Jay Singh — Weekly SSR: Bessent loses control of the 10-year as the debt crosses $40 trillion, gold's best week in a year and a three-sigma crypto move, Citigroup as the $165 rate beneficiary to buy on pullbacks, PRTH started at $5.52 against a $6 lowball take-private, and WBD's spread compressing 20% → 8.8%

Priority Technology is a payments plumbing company in Alpharetta, Georgia. It processes card payments for merchants, automates the bill-paying side for businesses, and rents out banking-and-payments infrastructure to other companies. Dull, but it throws off cash: about $75 million of spare cash last year and roughly $91 million expected this year, against a stock market value of only about $460 million. That is a cash return of roughly 20% a year on what you pay for the shares — cheap by any normal standard. The catch is around $920 million of borrowings, about four years' worth of earnings, which is high but not dangerous.

The reason it is interesting right now is a fight over who gets that cash. The founder and chief executive, Thomas Priore, already owns 55-58% of the company, and he has offered to buy the rest for $6.00-$6.15 a share in cash and take it private. Singh's objection is not the idea but the price and the timing: the offer came immediately after a weak earnings report knocked the stock from $7 to $4.87, so the bid was pitched into the hole it had just fallen into. At $6 you are being paid about five and a half times the company's annual cash generation — for a business that is still growing.

Two professional investors have publicly refused. Buckley Capital wrote to the board's special committee — the independent directors legally obliged to represent the shareholders who are not the founder — saying the company is worth $15-20 a share if you value each of its businesses separately. Steamboat Capital joined the objection. Even conservative estimates put fair value near $10.

That produces an unusually lopsided bet. The $6 bid effectively puts a floor under the shares, because a controlling holder who wants to buy is unlikely to let the price collapse below his own offer. Meanwhile the pressure on the committee makes a raised bid plausible — $7 would be about 25% above where Singh bought at $5.52, $8.50 about 50%. He calls that "quite asymmetric risk reward," has taken a starter position, and is ready to add.

Citigroup is one of the four giant American banks. For a decade it was the problem child — too many businesses in too many countries, poor returns, a share price stuck below the accounting value of its own assets. Under chief executive Jane Fraser it has been selling off the international consumer businesses, simplifying the structure, and building a wealth-management arm that earns steady fees rather than volatile trading profits.

The stock still carries the old reputation. It trades at about nine and a half times next year's expected earnings, and below 1.2 times tangible book value — meaning you pay barely more than the accounting worth of what the bank owns after stripping out goodwill. Goldman Sachs trades near 14-15 times earnings and Morgan Stanley near 16.5 times. Singh's point is simply that the gap is roughly 50% and Citi remains the cheapest of the mega-banks.

The link to the rest of the call is the interest-rate view. Banks make money on the difference between what they pay depositors and what they earn on loans and bonds, and that gap generally widens when rates stay high. If Singh is right that the Treasury cannot force long-term yields back down, then high rates persist, and the banks are the natural beneficiary of the same force that is hurting bonds. His price target is $165.

He is deliberately unenthusiastic about chasing it — "we normally wouldn't be buying this." The instruction is to wait: if high rates cause a broad market sell-off, that is the moment to buy Citigroup, because the sell-off and the reason to own it are the same event.

This is a merger arbitrage: Paramount Skydance has agreed to buy Warner Bros. Discovery, and the WBD shares trade below the agreed price because the deal might not complete. That gap is the "spread," and it closes as the deal becomes more certain. Singh was buying when the gap was about 20%; it is now 8.8% — most of the money has already been made.

What changed this week was the cinema industry. Cinema United, the trade body for theatre owners, had opposed the merger, fearing a combined studio would release fewer films and charge exhibitors more. It reversed and joined the three biggest chains in support, after extracting promises: keep or expand the number of films given full cinema releases, do not raise the fees theatres pay, and keep both studios' film libraries available. Regulators find it far harder to block a merger when the industry supposedly being harmed says it is fine.

Two other details protect the downside. Paramount agreed to pay WBD shareholders $7 million a day if the deal is not completed by October 1 — a "ticking fee," which means delay pays Warner holders rather than punishing them. And Paramount has asked a court to make the twelve states suing to block the deal post a $1.9 billion bond covering that fee, which would make the lawsuit expensive to continue. California's attorney general still says the merger is illegal and is refusing, but as Singh puts it, the market is looking past him.

GLD is simply a fund that holds gold bars, so owning it is owning gold without a vault. Gold had its best week in more than a year, with the fund moving from about $370 to $420 and bullion reaching roughly $4,700 an ounce.

