that, you have durable goods, initial jobless claims, continuing claims and UMich sentiment.

## Earnings Season Overview - 03:18

In terms of earnings, we're basically at the end of earnings season, but there's still a few stragglers left to report. We saw a big miss last week from Walmart. We'll see some retailers reporting this week as well. We see Dick's Sporting Goods on Tuesday. We also have some tech stragglers that still have to report. We have Intuit after the close. We have Zoom. We have Box after the close on Tuesday as well. On Wednesday, we have another retailer, Kohl's. Williams-Sonoma, JM Smuckers, Bath & Body Works. And after the close, the most important name reporting next week is Nvidia. After that, we have one of the biggest software benchmark companies, which is Salesforce, ticker CRM.

On the cybersecurity side, we have CrowdStrike and Okta. And on Thursday, we have Best Buy, where we'll see memory costs, how much those memory costs affect sales due to higher laptop costs and other electronic prices. We also have a number of Canadian banks reporting this week including CIBC and Scotiabank. On Thursday, we also have Dollar General. We'll see if they saw same-store sales decline like Walmart and Royal Bank of Canada and TD Bank. So that's a big Canadian bank reporting week. We also have another dollar store, Dollar Tree reporting pre-market on Thursday. Burlington Stores, another retailer on Thursday. And then after the close, we have Marvell. We have Iren on the data center side. We have Autodesk on the software side, Affirm, Ulta, Workday on the software side again, SentinelOne on the cybersecurity side, Elastic on the software side, and finally Gap on Thursday.

## Special Situations Update (Warner Brothers) - 05:08

In terms of the special situations update, shares of Warner Brothers finished Wednesday's session up 2%. As you know, this spread has come in very dramatically. We were adding when it was a 20% spread, now it's only 8.8 spread at the end of the week. So Warner Brothers is up 2% after Cinema United, who previously opposed the merger with P-Sky, changed course and joined the three largest theater chains in support of the merger. Cinema United is seeking a long-term commitment that Paramount would maintain or expand the number of films the studio offers in wide release, provisions that the merger would not increase the fees that theater owners pay to exhibit films, and continued access to the vast film libraries of Paramount and Warner Brothers. Bonta told reporters on Tuesday that he's happy to hear theater owner's views, but the fact remains that the proposed merger breaks the law. So we still have some anti-trust pushback.

They're worried about job losses, wage cuts, etc. But the market's looking past that given all the movie theaters are now in support of this deal. So the spread has come in from roughly a 20% spread all the way down to an 8.8% spread. So that's been a win for us on the special sits front. In other news, the Surface Transportation Board, STB, announced a decision adopting a procedural schedule for the consideration of a revised merger application for Union Pacific and Norfolk Southern. Final briefs are due on May 28th, 2027, so that's quite a long ways from now.

## Special Situations Update (Triple C, JBSNV) - 06:42

Shares of Triple C rallied as much as 12% last week after Bloomberg reported Copart is in talks to acquire the company with competition from several PE firms including GTCR and Veritas Capital. Lastly, JBSNV announced that it has submitted a non-binding proposal to acquire all of the outstanding shares of common stock of PPC. JBC owns approximately 82% of PPC stock currently.

## Warner Brothers/P-Sky Bond & Delaware DOI - 07:23

Early in the week, P-Sky requested a $1.9 billion bond from the 12-state coalition that was opposing the deal that had filed the antitrust lawsuit. The bond is meant to cover the potential harm generated by the ticking fee should the deal not close by October 1st. In a statement, the AG's office argued that Paramount is sophisticated and it knew that it would face regulatory review when it agreed to pay Warner Brother shareholders a $7 million per day ticking fee, and that California should not be made to bear the burden of that decision. In other news, the Delaware DOI provided an update in regard to the pending merger between BHF and Aquarian Capital. The department stated that the acquisition has generated significant national and international interest as well as the questions about the department's regulatory review process. The department is using its expertise along with carefully selected outside specialists to evaluate the proposed transaction and ensure it meets all applicable regulatory requirements.

You can see the summaries of the Warner Brother Cinema United deal acquiescence, the NSC UNP deal, the Triple C acquisition from Copart, the $1.9 billion bond requested by Warner Brothers, and the rejection from California, the AG. And then the LBRDK merger is expected to close on August 20th, which just happened.

## Defense Sector Spending - 08:52

There's a $500 billion funding request made for missile and air defense over the past week. The US Navy specifically requested RTX for Tomahawk missiles, giving them a $23 billion contract. So this is very positive for defense companies. You can see the 2027 budget request versus the 2024 budget request. We have almost 600 billion of Patriot interceptors, 400 billion of THAAD missile defense, 200 billion of Navy missile defense. Now, when you add all this up, it's almost a trillion dollars. So the 70 billion that we've spent on the Iran war is not accurate. We've deployed a lot of our defensive capabilities in the Middle East and our offensive missiles that need to be rebuilt and restocked. So our estimate is that this war has basically cost half a trillion dollars already in all the spent ammunition and interceptors and now we have to rebuild.

And that's going to be added to the budget deficit. So we could easily be seeing a 3 trillion plus budget deficit next year. With our current debt over 40 trillion, over 100% debt to GDP, we're approaching Greece type debt levels. And it's quite embarrassing to be frank. There's really no way that they're going to get this under control outside of monetizing the debt and, given what the bond market is doing to Bessent, by forcing the Treasury to do this quasi QE even further and the bond market is going to test that.

## Nvidia Price Hikes - 10:30

In other news, before we get into the long bond discussion, Nvidia is hiking prices by 15% plus to hyperscalers and other customers. Nvidia's biggest customers, which include hyperscalers like Microsoft, Amazon, Google, Meta, have been told that the prices of servers containing its artificial intelligence chips are going up more than 15% in many cases with memory chip costs soaring. The price hikes will impact systems including those with the flagship Vera Rubin and Grace Blackwell chips and will depend on the generation of Nvidia chips and memory configurations.

Companies who build servers under contract for large data center operators like Microsoft, Alphabet, and Google, and Oracle have already notified their customers of the forthcoming increases. The price hikes will go into effect on systems shipped early next year and will impact systems including those with the Vera Rubin and Grace Blackwell chips. The inability of the industry's most dominant company to hold the line on prices or absorb the growing costs shows how much leverage makers of memory chips like Samsung, SK Hynix and Micron have amid the surge in demand for AI infrastructure. In fact, if you look at the price of NAND and DRAM chips, it continues to go parabolic.

