Devin Merdin @DevinMerdin

US forces hit targets near the Strait of Hormuz overnight. Iran's Revolutionary Guard fired at multiple US aircraft. The US simultaneously told markets it is making progress on a peace deal. Oil barely moved. This is the market we are in. Guns firing in the Strait. Diplomats talking. Oil at $99. The market trying to decide which one to believe. For twelve weeks it has oscillated between the two. Every peace signal: oil down 5 to 15%. Every escalation: oil back up. Brent was $60 before the war. It is $99 today. The market has decided a 65% spike is almost normal. Meanwhile the economy is sending its own signals this morning. Salesforce reports today. Not Nvidia. Salesforce. The gap between AI infrastructure demand and AI application demand is one of the most important questions in the trade right now. Today's call begins to answer it. April new home sales also drop today. 30-year mortgage rates tracking a 30-year Treasury yield at a 19-year high. Housing affordability at worst levels in decades. Warsh. First full week. No public statement yet. HSBC called Treasuries a danger zone last week. Three data points. One morning. All asking the same question. How much of the resilience is genuine strength and how much is the gap between when a shock happens and when it shows up in the data? What history shows. The shock never announces itself. But Salesforce's language about enterprise spending intentions for the rest of the year will add one more piece to a picture that is becoming harder to ignore. Listen to the call. Not the EPS.

Markets close Friday at record highs. Eight consecutive weekly gains. This week asks three questions the rally has not yet had to answer. One. What does a genuinely hawkish Fed chair sound like in his own voice? Warsh took his oath at the White House on Friday. First since Greenspan in 1987. Trump said he wants him to be totally independent. You do not say that unless the independence is in question. This is his first full week with markets open. 30-year yields at a 19-year high. December hike probability above 50%. Every word gets parsed. Two. Does the consumer resilience run deeper than Walmart suggested? Costco and Dollar Tree both report this week. Costco tells us if trade-down is reaching higher income households. Dollar Tree tells us how stressed the bottom actually is. Walmart said uneven consumer strength among lower-income cohorts last week. That sentence needs a second data point. Three. Is the SpaceX IPO the signal or the noise? June 12 listing. $1.75 trillion target valuation. Largest IPO in history if it prices there. Multiple strategists already drawing dot-com parallels. June is historically the worst month for the S&P 500 in midterm election years. Minus 2.1% average. The market enters June after eight straight weeks up. The Warsh voice. The consumer data. The IPO signal. Against a backdrop of an unsigned Iran deal and a yield curve still steepening. What history shows. The eighth week of a rally is rarely where the thesis breaks. But it is almost always where the first cracks become visible to those paying attention. The investors who saw 2000 coming did not see it in the headlines. They saw it in the IPO pipeline and the rate environment. Both are in this week’s calendar.

The most interesting opportunities in markets are never in the assets everyone is watching. Here is what this environment is quietly creating. Energy infrastructure is up 25% year to date against a broader market up 2%. Not because of speculation. Because pipelines and terminals get paid whether oil trades at $60 or $100. Fee-based businesses with inflation-protected contracts do not need the conflict to be resolved. They just need the world to keep consuming energy. Short duration fixed income is paying real rates for the first time in fifteen years. 5% on short-term Treasuries was theoretical in 2021. With the FOMC confirming a rate hike bias this week, the case for patience at the short end has not been this strong in two decades. Energy supermajors with low-cost Permian production generate extraordinary free cash flow at current prices. Each $10 Brent increase adds billions to annual operating cash flow. They win if oil falls on a deal. They win more if it does not. None of these require the Iran situation to resolve in any particular direction. That is the point. The best positions in complex environments win across multiple scenarios. The worst ones require one specific outcome. What history shows. In every period of sustained geopolitical commodity stress, the investors who captured the most durable returns were not the ones who called the resolution date. They were the ones who identified which assets were structurally advantaged by the environment itself. The environment is complex. Complex environments reward those who read the map rather than watch the weather.

Week in review. Five days. More macro significance than most quarters. What we know now that we did not know Monday morning. The Fed changed regime. Warsh confirmed 54-45. FOMC minutes: four dissents, most since 1992. A majority now sees rate hikes as appropriate. December hike probability crossed 50%. It was below 10% in April. The 20-year Treasury auctioned at 5.047%, its first since Moody's stripped the US AAA rating. It tailed. Japan's 40-year yield breached 4% for the first time in thirty years. The UK Gilt at 5.71%. Germany's gas storage below 30%. Nvidia: $81.6B revenue. $91B Q2 guidance. The AI cycle is real. Intuit cut 3,000 jobs the same week. The displacement cycle is also real. Walmart beat estimates and fell 2.87%. Uneven consumer strength among lower-income cohorts. That is not a retail observation. It is a country observation. Oil priced a deal four times. No signed agreement. Tonight Iran's supreme leader ordered enriched uranium kept within borders. Trump gave Tehran more time. The S&P 500 is on track for its eighth consecutive weekly gain. The equity market looked at all of the above and went up. Markets can be right about the near term and wrong about the trajectory at the same time. What history shows. Every time equity markets have sustained a multi-week rally through a simultaneous bond repricing, a credit downgrade, a commodity shock, and a central bank regime change, the reconciliation with structural reality has not been gentle. The timing is always unknown. The direction has never been in question.