The reason is not fear of inflation as such but a judgement about a policy that will probably have to grow. The US Treasury said it would buy back its own long-term bonds to push borrowing costs down — at $4 billion per operation, perhaps $15-30 billion in total. In the same month the market had to swallow $742 billion of Treasury sales in a single week, and the government's own advisory committee warns of a roughly $1.45 trillion shortfall over the next two fiscal years. The defence is a fraction of the problem.

Singh's argument is that once a government reveals it is defending a level and brings too little firepower, the market keeps pushing until the government escalates or gives up. Japan escalated for years and ended up owning half its own bond market, with its currency absorbing the strain. If the US goes the same way, the dollar weakens — and a weaker dollar mechanically lifts gold, silver, farmland, property and crypto. "Precious metals are already looking past the flow and pricing the escalation path."

Hudbay is a Canadian miner of copper — and Singh confirms on this call that he owns it. Copper rose to about $6.60 a pound in the same week gold and silver jumped, and mining shares rose more than the metal because a miner's profit is the difference between a roughly fixed cost of digging and a rising selling price. That leverage cuts both ways, which is why the position sits alongside gold rather than replacing it.

The reasoning is the same one behind the precious metals. If a falling dollar is the release valve for an over-indebted government, then everything priced in dollars that cannot be printed — metals, land, property — is worth more in dollar terms. Copper adds a second engine: it is the physical input for electrical grids and data centres, so the same AI build-out that is straining the bond market also consumes copper.

RTX makes missiles and air-defence systems — Patriot interceptors, Tomahawk cruise missiles. The US Navy has just given it a $23 billion Tomahawk contract, part of a $500 billion emergency request to replenish stockpiles fired off during the Middle East campaign.

The plain point is that munitions are consumed. Every interceptor launched has to be rebuilt and paid for, and Singh's estimate is that the war has already burned roughly half a trillion dollars of ammunition against an official figure of $70 billion. When he adds the individual line items in the 2027 request — Patriot, THAAD and naval missile defence — they approach a trillion dollars. That is a multi-year order book for whoever manufactures them, and RTX is the named beneficiary.

The uncomfortable symmetry is that the same spending is his central reason for expecting deficits above $3 trillion and long-term interest rates that will not come down. The defence contractor is the direct beneficiary of the very bill that is hurting the bond market.

Micron makes memory chips — the components that store data while an AI processor works on it. An Nvidia chip is useless without a lot of memory attached, and only three companies in the world make it at scale: Samsung, SK hynix and Micron.

This week produced the clearest possible proof of where the power sits. Nvidia, the most dominant company in technology, told its largest customers that AI server prices are rising more than 15% because it cannot absorb the rising cost of memory. When a business with 75% profit margins passes the cost through rather than eating it, the supplier is the one setting the price. Memory prices for both types of chip continue to rise steeply.

Two supporting facts from the same week: UBS keeps a buy rating on Micron precisely because of price increases and new capacity, and Micron is committing $10 billion over ten years to new research labs in Boise on top of $250 billion of planned US investment — a company spending like one that expects the shortage to last.

SK hynix is the Korean memory maker that, with Samsung and Micron, controls almost all global production of the chips that AI processors depend on. It is on the same side of the trade as Micron: Nvidia's decision to raise server prices rather than absorb memory costs is an admission that the memory makers, not the chip designer, currently hold the pricing power.

What re-rated the shares this week was capital return rather than the chip cycle. The company announced a surprise share buyback of 30 billion — buying back its own stock, which shrinks the share count and lifts earnings per remaining share. It caught the market off guard, and JPMorgan kept its positive rating specifically because of it. A buyback of that scale is also a management signal: it is what a board does when it believes its own shares are cheap relative to what is coming.

Alibaba is China's dominant e-commerce and cloud-computing company, and it has now joined the American hyperscalers in raising outside money to fund AI infrastructure: an HK$80 billion (about $10.2 billion) share sale in Hong Kong, earmarked for data centres and chips.

The reason Singh stays constructive is policy rather than the raise. Chinese credit data for July was startlingly weak — new bank lending actually shrank, only the third monthly decline this century, and more than three times worse than expected. Weak lending means weak household and business demand, and in China that is usually the precondition for stimulus rather than a reason to sell: "which I think is going to result in some sort of a Chinese stimulus announced later this year."

So the position is a bet on two things arriving together: a policy response to a visibly weak economy, and a company already investing through the downturn in the one area — AI capacity — that global investors are willing to pay up for.