Major technology companies including Apple and Qualcomm have recently said they've been forced to charge more for their products because of chip shortages. Nvidia's accelerator processors are the heart of computers that create and run AI software. And their effectiveness depends on how much dynamic random access memory or DRAM they're paired with. The two Korean companies, Micron and Samsung, account for most of the world's production of that type of chip. Nvidia has also raised prices for its gaming-oriented PC graphics cards. The company has a gross margin of 75%, but that's likely a peak margin.

The price increases are also likely to add complexity to the industry's massive AI data center build out ambitions, to add on to power delays, project delays, labor shortages. Nvidia's reporting its earnings this week, so we'll see what they have to say about these price hikes.

## Open Source AI Models - 12:42

There's been more and more news about open source artificial intelligence models growing in importance. Rapid advances in open source artificial intelligence models, which are cheaper to run than closed source alternatives, have raised questions about who stands to gain and lose amongst the AI chain. Jim Covello, who's head of Goldman Sachs equity research, thinks that these smaller, faster, cheaper models could ultimately benefit the hyperscalers, the largest cloud computing providers. Open source models will more likely mean enterprises can properly implement AI in their organization, Covello said in his latest piece. By demonstrating that AI can boost corporate profits, open source models could in turn encourage more companies to adopt the technology, which is actually good for the cloud providers. He thinks it's good for the hyperscalers because it's more likely that you're going to be able to fill up this capacity that you're adding.

One technology bottleneck is the model optimization layer. Leading AI models are already powerful and capable. What companies lack is a way to route their queries to different models so that the highest consequence jobs go to the more expensive frontier models and the lower consequence queries are sent to open source models. That's going to be one of the big keys to unlocking the economic value of AI in the enterprise. As hyperscalers raise significant amounts of capital to build AI infrastructure, now including Alibaba, which is raising 10 billion overseas, investors are pressing harder on how and when they plan to see a return on their investments. Until relatively recently, every time one of these companies would announce higher capital expenditures, the market would reward that company. Over the last quarter or so, you've really seen a significant shift where the market is questioning this a lot more.

## Local vs. Frontier AI Models - 14:27

Now, local models are catching up with frontier models, but the frontier models continue to evolve. Open weights local models have closed the gap dramatically for standard everyday user queries. So Q&A, drafting, basic coding, single-turn summarization. For 70-80% of these routine tasks, a modern local model provides an experience comparable to a cloud LLM. The 89% figure below for 2026 is an optimized ensemble calculation rather than a real world single model deployment metric. So it's statistically inflated, but it assumes a perfect use of these open source models and perfect routing.

## Zipu AI's OX Alpha Model - 15:07

This OX Alpha is a 100% GLM model created by Zipu AI and it looks almost like Anthropic's Mythos class from very early results. It's very likely going to be called GLM6 and it's absolutely mogging every frontier model in cyber benchmarks. You can see how it ranks in the table below on page seven. First off, Zipu AI code names its models with the names of Chinese zodiac signs followed by Alpha. They did this in February of 2026 with GLM5, code naming it Pony Alpha and releasing it later as GLM5. The same was released on open router as stealth Pony Alpha, hence OX Alpha.

OX Alpha's tokenizer is one for one identical to GLM's. No other lab's tokenizer gets past 411 when it comes to tokenizer similarity. You can see in the next page that OX Alpha's tokenizer, how it compares to other players and other models that are out there. OX Alpha is free right now, unlimited usage and no guard rails on open router and open code. But if this model drops as an open source model and really turns out to be as good as Mythos on all coding and cyber benchmarks, it might just be the final nail in the coffin. And it could be another deep seek moment. However, they probably won't have enough GPU or compute for it to scale right now.

## Anthropic's Market Struggle - 16:37

There is news in the Financial Times that Anthropic's best AI model struggles to attract users as cheaper tools thrive, including Anthropic's other models. Corporate demands have been shifting. Anthropic's flagship frontier model Fable 5 has plateaued at roughly 11% of total corporate spend on Anthropic tools. This is because of a rise of cheaper alternatives. Enterprise clients are actively routing work away from the costliest frontier models towards cheaper, good enough options, including lower tier models like Opus 5, as well as open weight alternatives.

In terms of business model implications, this trend challenges the assumption that technical capability leadership automatically translates into revenue, forcing AI labs to rethink pricing structures ahead of expected market exits and IPOs, like Anthropic's IPO expected for October.

## The Fed, Bond Market, and the US Debt - 17:30

Very importantly this week, Bessent pleaded with the bond market to calm down three times, promising additional long-end purchases as the US debt load tops 40 trillion. So now, according to the Financial Times, no fewer than three times in three weeks, Bessent has given the message to investors to please stop selling his bonds, but they're selling his bonds anyway, taking borrowing costs to extraordinary heights in the same week the US debt crossed 40 trillion for the first time. Worse, investors are now selling the currency too, as the dollar fell. The dollar took a sizable hit after the buyback announcement, which resulted in a big rally in gold and a three sigma rally in crypto, specifically in Bitcoin, Ethereum, and Solana.

On Sunday, US Treasury Secretary Scott Bessent scheduled a press conference for Monday at 2:00 p.m. tomorrow Eastern at the Treasury Department. The primary triggers and market context surrounding this announcement include one, the reacceleration of the bond market sell-off. The failed intervention. The Treasury department attempted to curb long-term borrowing costs by announcing plans to at least double its long duration bond buybacks, purchasing 4 billion per operation in 10 to 30 year treasuries. For those of you who don't know, this was to start on September 9th and the US government already owns about 50% of all long-term bonds. Not much dissimilar to Japan. While long-dated treasuries initially rallied on the news, the gains completely unwound by Friday. Yields snapped back towards multi-year highs with the 30 year hovering around 5.25% as investors dismissed the buyback strategy as a band-aid on a bullet hole.

Two, fiscal debt and deficit concerns. Right now we have a 40 trillion dollar debt wall. Total US national debt surpassed 40 trillion last week, coinciding with the CBO projected annual budget deficit of 6% of GDP. Bessent has indicated that Monday's announcement will include new measures aimed at fiscal consolidation to demonstrate a path towards reducing federal deficits. Escalating Iran tensions, and now a trade war with Canada. In addition to debt management, Bessent stated earlier in the week that Monday's press conference will detail stricter economic sanctions against Iran and potential secondary enforcement against nations aiding Iranian oil flows.

## Congressional Testimony & Fed Mandate - 19:45

Congressional testimony by Jeffrey M. Lacker, a former president of the Richmond Fed delivered during the House Financial Services Committee last week, was titled "Revisiting the Treasury Fed Accord" and directly challenged Warsh and Bessent. The highlighted passage carried a major implication for the institutional boundary between the Federal Reserve and the US Treasury, and the core meaning of the highlighted passage, which we took a screenshot of on the next page on page 11, basically redefines a third leg of the Federal Reserve's mandate. So under the Federal Reserve Act, the Fed has a statutory dual mandate for maximum employment and price stability. That legally includes a third often overlooked mandate containing a moderate long-term interest rate.