The oil market has priced an Iran deal four times since April. April 8: Brent fell 16% on ceasefire news. Oil recovered to $110. May 6: Oil plunged 15% on MOU reports. Trump said too soon to sign. Oil recovered to $102. May 20: Oil fell 5.6% on final stages talk. Tonight: Oil selling toward $96 on a TV station draft. No White House statement. No Iranian signature. Four times. Same trade. Same outcome. The Strait is still closed. Inventories fell 4 million barrels a day in March and April. Germany's gas storage is below 30%. A signed deal reopens the Strait. It does not refill the inventory. What history shows. Every major geopolitical supply shock prices resolution before it arrives. Repeatedly. The structural floor is never found at the hope price. It is found at the supply-reality price. The supply reality has not changed tonight.

Nothing good happens above 4.5%" is the line that should be on every portfolio manager's wall right now. But the move this week was not just the US 10-year crossing that level. It was the synchronisation. UK. Canada. Germany. Japan. All making new cycle highs simultaneously. That is not a US rates story. That is a global repricing of what governments should pay to borrow. Japan's 40-year yield crossed 4% for the first time in three decades this week. Japan holds $3.7 trillion in foreign assets accumulated during thirty years of near-zero rates. As those yields rise, that capital starts coming home. History shows that when the world's largest creditor nation begins repatriating at scale, the left tail in every duration-sensitive asset class gets significantly fatter. The left tail is not awakening. It has been awake all week.

Tax cuts for exporters do not fix a currency crisis. They fix a competitiveness problem. Turkey has both but only one is being treated. The Lira has lost over 80% of its value against the dollar in the last decade. That is not a tax rate problem. That is a monetary credibility problem. When a central bank is perceived to operate under political pressure rather than price stability mandate, no fiscal package closes that gap. History shows that countries which attempt to substitute fiscal stimulus for monetary credibility during a currency crisis delay the adjustment and increase its eventual cost. Turkey is not unique in this. Argentina tried it. Egypt tried it. The Lira is writing the same chapter. The uncomfortable question is not what tax rate Turkish exporters should pay. It is whether the institution responsible for the Lira's value is trusted enough to anchor it. Until that question has a different answer, the balloon in that chart keeps rising.

The $80 billion buyback and the 24x dividend increase are real. So is the 10-year Treasury at 4.59%. So is the FOMC majority that now sees rate hikes as appropriate. So is the cost of capital that rose from 13.8% to 16.5% in four quarters. Nvidia is compounding at a scale that has no modern precedent. That is not the question. The question is what multiple the market should pay for that compounding in an environment where money is no longer free. History shows that the best companies in the world can be the wrong investment at the wrong price. In 1999 Cisco was also the future. The pattern that matters is not the earnings pattern. It is the rate environment pattern.

4am deal on. 6am deal off. 1:30pm deal on. This is not a market. It is a news cycle with a futures strip attached. The traders who cannot find conviction are looking at the wrong signal. The FOMC minutes, the 20-year auction, the JGB yields, the inventory data, none of those changed between 4am and 6am. The structural picture has not moved. Only the headlines have. History shows that when a market loses the ability to price fundamentals and starts pricing rumours instead, the eventual return to fundamentals is not gentle.

Every element of this agreement is designed to be market-friendly on announcement. Complete ceasefire. Sanctions lifted. Strait reopened. Clean and comprehensive. Which is exactly why it should be questioned hardest before being treated as confirmed. Agreements this comprehensive do not emerge overnight. Leaked drafts from regional TV stations are not signed agreements.

Oil is selling off on a draft agreement obtained by a regional TV station. Not a signed agreement. Not a White House statement. Not a UN confirmation. A draft. The market just priced peace from a headline. The inventory data priced something very different all week. Watch what oil does in 72 hours when the draft meets the reality.

Point 5 guarantees freedom of navigation in the Strait of Hormuz. That is the only line the oil market cares about tonight. But a draft agreement obtained by a TV station is not a signed agreement. The Strait has been closed for weeks. Germany's gas storage is below 30%. Global oil inventories fell 4 million barrels a day in March and April. A ceasefire reopens the Strait. It does not refill the inventory. Watch what oil does at the open. Then watch what it does in 72 hours.