Nvidia designs the processors that train and run AI. This week it told its biggest customers — Microsoft, Amazon, Google, Meta — that the price of its AI servers is going up more than 15%, on the systems shipping early next year.

A price rise usually reads as strength. Singh reads it both ways. The company can raise prices because demand exceeds supply, which is genuinely good. But the stated reason is that memory chip costs are soaring and Nvidia cannot absorb them — and a business with 75% gross margins that still has to pass costs on is telling you that its suppliers, not it, hold the whip hand. Hence his conclusion that 75% is "likely a peak margin": the most profitable quarter is probably behind rather than ahead.

There is a second-order cost too. The customers building data centres are already fighting power shortages, construction delays and labour scarcity; a 15% price increase on the most expensive component makes those projects harder to justify. That is the seed of a slowdown in the very spending that has driven the shares. The stance is neutral, not negative: demand is real, analysts expect a $3-4 billion revenue beat, and big funds are still underweight the stock.

Broadcom designs custom AI chips for the largest technology companies. The item that caught Singh's eye is not a product but a financing: the company is reportedly seeking more than $60 billion of debt tied to its latest AI deal, with the total package potentially approaching $100 billion.

What matters is the market's reaction in the credit market rather than the stock market. Broadcom's credit default swaps — effectively the insurance premium other investors pay to protect themselves against the company failing to repay — spiked. That is bondholders repricing the risk of borrowing on this scale to fund equipment for a single customer relationship.

It belongs to the call's larger theme: the AI build-out has moved from being funded out of profits to being funded with borrowed money, and the credit market is beginning to charge for it.

Nebius rents out AI computing capacity — it builds data centres full of Nvidia chips and lets other companies use them. Its biggest contract is a roughly $17.4 billion agreement with Microsoft, and this week it cleared a real obstacle: local approval for the expansion of its Vineland, New Jersey site, without which the contract could not be delivered.

The share price nonetheless came under pressure, for a purely mechanical reason. The company is raising $4.5 billion through convertible bonds — debt that can later turn into shares. The funds that buy these bonds are not making a bet on the company; they buy the bond and simultaneously sell the shares short to neutralise the equity risk, harvesting the difference. That short selling pushes the stock down at issuance regardless of how the business is doing.

So the week contained one genuinely good event and one temporary technical drag, which is why the stance is neutral rather than a change of view.

Anthropic is one of the two leading American AI labs and is expected to go public in October, with Morgan Stanley, Goldman Sachs, JPMorgan and now Citigroup running the listing. Its growth is remarkable: revenue on an annualised basis went from about $9 billion at the end of 2025 to $47 billion in May and more than $65 billion in July, with the business now producing positive operating profit.

The problem sits inside those numbers. Its most advanced and most expensive model, Fable 5, has stalled at roughly 11% of what corporate customers spend with Anthropic. Companies are deliberately routing routine work to the firm's cheaper models, or to free open-weight models they can run themselves, and saving the expensive one for jobs where a mistake would be costly. The uncomfortable implication, in the report's words, is that "technical capability leadership" does not automatically become revenue — being the best is not the same as being paid for being the best.

There is a second warning in the same week: the $65 billion figure was actually below what other data had implied (75-80 billion), and JPMorgan blames that shortfall for a sharp reversal in technology momentum stocks on Tuesday. Extraordinary growth, an unresolved pricing model, and a valuation about to be set in public.

OpenAI, the maker of ChatGPT, reported second-quarter revenue of $6.7 billion — up only 18%. For a company that needs to spend hundreds of billions of dollars on computing capacity, a growth rate that starts with a one is a problem, because the spending is committed years in advance while the revenue has to keep compounding to justify it.

The scale of that commitment appeared in the same week from the other direction: Nvidia invested $1.5 billion in an energy developer to secure 4.25 gigawatts of electricity for OpenAI's campus in Ohio, "because energy is the main bottleneck." Getting power, not chips, is now the binding constraint.

Singh does not take a directional view — OpenAI is not listed — but treats the slowing growth as evidence for his broader argument that the market has stopped rewarding AI spending automatically and started asking when it pays back.

Paramount Skydance is the buyer in the Warner Bros. Discovery deal, so its actions determine what happens to Singh's arbitrage position rather than being a holding in their own right.

Two moves this week. It won over the cinema owners by promising to keep releasing films in theatres, not to raise the fees exhibitors pay, and to keep both studios' libraries available — which removed the industry's objection that regulators had been relying on. And it went on the legal offensive, asking a court to require the twelve states suing to block the deal to post a $1.9 billion bond, on the grounds that their lawsuit is what would trigger the $7 million-a-day penalty Paramount owes Warner shareholders for a late closing.