Minimizing the term premium is something he discusses as well and should be interpreted specifically as minimizing the inflation risk premium and macroeconomic uncertainty premium built into treasury yields. And number three, price stability lowers borrowing costs. Lacker argues that the best way the Fed can keep long-term yields low for the US Treasury is not by capping yields or buying government bonds, which is QE, but by maintaining tight control of inflation so that bond holders don't demand a higher risk premium to hold US debt.

## Lacker vs. Bessent & Warsh on Policy - 21:00

So what he's doing is he's pushing back against fiscal dominance. As the US government debt expands and interest expense surges, there is a growing political pressure for the Fed to use its balance sheet to hold down long-term yields, which is also called yield curve control, which Bernanke started after 2008 with Operation Twist, or targeted Treasury purchases. Lacker's testimony pushed directly in the opposite direction, warning that using central bank liquidity to fund or stabilize federal deficits degrades monetary policy independence. He also revisited the 1951 Treasury Fed accord. Prior to 1951, the Fed was forced by the Treasury to cap government borrowing rates to low levels to fund World War II debt. The historic 1951 accord ended this yield peg and established modern Fed operational independence.

Lacker is advocating for a new accord that explicitly prohibits the Fed from stepping into secondary Treasury markets under the guise of yield management or market functioning backstops unless strictly necessary, forcing Congress and the Treasury to absorb the market price of their fiscal policies. If Congress adopts this framing, the Fed would move away from long-duration bond interventions like QE or Operation Twist to revert to a bills only or pure short-term framework. The Treasury would no longer be able to count on the Fed balance sheet expansion as a buyer of last resort to absorb term risk or compress long-term yields. Jeffrey Lacker's testimony functions as a direct critique of the economic playbook being pursued by Treasury Secretary Bessent and Fed Chair nominee Kevin Warsh, which is what resulted in the market to second guess the ten year.

While Lacker supports the conceptual idea of a new accord in theory, he's warning against how Warsh and Bessent intend to implement it. Lacker rejects Treasury interventions to cap yields, whereas Bessent has been actively managing long duration debt via buybacks and maturity manipulation to keep long-term yields down just like Yellen used to. Lacker has basically said that the only legitimate way to moderate long-term interest rates is by eliminating inflation risk, which is getting worse with this new tariff war and the Iran war. Warsh's proposal around the Treasury Fed accord argues that the Fed should maintain strict independence on setting short-term interest rates while collaborating with the Treasury's non-monetary balance sheet policy and emergency swap lines. Lacker has publicly pushed back on Warsh's version of the accord: according to the reporting by Briefs Finance and the analysis by the Brookings Institution, Lacker warned that Warsh's arrangement risks turning into a less constructive agreement that lets the Treasury use the Fed's balance sheet to bypass Congress, perpetuating bad practices and compromising the Fed's independence.

Lacker's position even back in March of 2026, before the House Financial Services Committee, he advocated that any new pact must legally force the Fed to adopt a bills only strategy. You can see highlighted what he's trying to say on page 11.

## US Government Debt Analysis - 24:03

On page 12, you can see the world's largest government debts in 2026, with the US twice as large as China, including its shadow debt and its provincial debt, and roughly four times the size of Japan, 10 times the size of France, 15 times the size of India. The president last week said that the United States is considering buying sizable amounts of Bitcoin on live television, which has yet to be supported as well, which helped push cryptocurrencies higher.

## Market Reactions to Treasury Yields - 24:31

The surge in the 30-year Treasury yield to multi-decade highs near 5.29% last week, its highest level since 2007, is driven by a confluence of fiscal, monetary and market structural pressures. The unprecedented fiscal deficits and Treasury supply: with a non-war deficit of 6%, the government faces massive ongoing budget deficits which will increase by half a trillion next year with requests to build our missile defense systems after what we've spent in the Middle East, requiring the Treasury to issue relentless supplies of long-dated bonds.

Following the Federal Reserve's decision to hold benchmark interest rates unchanged last month, Central Bank communication signaled a hands-off approach to guiding the yield curve. The market responded by steepening the curve, pushing long-term yields sharply higher while short-term yields held relatively steady. Sticky inflation readings have diminished expectations for aggressive rate cuts, forcing market participants to price in higher long-term neutral rates. Investors are demanding additional compensation or term premium to hold fixed income over a 30-year horizon given long-term purchasing power risks.

Crowded capital markets and alternative debt: corporate issuers, particularly major tech companies raising trillions of dollars to finance massive AI projects, have flooded the corporate bond market. This creates direct yield competition for sovereign US debt, forcing US Treasury yields to stay higher for institutional buyers like pension funds and insurers. You can see the US 30-year over the last 20 years effectively peaking last week. The US federal debt, as you can see in the following graph on page 14, has reached 40 trillion. I had expected yields to effectively start to peak, but the new expenditures of 600 billion on missile defense systems and a new military budget which is expected to be over a trillion and a half has raised deficit expectations over 3 trillion next year and it's made it harder to find a peak in yields.

## JPMorgan and Evercore on Bond Buybacks - 26:42

Total interest on US debt is now 1.4 trillion, which is actually higher than our present year's defense spending. On Wednesday, JPMorgan issued a warning that the Treasury's expanded bond buybacks will lower long-term yields only temporarily, it won't solve the real problem. The US is running a 6% budget deficit near full employment, national debt has surpassed 40 trillion and JPMorgan estimates a funding gap exceeding 3.5 trillion in the coming fiscal years, the highest deficit in history. Without meaningful fiscal reform, attempts to suppress yields will continue to undermine the Treasury's regular and predictable approach, raise the term premium, and ultimately push long-term borrowing costs even higher. They call it just a bandaid. And you can see there's an immediate impact on long-end yields on page 15, but then on the following page, page 16, the 10-year yield then quickly within two days broke that decline. Evercore is also skeptical that the administration can realistically do anything at this point on the deficit that would be material. We'll see what Bessent has to say tomorrow at 2:00 p.m.

The US right now is spending 100 billion more on debt interest than defense, according to Yardeni. Since Bessent took office and the Iran war started, the 10 year's gone from 4% all the way to 4.7%. And the US Treasury planned to sell a net 97 billion of long-term bonds in the market versus the 4 billion of buys, which shows that the buys are really de minimis.

