It was just two weeks ago that the AI doom campaign bubbled up.

OpenAI had just released its GPT-6 model. 

Three days later, Jensen Huang said AGI had arrived.

Hedge fund legend Paul Tudor Jones then penned an op-ed in the Wall Street Journal sounding the alarm on AI safety. And two leading frontier model CEOs followed it by publicly calling for a slow down on AI model development.

The media and politicians amplified the case.

As we discussed at the time, you can’t slow down in this race, because whoever gets to human-level intelligence first sets the standard, attracts the talent, decides what gets embedded in governments, in banks, in power grids, in weapons. It’s the difference between AI that serves humanity and AI that serves the Chinese Communist Party.

The President said as much today.

Let’s talk about the meeting he convened in the White House this afternoon. 

Fair to say this was short notice, and yet more than thirty leaders were there, from every major company in the AI stack, representing roughly $26 trillion in public market value, nearly $30 trillion adding in private market valuations (OpenAI and Anthropic).

So, this is a room full of the people controlling companies with a combined value nearly the size of the entire U.S. annual economic output. It’s the companies building the technology, buying the infrastructure, financing it, and bringing it to customers — all sitting at the same table.

The result: Self-regulation. The group signed a voluntary safety accord covering internal controls, independent audits and board oversight. Trump called it “morally binding.”

No slow down! 

Meanwhile, Xi just left a two day visit to Washington with these formal remarks: “Both China and the United States are leading nations in artificial intelligence.” Competition “should be a race of catching up with one another, not a wrestle in which one either wins or loses.”

For the reasons we discussed above, Trump sees it as “whoever wins super intelligence, wins.” And he said just that today.

As we’ve discussed along the way, the more abundant intelligence becomes, the more valuable the scarce inputs become.

The U.S. is keeping its foot on the accelerator. That leaves physics as the only speed limit. That’s power, memory and equipment (physical inputs) that determine how fast this goes.

For that reason, you want to own scarcity in a world of abundance.

The 10-year Treasury yield traded to another two decade high today.

Last week we looked at the 5% level in the 10s, which has been kryptonite for financial stability in the recent past. But that recent past didn’t come with hot growth. It didn’t come with a productivity boom. And it didn’t come with a historic investment to build out the infrastructure for the fourth industrial revolution.

Keep in mind, the U.S. is now borrowing at 5.2%, but is growing (based on the Atlanta Fed model) above 8% (nominal). When an economy grows faster than its government’s borrowing rate, the debt burden shrinks relative to the economy.

And the move in yields is less about how much the government is borrowing, and more about who else wants to borrow. AI data center builders are competing with the Treasury for money.

Today, the Treasury Secretary hired a man who has been making the same argument.

David Zervos, Jefferies’ chief market strategist for nearly two decades, is joining Bessent as a counselor at Treasury. Just earlier this month, Zervos described the move in yields as “a large capital call driving real rates higher on expectations of very significant productivity growth. It’s about as positive a story that you could associate with higher yields as any story possible.”

In August, on the Treasury’s bond buybacks, Zervos said: “I don’t see how you could fight this when the firepower and the cards are all sitting in the Treasury Department.” That statement comes from someone who now sits inside the Treasury Department, with the firepower.

On a related note, we get August PCE on Wednesday (the Fed’s favored inflation gauge). Warsh already told us the number at his press conference: 3.6%.

The bigger inflation report is in Europe.

The euro zone reports September inflation on Friday, which is expected to jump to 3.7% from 3.2%. That’s the energy shock at work. And Europe’s energy premium ticked back up today, after Trump rejected Iran’s Hormuz proposal over the weekend. This all puts more pressure on the ECB to hike again when they meet next month (which would be a third time since June).

Lastly, we talked last week about Trump’s success in fortifying the Western Hemisphere. On Sunday, Brazil votes. It’s the largest economy in South America, and one of the three big countries in the hemisphere not yet aligned with Washington.

The 10-year yield closed at 5.11% today. The highest since 2007.

A lot was made of the move, 14 basis points on the day, the biggest since April of last year.

Here’s what that looks like against every trading day since 1999.

As you can see, it’s not an extraordinary move in yields.

What about in percent? 

Not extraordinary.

But the level? It’s 5.11%. A nineteen year high.

That said, the Atlanta Fed’s running estimate now has the economy growing 5.1% in the third quarter, in real terms. Add inflation, and nominal growth is running north of 8%.

So, the U.S. borrows at 5 and grows at 8. When an economy grows faster than its government’s borrowing rate, the debt burden shrinks relative to the economy.

When the 10-year was trading above 5% in 2007, nominal growth was under 5. So, same yield but different economy.

This is what Warsh has been describing since he took the chair. The global savings glut is over. It’s been replaced by a global investment surge. Rates are high because capital is being competed for, not because the Fed is choking anything.

And if we look back at the 1997 analogue we talked about last week: Greenspan hiked into strong growth and rising productivity. It didn’t stop the boom. 

Now, let’s compare the above to what’s happening in Europe.

The ECB’s own projection has Europe growing just 0.9% this year.