California's attorney general refused, arguing that Paramount is a sophisticated buyer that knew it faced regulatory review when it agreed to that penalty, so California should not have to underwrite its choice. The fight over who bears the cost of delay is exactly what an arbitrageur watches.

CCC Intelligent Solutions provides the software that insurers and repair shops use to process car-crash claims. Its shares jumped as much as 12% after Bloomberg reported that Copart — the online salvage-vehicle auction company, which sits in the same accident-repair ecosystem — is in talks to buy it, competing against two private equity firms, GTCR and Veritas Capital.

Singh flags it without taking a position, and the reason it is worth flagging is the shape of the situation rather than the company. A strategic buyer that can extract genuine synergies bidding against two financial buyers who cannot is the setup most likely to produce a real auction and a real premium. It goes on the board to watch for a formal proposal.

Pilgrim's Pride is one of the largest US chicken producers, and 82% of it is already owned by the Brazilian meat group JBS. JBS has now made a non-binding proposal to buy the remaining shares.

This is the same structure as the PRTH situation on the other side of the call: a controlling owner offering to buy out the minority. The tension is identical — the buyer wants to pay as little as possible, and because it already controls the company, there is no realistic rival bidder. What protects the remaining shareholders is the board's special committee of independent directors, whose legal duty is to negotiate on their behalf and refuse a price that undervalues the business.

Singh notes it as an event rather than a position, but the mechanics to watch are the ones he spells out in detail for PRTH: does the committee extract a bump, and do activists show up to make refusing easier?

Liberty Broadband's merger completed around August 20 — the deal Singh had been tracking is simply done, which is why it appears as a one-line item in the special-situations round-up.

For an arbitrage book this is the ordinary end state: the spread between the market price and the deal price disappears at closing, the position converts, and the capital is freed for the next situation. It is housekeeping rather than a new idea.

Brighthouse Financial, a US life insurer and annuity provider, has agreed to be bought by Aquarian Capital. Because insurers are regulated at state level, the deal needs approval from Delaware's Department of Insurance, and that regulator gave a progress update this week.

The substance of the update is that the deal "has generated significant national and international interest," along with questions about how the department conducts its review, and that it has hired outside specialists to help evaluate it. Insurance regulators are ultimately protecting policyholders — people whose retirement income depends on the company still existing in thirty years — so a private-capital buyer attracts scrutiny about who ends up owning those promises.

None of this changes the odds; it is a reminder that this particular spread is governed by a regulatory calendar, not a commercial one.

Union Pacific is attempting to merge with Norfolk Southern, which would create the first true coast-to-coast US railroad. Railroad mergers are approved not by the usual antitrust agencies but by the Surface Transportation Board, which has its own procedure and its own pace.

This week the board set that pace, adopting a schedule for the revised application in which final written arguments are not due until May 28, 2027 — "quite a long ways from now."

For anyone holding the deal, the date is the news. An arbitrage position's return is annualised: the same gap between the market price and the deal price is an excellent return over six months and a poor one over two years. A long regulatory calendar does not make the merger less likely, but it does make waiting for it much less attractive.

Norfolk Southern is the eastern railroad being acquired in the Union Pacific merger, so it is the side an arbitrageur would actually own — buying the target below the agreed price and waiting for approval.

The same Surface Transportation Board schedule applies: a revised merger application, with final briefs due May 28, 2027, and a decision after that. The deal is proceeding rather than stalling, but the money is now tied up for a long time, which is the whole of the observation.

Alphabet appears here as the lead example of what Singh calls the big-tech bond paradox. It has issued a 100-year bond — borrowing money repayable a century from now — at roughly 6%. The proceeds fund data centres.

Why this matters beyond Alphabet: pension funds and insurers have to put money somewhere safe for the long term, and have traditionally bought US government bonds. Alphabet's paper pays two to three percentage points more, and few credit investors believe Alphabet is going bankrupt. So money leaves government bonds for corporate bonds, which pushes government borrowing costs up. The paradox is that the AI investment supposed to grow the economy out of its debt problem is being financed by taking money away from the government's own debt.

Separately, Alphabet is moving all Pixel phone, watch and earbud manufacturing out of China to Vietnam and India from 2027 — a supply-chain hedge against exactly the kind of tariff escalation the rest of the call describes.

Meta is one of the five giant technology companies whose AI spending has outgrown the cash their businesses produce, so they have started borrowing. Meta and Oracle have issued long-dated bonds paying 6-8%.