## Jefferies' Analysis of Treasury Buybacks - 28:04

Zervos, the head of macro at Jefferies, wrote this on bond buybacks. The US Treasury's announcement to upsize its long-end buybacks on Wednesday has created plenty of controversy and confusion amongst the financial markets. To dispel some of this confusion, he has outlined below some of the bullet points on how he sees operations impacting the macro landscape. Further, given the growing number of market surprises in August — the buybacks, the Yen intervention by Bessent and the numerous high profile losses — it's probably a good time to hold a firm line macro call in his opinion. The exact timing of the buyback announcement was a shock to some given that the refunding announcement was only two weeks ago. However, the tool of buybacks should be seen distinct from the long-standing practice of announcing stable and predictable auction schedules. Buybacks are tactical. They're designed to counter shifts in short-term market conditions. Alternatively, auction schedules are built to foster secular stability.

Tactical buybacks were first floated by the Treasury Secretary in April of 2025 during this administration in response to aggressive foreign selling of US Treasuries. And even just the threat at that time worked effectively to settle yields after the large spike in the 10-year Treasury from 3.9% to 4.6% after tariffs. Given that experience, I would argue that Treasury holds all the cards here. Those fighting tactical buybacks will soon find out that there's only one bond vigilante out there and it's the Treasury Secretary who can continue to ramp up the amount of bonds they buy back. Tactical buybacks targeting the long end of the curve are nothing more than Operation Twist. The only difference from previous twists like Bernanke's is that historically the Fed has been the initiator of such actions. As of now, the Treasury has largely wrestled that tool away from the Fed. In some ways, the lack of Fed action back in April of 2025 when the rate market conditions went hyper-spastic almost surely led the Treasury to take a more forceful lead with twist operations.

While some folks are calling buybacks QE, this is not entirely correct. There are, however, some strong similarities. QE is the purchase of long-dated US Treasuries funded with reserve creation. And only the Fed can create reserves. However, you can think of a reserve as a one-day T-bill that only banks can hold. The mechanics therefore look very much like a twist operation in so far as long duration bonds are swapped for short duration bills. And the market implications are also quite similar with regards to lowering of term premia. The difference with QE, however, is that the printing of new high-powered money shifts inflation expectations. With QE, there's the creation of new government liabilities that are unconstrained by fiscal considerations like a debt ceiling. Thus, QE should be seen as more powerful and inflationary, given that it can be used to monetize the debt. QE generates more of a psychological effect with inflation expectations than a tactical Treasury twist. What the Treasury is doing is simply issuing short-term bills to buy long-term bonds. So they're not creating money. That's the primary difference.

That said, these specific Treasury twist operations do create some room for fiscal expansion. And as such, they also have the potential to generate some QE style reflationary or inflationary impact. Remember, the Treasury is buying long-dated bonds with $50, $60 and $70 prices, old issues with low coupons. This LME reduces overall debt outstanding and leaves room for fiscal expansion. With less fiscally binding constraints for the Treasury, the market may look at these operations in a more similar vein to fully fiscally unconstrained exercises like QE. To that end, the power and the punch of these Treasury twists may be even more than many in the market are predicting.

He ends with a last observation on the buyback announcement. Thus far, the range on the 10-year yield since the presidential election in 2024 has been 90 bips. Going back in time, it's very difficult to find periods of nearly two years with this much stability. One would need to go back to the conundrum period in the mid-2000s to find anything similar. So for all the talk of dollar debasement, credibility risk around Fed independence, fiscal recklessness, tariff Armageddon, geopolitical risk and every other Chicken Little style warning, a 90 basis point range in the tens for some 22 months is quite remarkable. Risk markets love stability in long end rates, and this administration fully recognizes that. The tactical buybacks are yet another business-friendly policy like deregulation that should support risk assets and economic growth more broadly. Now, in my opinion, as long as Bessent in the 2:00 p.m. speech tomorrow is able to cut the deficit and increase the amount of bonds he buys back, because 4 billion in an auction simply isn't enough.

## The Big Tech Bond Paradox - 32:49

Now, let's talk about the big tech bond paradox because US tech is also issuing trillions of debt. So the US federal debt, as you know, hit 40 trillion. The only way to deal with this is growth because no political party is ready to cut spending. So how are we getting that growth? Well, it's AI CAPEX. That CAPEX is coming from big tech companies like Meta, Google, Microsoft, Amazon and Oracle. But then how are those big tech companies funding their CAPEX? Well, initially, it was free cash flow, but that dried up. So now it's by issuing bonds. So Google issued a 100-year bond with a 6-year yield. Meta and Oracle have issued bonds with 6-8% yields at the long-dated side as well. The US 10-year Treasury yield is now above 4.7%, the 30-year is above 5.2%. The US sells these treasuries in order to fund the government. People then buy them with the expectation that this is the safest return they can get because the US will never default on their debt. The yields that are currently being offered are the highest in decades. So what's the problem? Well, the big tech companies are giving bond yields that are two or 300 basis points above what the US Treasury is offering, which is an issue because if you are a credit investor and don't think that Meta and Google are going out of business, why would you not buy their debt over US government debt?

As a result, people are selling US Treasuries to buy hyperscaler debt, causing yields to go higher, and then buying big tech corporate debt. The paradox in all this is that Big Tech needs to issue this debt in order to continue to spend on CAPEX, and that same CAPEX growth is what's supposed to solve our debt issues. If Big Tech stops spending on CAPEX, we wouldn't have any growth. But in order for them to get the growth, they have to issue bonds with high coupons and take away money from the long end of the curve of US Treasuries, causing the highest yields we've seen in 20 years. The simple way to resolve all of this is to end the Iran war because oil prices will go down, inflation expectations will go down, and credit markets will buy up US Treasuries yielding 4.5 to 4.7% since they will be getting a great yield on lower inflation expectations. The problem is the war hasn't stopped for months and the market doesn't think it's stopping anytime soon, which means until it does, we have to deal with higher oil prices, higher yields on bonds, and more uncertainty for stocks if the credit market continues to scream that yields need to come down. And now on top of this, we have a trade war with Canada. And by the way, right after we see this November midterm elections, you better believe we're going to have even more trade wars. So it's just complete nonsense coming from the executive branch of government, which is working against the Treasury, which is working against the Fed, and which is working against the American people.

## Impact on Gold and Precious Metals - 35:22

Now, how Bessent's remarks could affect gold and precious metals. When an official actor reveals where their pain threshold is, and then sizes the defense too small to be decisive, the market's rational move is to lean on it until they escalate or fold. If you take the Sterling in 1992, the Swiss National Bank's Euro peg in 2015, or the Bank of Japan's yield curve control in 2022 to 2023 when Japan yen volatility spiked, two of those folded. The Bank of Japan instead escalated repeatedly. The Bank of Japan ended up owning half of the JGB market, which is probably what's going to happen in the US. And the pressure never disappeared. It just used the Yen as a release valve, which means the dollar should also fall if we go in that direction, which is positive for gold, which is positive for silver, which is positive for real estate and agricultural land, and it's also positive for crypto.