While everyone watched Nvidia and Walmart this week, the most important bond market event on earth went almost unnoticed. Japan's 40-year government bond yield has breached 4% for the first time in over three decades. Japan was the anchor of global low rates for thirty years. Its institutions accumulated $3.7 trillion in foreign assets, mostly government bonds, because domestic yields were near zero. As Japanese yields rise, those assets come home. Goldman Sachs estimates that every 10 basis points of JGB shock propagates 2 to 3 basis points of pressure onto US Treasury yields. The UK Gilt at 5.71%. The US 20-year at 5.047%. Japan's 40-year at 4.24%. This is not four separate bond market stories. It is one. The thirty-year experiment with near-zero rates in the world's largest creditor nation is ending. Every borrower on earth is going to feel it. Japan's debt to GDP is 260%. The United States is at 100%. What history shows. Every time a major creditor nation repatriates capital at scale, long-duration bonds reprice lower, heavily indebted currencies come under pressure, and hard assets absorb what has nowhere else to go. History does not repeat. But it arrives wearing the same clothes. You do not need me to tell you what to do with that.

Wednesday: oil fell 5% on Iran deal hopes. Today: oil is gushing higher. S&P is slipping. Also this week: Nvidia reported $81.6 billion in revenue. AI infrastructure is booming. Intuit cut 17% of its workforce. 3,000 jobs. The products AI is replacing no longer need the same headcount to run them. Two headlines. 24 hours apart. The same industry. Completely opposite directions. The AI boom is real. So is the displacement it is creating. Somewhere between those two headlines the Fed has to set interest rates. Good luck to them.

This week. Two earnings reports. One story. Nvidia beat. Stock rallied. AI trade confirmed. Walmart beat. Stock fell 2.87%. The reason: uneven consumer strength among lower-income cohorts. Walmart serves 90% of US households. When it uses the word uneven it is not describing a customer segment. It is describing a country splitting in two. Both companies beat estimates. Both stocks fell on the day. The market was not reading the headline. It was reading the sentence nobody highlighted. The K-shaped economy is no longer a theory. It showed up in two earnings calls this week and the market priced it immediately.

Japan spent $70 billion defending the yen. The yen gave it all back. This is what happens when you fight a structural problem with a tactical tool. The Bank of Japan faces the same constraint as every other major central bank right now. The US FOMC minutes confirmed hike language last night. The ECB faces an energy crisis. And the BoJ is trying to hold a currency together with intervention while the yield differential between Japan and the US grows wider every week. You cannot buy your way out of a rate problem. USD/JPY at 159 is not a trading level. It is a policy failure in slow motion.

In 2023 13-14% sounded absurd. In 2026 the FOMC minutes confirm a majority of participants see rate hikes as the likely next move. The 20-year Treasury just auctioned at 5.047%. The US lost its final AAA credit rating last week. Nobody laughed at Volcker either. Until they did. Until they had to stop. The question is not whether Santelli was right. The question is whether the people running the institution have the policy space to do what Volcker did. In 1981 federal debt was 25% of GDP. Today it is 100%.

Some of us have been paying attention. Four of the world's largest bond markets hitting multi-decade highs simultaneously is not a coincidence. It is a coordinated message from the people who lend money to governments. The message: we want more to hold your debt than we did a decade ago. Significantly more. Last night the US government paid 5.047% to borrow for 20 years in its first auction since Moody's stripped its final AAA credit rating. The UK 30-year Gilt is at 5.71%. Germany's gas storage is below 30%. The FOMC minutes confirmed a majority of participants now see rate hikes as the likely next move. The 1970s did not announce themselves either.

Last night two things happened simultaneously. Nvidia guided Q2 at $91 billion. The AI trade is back on. Asia is up 7% this morning. The FOMC minutes confirmed in writing that a majority of participants believe policy firming would become appropriate if inflation persists. Policy firming is the Fed's language for rate hikes. The market chose the Nvidia story overnight. This morning Walmart reports. That number will tell you whether the half of the economy that does not build AI infrastructure got the same memo. Two economies. One morning. Watch both numbers.

Nvidia just reported. $81.61 billion revenue. Beat. $1.87 EPS. Beat. $91 billion Q2 guidance. Crushed the whisper. $119 billion supply commitments. Dividend increased 24 times over. $80 billion buyback. Zero China revenue assumed. The AI trade is real. These numbers prove it. Now the question nobody is asking. This reported into a tailed 20-year Treasury auction. FOMC minutes revealing a committee already losing the inflation argument. Rate hike probability above 50%. Nvidia proved the AI demand cycle is intact. The bond market is proving the financing cycle is tightening. Both things are true. Only one of them decides the multiple.