The number to hold on to is the gap: two to three percentage points more than the US government pays. For an insurer or a pension fund choosing where to park money for decades, that gap is enormous and the perceived extra risk is small. The money moves — and every dollar that moves is a dollar not buying Treasuries, which is part of why long-term interest rates keep rising.

Meta is also among the customers absorbing Nvidia's 15% server price increase, so it is paying more for the equipment it is now borrowing to buy.

Oracle sits in the same group: a company funding enormous AI data-centre construction by issuing long-dated bonds at 6-8% rather than out of profits, and one of the large operators whose server suppliers have already warned about Nvidia's price increases.

Singh's interest is structural rather than in the stock. Each of these bond issues competes directly with the US Treasury for the same pool of long-term savings, and the more the technology companies borrow to build, the higher the government's own borrowing costs go. Oracle is one exhibit in that argument.

Microsoft appears three times in this call without ever being a recommendation. It is one of the hyperscalers told that its Nvidia server bill is rising more than 15%. It is one of the five big technology companies now funding AI construction with borrowed money rather than cash flow. And it is the customer behind Nebius's $17.4 billion cloud contract, whose New Jersey data-centre approval cleared a major hurdle this week.

Taken together these are the load-bearing facts of the call's central argument — that AI spending is now large enough to move the government bond market — rather than a view on the shares.

Amazon is another of the five hyperscalers absorbing Nvidia's price increase and funding data centres with debt. Two other items attached to it this week: Morgan Stanley sketched a path to $500 a share by 2027 on the possibility that AWS, its cloud division, eventually reaches $1 trillion of annual revenue over eight to ten years; and Amazon committed more than $2 billion to Latin American content and live sports for Prime Video.

Neither is presented as a reason to buy. The cloud number is the bull case that justifies the spending, and the content spending is a reminder that the capital budget keeps growing on several fronts at once.

Moderna and Merck reported genuinely good news: their personalised mRNA cancer vaccine, given alongside Merck's Keytruda, reduced the chance of melanoma coming back or spreading, and the trial was stopped early because the results were already convincing. The stock rose 175% in a day.

Singh shorted it — betting the price would fall — not because he disputes the science but because a 175% single-day move is a squeeze, not a valuation. A large part of that move comes from investors who were betting against the stock being forced to buy it back at any price, which produces a spike that unwinds once the forced buying finishes.

That is what happened: the shares fell about 30% from the peak. He describes the exit precisely — the first half was covered "around break even," the rest as it fell, for a high-single-digit percentage gain overall. A modest profit on a fast trade, and an illustration that the discipline is in the covering rather than the shorting.

BioNTech is the German biotech best known for the Pfizer Covid vaccine. It rose 22% during the same week — not on any news of its own, but because Moderna's cancer-vaccine result dragged up everything that sounds similar. Singh's phrase is that it was "up 22% for no reason."

He kept this short when he closed the Moderna one, and thinks it is the better of the two. The reasons are company-specific: senior management has been leaving, and BioNTech has nothing like Moderna's success in cancer vaccines. So the shares rose on a resemblance that does not hold — a sympathy move rather than a re-rating.

The position is being managed rather than held rigidly: half was covered when the stock fell to about $107, and it bounced back into the weekend. "We might add to that short again" — the plan is to sell more into any further strength that the fundamentals do not support.

Walmart fell 8% in a day — its worst session since May 2022 — after reporting the slowest growth in sales at existing US stores in more than six years: 2.6% against the 3.7% expected.

The significance is bigger than the company. Walmart is where American households on tight budgets shop, so its sales are among the cleanest readings available on whether ordinary consumers are still spending. A miss of that size says they are pulling back, which fits Goldman's forecast that inflation-adjusted consumer spending slows to 1-1.5% in the second half of the year, and fits the mortgage-rate squeeze created by a 10-year Treasury yield near 4.7%.

Singh uses it forward-looking too: the question he brings to the coming week's discount retailers is whether Dollar General and Dollar Tree show the same weakness. If they do, the slowdown is confirmed across the whole low-income consumer, not just at one retailer.

Super Micro builds the servers that house Nvidia's AI chips — a good business in a good market. The problem is trust in what the company says.

This week it announced that an independent investigation had found no evidence that current senior management knew about an alleged scheme to divert Nvidia GPUs to China in breach of export controls, blaming two former employees and a contractor. Singh's entire response is four words of sarcasm: "Well, I'm sure that's true. Huh."

The point is one he has made about this name before: being in the right industry does not compensate for not being able to rely on the numbers or the disclosures. When an investigation a company commissions into itself clears its own management, it is not the same as independent proof.