## Analyzing the Treasury's Actions - 36:18

Now, price what the Treasury has announced: 4 billion minimum per operation, a fixed calendar, cap size, call it 15 to 30 billion across the whole window against a market that absorbed 742 billion of treasury sales in a single week this month. And a CBAC warning of about 1.45 trillion funding gap in addition between fiscal 27 and 28. Precious metals are already looking past the flow and pricing the escalation path. Programs like this have a habit of growing. You can see GLD spiked from 370 all the way to 420, with gold in ounces rallying to about $4,700 per ounce last week, and silver rallying back to about $69 an ounce. And copper also rallied to about 6.6 a pound. And copper stocks like HBM have also done quite well, which we are holders of.

## White House Crypto Summit Recap - 37:12

Then we have the White House crypto summit recap. President Trump, we're on page 20, said that the US is considering buying sizable amounts of Bitcoin and other cryptocurrencies, which has not yet been approved. He's relying on Paul Atkins of the SEC and other experts to make this decision. Trump is calling on Congress to pass the crypto clarity act, which is not going to pass because of all the corruption in the executive branch. Trump has also said that the US is ensuring it remains the undisputed leader in Bitcoin and crypto. Hyperliquid's HYPE pumped 15% to 69 after Trump says the CFTC is working to bring it to the US. SEC chair Paul Atkins said that we will ensure the greatest advances of the technological frontier are realized right here in America. The Gemini co-founder said America should lead crypto and win the market. Of course they would. And Trump says he ended the war on crypto once and for all, like the 50 other wars he said he stopped.

## Consumer Spending Slowdown (Goldman Analysis) - 38:05

Consumer spending is slowing according to Goldman this week, Goldman Sachs Asset Management. The US consumer spending growth is forecast to slow in the second half of 2026 after a surge in tax refunds in the spring likely provided a temporary boost to the consumer, according to Jan Hatzius. Their economists expect real inflation adjusted US consumer spending to grow only 1 to 1.5% in the second half of 2026, down from 1.8% in the first half. With the Strait of Hormuz still closed, elevated energy prices could pose an additional risk to the outlook for consumer spending. A renewed spike in gasoline prices which are already high because of crack spreads would further hurt consumers, especially those with lower and middle incomes. Strong business investment and the effect of earlier large equity wealth gains should continue to support overall US GDP growth, albeit at a pace slightly below its potential. GS forecast US real GDP to expand at only 2.1% this year.

In addition, Hatzius notes that US inflation data has improved meaningfully over the last two months because of real estate prices coming down, for example, and with temporary drivers like tariffs, software and accessories and energy that have kept inflation elevated set to fade, we'll see. GS says it doesn't expect the Federal Reserve to hike rates in 2026 or in the upcoming meeting in September. In a sign that policy hawks have become louder, three members of the FOMC dissented in the vote to hold the key policy rates steady this July. But after two months of materially softer jobs and inflation data, it's hard to see any of the doves shifting towards hikes, Hatzius writes. He notes that it's now very unlikely that the FOMC will hike rates in its September meeting, and the market pricing for the key policy rate seems too hawkish. The US Treasury's plan to double its purchases of long-dated government bonds could prompt investors to allocate more money to longer maturity securities in their opinion. So Goldman's taking a very dovish approach.

The announcement is part of a bigger story as the Treasury takes a more activist approach in managing the issuance profile of government debt, including signals that it may tilt future issuance towards shorter maturities and use those proceeds to buy longer term maturities, according to Mike Mitchell, head of US Treasury and inflation trading at Goldman. If the Treasury repurchases the full amount outlined, it would buy back a third of the long-dated debt it issues, Mitchell told Tony Pasquariello, the global head of hedge fund coverage in a podcast. It remains to be seen whether a change in the issuance mix will have a lasting impact on the trajectory of longer maturity Treasury yields according to Mitchell. Investors are demanding more term premium or compensation for holding longer dated debt amid fiscal risks and heavy supply of longer maturity debt from governments and companies. With the term premium around the middle of its long-term range, he expects it to drift higher, which is contrary to what he said earlier. A more durable change would likely require a fiscal consolidation through some combination of reduced spending or higher taxes or further good news on inflation.

The latter could reduce market expectations of a rate hike in the medium term. Otherwise yields on longer dated maturity treasuries could climb until they reach a level where there's enough of a value proposition for investors or they damage growth. Mitchell thinks that this level could be 5.2 to 6% on the 10 year.

## PRTH Special Situation Analysis - 41:23

Last week we added a starter position in PRTH at 552 per share. The company founder actually bid over $6 a share to take the company private, which is only five and a half times cash flow. We think that this was a lowball bid. There's an activist in the company called Buckley Capital Advisers who sent an open letter to the board special committee stating that the $6 bid drastically undervalues the company. Buckley released a sum of the parts model valuing PRTH at $15 to $20 per share. For those of you who don't know, Priority Technology Holdings or PRTH provides payment and data processing services. The company offers consumer and commercial payments, automated accounts payable, integrated banking and payments infrastructure and electronic fund transfer solutions. The company's based in Alpharetta, Georgia, has a $500 million market cap, has a total enterprise value of 1.4 billion, has 900 million of debt, which is the main issue, which is about four times debt to EBITDA. The company is a cash cow. It did about 75 million of free cash flow in 2025, is expected to do 91 million of free cash flow this year. 91 million on its 460 market cap is a levered free cash flow yield of about 20%. So it's arguably quite cheap where we bought it, as long as the company continues to grow.

The chairman and CEO of the company, Thomas Priore, alongside investors, submitted a non-binding take private proposal to acquire the remaining stock he doesn't already own at $6 to $6.15 per share in cash. He already owns about 55 to 58% of the company. Priore submitted the bid right after PRTH released a soft Q3 earnings on purpose, which caused the stock to plummet from 7 down to 487. The $6 to $6.15 offer was anchored right between the pre-crash price and the post-earnings lows, effectively creating a low premium environment above the distressed level.

In addition to Buckley Capital Partners, which stated the company's worth $15 to $20 a share, Steamboat Capital Partners also jointly submitted a formal letter opposing the takeover at $6 a share. Independent conservative estimates place the fair intrinsic value of this company closer to $10 a share, which is a 40% premium to where it — more than, sorry — based on the 561 close, $10 would be a 78% premium to the close.

Now, the board's special committee is tasked with negotiating on behalf of minority shareholders, and they face intense public and legal pressure from activists not to approve the $6 offer without extracting a bump. With the stock hovering at around 561 at about five and a half times cash flow, the founder's bid acts as a structural price floor, at least a small discount to that bid. The primary upside driver: there's some upward asymmetry here because the minority shareholders are represented by the special committee, you could easily see a bump to a $7 or $8 offer from the CEO. A $7 offer versus the 561 would be roughly 25% upside. An 850 on the 561 Friday close would be a 50% upside. So it's quite asymmetric risk reward here.

If you look at the following page, you can see where we bought it in the middle of the day on August 21st. We're positioned to buy more. The only risk here is you can see that the company does have a moderate amount of debt at about 1 billion minus the 120 of cash. So roughly 920 of net debt divided by the 230 of EBITDA this year, it's about four times levered, which is not egregious, but it is moderately high leverage.

## Citigroup as a Value Play - 45:17

Another name which is less risky on pullbacks that we think is interesting is Citigroup, which is still a value stock, trading at below 1.2 times price to tangible book value and nine and a half times forward PE, a big discount to peers. The company's cleaned up its balance sheet, it's building its wealth management franchise. And if you look at multiples, for example on PE, Goldman trades at roughly 14 to 15 times earnings on a forward basis. Morgan Stanley currently trades at 214 a share, which is roughly 16 and a half times forward PE. So this stock trades at about nine and a half PE, so it's still like a 50% discount to Goldman and Morgan Stanley and remains the deepest discount for money center mega banks. Now, we normally wouldn't be buying this, but if you think that interest rates are going to stay high, banks should benefit and Jane Fraser has done a wonderful job turning Citi around. So we think that if there's a market pullback due to high interest rates, Citi could be an interesting name to add. We have a model with key drivers documenting the company's turnaround that we've shared in the SSR Q&A tab.

## Iran War Update - 46:38

Now, in the Iran war update, Trump advised officials to stop negotiating with Iran last week after the latest deal with Oman and the risk of an offensive strike. You can read this very long rant on page 26.

## Goldman on Oil and Refinery Runs - 46:52

Goldman is also now putting up a red flag on oil and refinery runs right as Canadian tariffs restart. The worst thing Canada could do — because we buy all our sour crude and cheap crude from Canada for our complex refineries — the worst thing Canada could do in this tariff war is to stop sending oil here. But what Goldman is saying is that they've been seeing diverging margins and runs. Refined product supply and product margins typically move together along the supply curve as demand shifts, but an unprecedented supply constraint has driven sharp divergence. Global refinery runs are down 7 million barrels a day, while margins remain near record highs.

What they're saying is that capacity outages for refineries are putting a ceiling on runs and that's what's driving crack spreads up despite oil prices remaining moderately low. That's why gasoline prices are so high, because crack spreads have been driven higher. Margins and runs usually move together. So in the short run, elevated refinery runs tend to reflect refiners' expectations of strong demand and high profitability and therefore coincide with higher margins. The positive relationship between refinery runs and margins reflects the typical upward sloping supply curve. Higher margins in turn incentivize refiners to lift supply, leading to builds in refined product stocks that eventually rein in margins, but we're not there yet.

A $10 per barrel rise in the US average of gasoline and diesel margins tends to boost global refinery runs by .6 to .8 million barrels per day the following month. But because of the issues we have in the Middle East, these refinery runs have not increased despite refinery margins being very high.

## US-Canada Trade Talks Collapse - 48:36

So US Canada trade talks basically fell apart at the last minute this week with 50% fresh tariffs on billions of dollars of Canadian goods taking effect and the Prime Minister Mark Carney promising to retaliate. The US tariffs kicked in on Saturday on hundreds of items the US buys from Canada, such as plywood, liquor, electrical equipment and hockey gear, totaling around 20 billion. Carney said he suspended talks and his government would match these duties dollar for dollar to protect his workers. One of the issues is that Canada imports a lot of goods from China and the US is trying to protect itself from trans shipment.

The two sides blamed each other for the collapse. The US Trade Representative Jamieson Greer and the Canadian negotiators made 11th hour demands that upended a draft deal worked out over days of negotiations. Key drivers behind the breakdown: section 338 catalyst, the punitive 50% tariffs from President Donald Trump invoking section 338 of the tariff act of 1930, which was a depression era statute to target countries deemed to have discriminated against US commerce. The US is targeting 20 billion of goods from Canada, or about 28 Canadian dollars worth of goods. Specific sectors include dairy, liquor, automobiles, plywood, electrical equipment, and hockey gear. The move zeros in on areas where Washington claims Canada unfairly penalizes US markets, notably provincial bans on US alcohol sales and dairy market access restrictions. This tariff has no USMCA carve outs like the other ones did in the past, which is likely worse in this tariff war.

Jamieson Greer in the US stated that Washington offered Ottawa the best treatment of any major exporter, which included proposed cuts to tariffs on Canadian steel and aluminum down to 25%, autos down to 15%. Greer accused Canada of scuttling the deal by introducing new 11th hour demands, backing out of prior commitments and refusing to lift its own retaliatory measures on American products. Prime Minister Mark Carney suspended negotiations and recalled his diplomatic team, stated that last-minute changes in the US proposals were unfair, uneconomic, and called into question the reliability of any deal. Carney emphasized that Canada would not accept a deal at any price and vowed to match the 20 billion tariffs dollar for dollar with retaliatory measures set to take place on September 8th, targeting US steel, dairy and consumer sectors. The targeted economic damage, while 20 billion is not huge, it's 5% of Canada's total exports to the US, it's just the start. Trump has said that he would escalate if this is not turned back.

## Details of the US-Canada Trade Dispute - 51:18

You can read Carney's objective on the press release on page 29 and Greer outlines the deal that Canada rejected on page 30 to see both sides. In an interview, Jamieson Greer, President Trump's trade representative laid out details of what the US had offered Canada before the deal crumbled. The US had offered to reduce its tariffs on steel, aluminum and autos, eliminate a recently imposed tariff on Canadian lumber before negotiations suddenly collapsed. In an interview with the New York Times, Greer detailed previously confidential and unreported elements of the US trade offer to Canada, saying those measures would have given Canada the most preferential treatment of any trading partner.

Trade talks between the countries suddenly soured late Friday night following a week of talks in which negotiators appeared hopeful for an agreement. Mr. Trump abruptly threatened Canada with a 50% tariff on about 20 billion of its exports. Those tariffs were set to go into effect on August 19th, but on the night before, Trump delayed the tariffs until August 22nd. But over following days and particularly late night through Friday, the consensus crumbled. Greer said the Trump administration had pledged to eliminate the 10% tariffs on softwood lumber that Trump imposed last year using a provision known as 232. The US had also offered a 25% tariff reduction that Trump put in place last year. The US offered to lower tariffs on metals. On aluminum, the US had offered to reduce its tariffs from 50% to 25%. For products made with steel and aluminum like golf clubs and beer cans, the United States offered to cut those tariffs anywhere from 10 to 25%, with the lowest tariffs set at 15%.

Greer said the US had also offered to suspend the 50% tariffs that it brought into effect early Saturday morning on Canadian dairy, wine, and hockey sticks and other goods. Obviously, we would completely suspend the 50% section 338 tariffs that were going to apply to 5% of Canadian exports, he said. Carney said that Canada had walked away from negotiations not just because it assessed that what was being offered by the United States negotiation team was insufficient, but also because US officials were asking for concessions that Canada was not prepared to make. So these negotiations are likely still happening, but what's worrying to me is that this is all happening while the Iran war is taking place with inflation fears, term premium fears, and US 30-year yields and 10-year yields making new highs. If anything, this tariff war should have been paused until the Iran war was resolved. But unfortunately, we have tariffs and oil prices being a risk going into midterm elections now.

## US ETF Flows and Market Euphoria - 54:14

Despite all of this, US ETF flows have accelerated this year, with retail euphoria driving much of this. Investments in US listed exchange traded funds on page 31 are expected to top 2 trillion this year, a 40% jump over 2025 according to Goldman Sachs. In the first half of 2026 alone, investors poured more than 1 trillion into US listed ETFs. The surge is being largely driven by an unprecedented wave of product innovation, including levered ETFs and tax efficient ETFs. More than 1,100 new ETFs launched this year, and that record is on track to be broken by year end, pushing the total number of US listed ETFs past 6,000. There are going to be more ETFs than stocks in the Russell indices, which is unbelievable. We're now seeing some of the most advanced active management strategies in the markets expressed within the ETF wrapper. These range from levered funds to innovative fixed income offerings to structured derivatives. And you can see the pace of ETF inflows on page 31.

## Weekly Market Recap (Day by Day) - 55:18

Now to do a quick weekly recap day by day from Monday through Friday of the most important events of last week. On Monday, Nvidia was backing OpenAI's Ohio campus with a $1.5 billion investment in SB Energy because energy is the main bottleneck, in support of 4.25 gigawatts of AI infrastructure. Anthropic's revenue run rate reportedly surged ahead of its IPO, rising to more than 65 billion in July of 2026 from 47 billion in May and 9 billion at the end of 2025. The company also posted over 11.5 billion in preliminary revenue for its latest completed quarter, up from 787 million a year ago, while generating positive operating income and EBITDA. The company is supposed to IPO in October.

US interest expense on the national debt reached a record 1.4 trillion with debt servicing costs nearly tripling since 2020. If rates stay near current levels, interest payments are projected to rise to 1.7 trillion by November 2028, which would be the highest percentage of tax revenues in history, putting it on pace to overtake social security and national defense expenses. This is why we have to cut rates. Evercore ISI's Julian Emanuel says the S&P 500 could reach 9,000 over the next 12 months, while estimating a 7750 base case target, which is basically where we are now. He argues that the usual bull market killers — recession, sharply higher long-term yields, and extreme investor chasing — still have not shown up. I would argue that sharply higher long-term yields are a big risk. At the same time, 121 S&P 500 stocks now have negative beta to the index, the highest number since the 2000 to 2001 dot-com unwind.

Fabrinet, FN, reported revenue of 1.32 billion, which was a big beat last week. The top 10 most active options contracts were in Nvidia, Tesla, Micron, SpaceX, Apple, Amazon, Meta, Microsoft, Intel, and AMD last week. Tesla reportedly is preparing to launch its purpose built Cybercab in Austin as soon as this month. Uber is investing in Zipline as the companies expand into drone delivery for Uber Eats, which is quite exciting. The goal is to reach 1 million drone deliveries by the end of 2029. China's credit data weakened very sharply in July with new loans falling by 50 billion, only the third monthly decline this century and more than three times worse than expected, which shows you how weak the Chinese economy and household demand is, which I think is going to result in some sort of a Chinese stimulus announced later this year.

Morgan Stanley sees Amazon with a bull case path to 500 by 2027 driven by AWS potentially scaling towards 1 trillion in annual revenue over the next eight to 10 years. UBS expects Nvidia to beat its quarter two revenue estimates by roughly three to four billion next week, with revenue reaching 94 to 95 billion. It also sees guidance of 107 to 108 billion. And analyst Timothy Arcuri sees Blackwell demand remaining stable while Rubin units are starting to layer in ahead of a bigger fiscal fourth quarter step up as Rubin gets to 500,000 GPU units per month. On Tuesday, Nebius got approval for its Vineland, New Jersey data center expansion, a major hurdle tied to its $17.4 billion Microsoft cloud deal, which is insane.

Tuesday's sell-off was driven by macro pressure. The VIX had been sitting near a two-year low, leaving stocks vulnerable to any pickup in volatility with the 10-year spiking to 4.7%. Trump said the US was not holding talks with Iran. The move was especially sharp in tech with many semiconductor names on Tuesday down 8 to 10%. Anthropic is reportedly preparing to give CEO Dario Amodei and co-founders enhanced voting rights ahead of the potential IPO in late September, October. OpenAI said revenue reached 6.7 billion at the end of Q2, up only 18% and the market was very worried that OpenAI's growth rate is decelerating given hundreds of billions of dollars it needs to spend on CAPEX.

July economic data came in mixed. Import prices fell .4%. Housing starts came in at a very low 1.239 million versus 1.35 million due to mortgage rates, down 12%. ADP said private employers only added 9.5 thousand jobs per week. UBS remains bullish on Micron with a buy rating given price increases and future capacity additions. Onchain tokenization equity trading volume hit a record 9 billion in 2026, given the tokenization of stocks, with Jupiter leading the Solana based tokenization equities, 95% quarter over quarter. Bank of America's August fund manager survey shows investor bullishness approaching extreme levels, which is quite dangerous. A net 56% of managers overweight equities, the highest since November of 2021.

Google has reportedly told suppliers it plans to move production of all Pixel smartphones, watches and earbuds out of China starting in 2027, mainly to Vietnam and India, which is going to be a big boom to the Vietnam and Indian economies. Morgan Stanley says active fund managers remain underowned in MegaCap tech, with the gap widening to -129 basis points in Q2. Nvidia is actually underowned by big funds according to them. On Wednesday, Moderna and Merck said their personalized mRNA cancer vaccine intismeran, combined with Keytruda, reduced melanoma recurrence and spread in late stage trials. The phase three study was stopped at its first interim analysis after positive results, which resulted in a 175% surge in MRNA stock, which we shorted and, from the peak, that stock fell about 30%. We covered it, I think we made about a high single digit percentage because we covered the first half around break even.

Crypto rallied very sharply on Wednesday. Bitcoin jumped roughly 8%, briefly touching 78,000, and then rallying to 77,000 later in the week. The move liquidated an estimated 1.5 billion of short positions, driven by a mix of low yield pressure from Treasury buybacks, regulatory optimism, a decline in the US dollar, and the White House crypto meeting. Nebius announced a proposed private offering of 4.5 billion of convertible senior notes, which put pressure on the name given the convertible hedging — when you issue a big convert, there's usually convertible arbitrages that buy it and short some of the stock. On Wednesday, the US national debt also crossed about 40 trillion with 10 trillion just added in the last 4.5 years.

The US Treasury announced it would double its long-term government bond buybacks on Wednesday. Starlink also applied for an India approval of its gen 2 network. Goldman is preparing for roughly a $1.15 billion junk bond to finance Virginia data centers linked to CoreWeave. Nvidia reportedly discussed investing in AI data supplier Mercor on Wednesday, that would value the startup at 20 billion. Weibo reported a big beat on earnings. President Trump is said to reduce tariffs on automobiles from Canada, but that reversed today. US margin debt actually fell by 85 billion in July to 1.42 trillion, the largest monthly decline ever, given the big deleveraging in semiconductor stocks and the same deleveraging we saw in Korea. But that margin's already going up again in August.

On Thursday, crypto markets saw their seventh largest short covering event ever with 3.5 billion in leverage positions wiped out in 24 hours. Despite the forced liquidations, the total crypto market actually added roughly 280 billion in the same period as Bitcoin crossed 72k and Ethereum moved to 2300. On Thursday, Treasury Secretary Bessent said the US Treasury will assess bond market conditions and take further action if needed. The market didn't like that. Anthropic is reportedly adding Citigroup to its IPO bank lineup, making the IPO more real with Morgan Stanley, Goldman Sachs, and JPMorgan among the lead banks working on the listing. Anthropic is actually trying to raise more than SpaceX.

Deutsche Bank reiterated its buy rating and 200 price target on Palantir. Amazon plans to invest more than two billion across Latin America through Prime video, focused on original content and live sports. Super Micro said an independent investigation found no evidence that current senior management knew about the alleged export diversion scheme involving two former employees and a contractor, basically selling Nvidia GPUs to China. Well, I'm sure that's true. Huh. Walmart fell 8% on Thursday after a slow same store sales print. The worst day since May of 2022 and reporting its slowest US comp sales growth in more than six years. US comps rose just 2.6%, missing the 3.7% estimate, the lowest comp sales estimate in several years.

SpaceX and AST SpaceMobile reportedly interested in acquiring 800 megahertz spectrum licenses valued at about 6 billion. The preliminary bids are due in early September. The median value of US consumer stock market investments climbed to roughly 350k per person, including holdings in individual stocks, mutual funds, 401Ks and IRAs per Schwab. That figure has more than doubled over the last two years and is up 150k or 75% since January alone, which is unbelievable. Micron is committing 10 billion over the next decade to build Micron research labs, a new long-term US research hub based in Boise. Micron has already planned 250 billion of US spending on R&D and CAPEX. And Broadcom is reportedly seeking more than 60 billion in financing tied to its latest AI deal, which resulted in a huge spike in its CDS, with the full package potentially nearing 100 billion. The structure could include 60 to 70 billion of senior secured debt plus 30 billion of junior debt.

On Friday, crypto markets saw their seventh largest liquidation. Alright, let's jump to page 37. UBS expects further volatility. After July FOMC, they stressed caution ahead of a busy macro calendar over the next several weeks, especially amidst a seasonally volatile and liquidity deficient period. September FOMC and midterm elections are the two notable risks that have risen in terms of pricing. Lastly, investors who believe midterm election risk pricing might continue to rise or remain elevated should sell the VIX October puts to buy November puts. S&P 500 stocks with dividend yields greater than the 10-year Treasury yields hit a record of 63.4%, excluding the COVID-19 crash back in July of 2016. So effectively, stocks are the most expensive to bonds than they've been since the great financial crisis because of where yields are today. Goldman is showing high refinery outages in the Middle East and in Russia being the key drivers of low global refinery runs and crack spreads. That's on page 39. The MRNA melanoma cancer vaccine driven short squeeze recap is on page 40 and my reason for shorting Moderna stock down to 140 is written up there in a paragraph. We think that BioNTech could be a better short, which was up 22% for no reason. We did cover half of it. It did fall to like 107 and then it bounced back at the end of the week. We might add to that short again. Small cap squeezes across biotech and crypto: you can see all the names that have squeezed last week on page 41, maybe to look at some ideas. Duolingo gained after inadvertently disclosing user growth data. So that was a big leak, on August 18th. All correlations are breaking down according to Goldman. Long-short correlations have been breaking down all year. You saw that with the Citadel acquisition of the situational awareness portfolio, which by the way, they've already sold 80% of this past week. JPMorgan's Brian Hevey explains why tech momentum plunged on Tuesday. The main driver of the momentum reversal was that Anthropic's 65 billion ARR number was below various data sources that implied higher growth of 75 to 80 billion.

Walmart's big miss in same-store sales, you can see graphically on page 43 and according to the RSN Focus Fund, the highest concentration in the top 10% of US companies which owns 78% of US market value. That graph is also on page 43 and on page 44. Despite all the fear we have about government debt, US corporate debt as a percentage of market cap is still very much contained. SK Hynix upgrades because of the sharp price increases in NAND and DRAM and the fact that SK Hynix announced a surprise 30 billion share buyback, caught the market off guard. JPMorgan has now kept an overweight on SK Hynix because of that massive buyback plan. And Alibaba announced an 80 billion Hong Kong dollar capital raise, a 10.2 billion US share placement in Hong Kong. The company will be raising cash just like US hyperscalers to invest in AI infrastructure.

## Q&A Section - 01:07:06

So quite a lot that we covered during this call. I do want to get to some Q&A. The interesting special situations: we talked about PRTH and Citigroup, which should benefit from higher interest rates if rates continue to go higher, as being one of the cheapest money center banks. Remember for that one to buy on pullbacks. We might short some more BioNTech. Warner Brothers has played out really well with the spread compressing from 20% to 8.8%.

Let's just quickly go to the Q&A. I don't see any questions. So what I'll do is I'm going to upload the recording and if you have any further questions, you can just throw them in the Q&A tab and I'll address them in the morning. Thanks for listening. Have a wonderful evening and a successful trading week this week.