Add inflation at 3%, and nominal growth is about 4. The German 10-year closed today at 3.57%. Germany borrows at its growth rate.

Italy borrows at 4.5%, with growth near zero, on a debt stock of 135% of GDP. Its nominal growth is 3, maybe 3.5. Its borrowing cost is a full point above it.

The U.S. has three points of room. Italy has a point of deficit. That’s a debt burden shrinking relative to the economy, versus a debt burden growing.

There’s a study on this that looks at sixteen countries over a 145-year period (here). For most of those 145 years, governments borrowed below the rate their economies grew. That’s what makes sovereign debt sustainable.

In the case of Italy, through the 1980s and early 1990s it borrowed above its growth rate. Its debt went from roughly 55% of GDP to roughly 100%. It ended with the lira forced out of the European exchange rate system.

Italy is back in that configuration. 

Yesterday, we talked about Xi’s upcoming visit, and the year Trump has spent pulling countries away from China’s influence and back toward the United States.

Today, Trump stood in front of the UN General Assembly and laid the whole case out.

A year ago, his UN address started with an escalator that stopped and a teleprompter that failed.

This one started with an accounting of the past year.

Maduro is in U.S. custody. The Venezuelan agreement covers 65 billion barrels of reserves. The Greenland security agreement was signed today, with two more American bases going in. NATO is committed to 5% of GDP by 2035. And he named the instrument behind the wars he says he has settled: “the use of tariffs and the threat of tariffs.”

Then he rejected the globalist agenda. 

On the shipping levy the International Maritime Organization tried to impose: “There is no global government and while I’m president there will be no global taxes.”

On AI, the United States “totally rejects any attempt to construct a globalist scheme to control” it.

And on the UN itself, he took credit for cutting its budget and its staff.

For perspective, the takeaway of last year’s UN meeting was a world that was wiring up two systems. One that attracts capital around abundance, around energy and industry and innovation. And the other that demands compliance, controlling access to money and markets.

A year later, and the UK has scrapped the digital ID program. Europe, though, is still building the digital euro pilot (central bank digital currency).

Importantly, the UK didn’t reverse because Trump asked. It reversed because the policy consequences resulted in political change, and political change resulted in policy change. 

That dynamic is slowly working through the rest of Europe, with the energy supply shock acting as the catalyst. 

Xi arrives tomorrow (meetings through Thursday). 

China got almost no direct attention in today’s speech beyond the AI race. As we discussed yesterday, keeping the China talks stable buys the Trump administration the time to finish everything around them.

  

With Xi due to arrive in D.C. this week, let’s revisit the May meeting and what has happened between then and now.

Remember, the May meeting was supposed to take place in April. It was pushed “5-6 weeks” on the condition that China help unblock Hormuz. That didn’t happen.

Still, Trump went to Beijing, and with a large delegation that included Jensen Huang and Elon Musk. He posted that his very first request would be for Xi to “open up” China.

“Opening up” China has been the false promise of the past few decades. 

It hasn’t happened. The result has been a Chinese economy that grew 50x in 35 years while ours grew 5x. It was a transfer of wealth and power, built on a cheap currency, an export monopoly, and the recycling of American consumer dollars into Treasuries so Americans could buy more cheap stuff.

“Opening up” China would require Xi to abandon the economic model that made China a superpower. Not going to happen.  

So, what came out of Beijing?

On paper: a Chinese commitment to buy $17 billion a year of American agriculture through 2028, an order for 200 Boeing jets, a new “Board of Trade” for non-sensitive goods, and a promise to “address” American concerns on rare earths.

The White House announced the $17 billion. Beijing’s commerce ministry confirmed only that it would “expand agricultural trade broadly.” No number.

Not on paper was the H200 chips, Taiwan, Iran, Hormuz. And on the November trade truce, Bessent said afterward that the U.S. was “not in a rush to extend” it.

That was May. And Trump invited Xi to Washington for September, before the truce expires on November 10.

So, Xi is due. What has happened between the two meetings?

As Bessent reported this morning in a CNBC interview, China’s deliverables from the last meeting “have not been completely fulfilled.” 

And the leverage has shifted.

This time Xi comes to Trump, on American soil. 

The U.S. has increased its control over global oil supply.

The dollar has been reaffirmed as the center of the global financial system. The new stablecoin framework creates a new source of demand for dollar assets and Treasury securities.

And the “isolate China” endgame that we’ve talked about since early last year, now looks to be in-motion.

The Trump administration launched Economic Outcast in late August, a program that has removed banks from the dollar system for handling Iranian money, but for now with an obvious name unspoken.

The operation itself is actually called Economic Outcast. 

And then, earlier this month, at a G20 hosted by the United States, nineteen members put their names to a document about repairing global trade imbalances, and China sat alone on the other side of it.

The “isolate China” pieces seem to be moving into place.

That said, in dealing with China, the plan thus far has been preserving stability. Bessent said as much in a CNBC interview this morning.

Maintaining stability buys time to continue working toward the endgame.

What’s the sequencing to the endgame?

You don’t isolate the second largest economy in the world by yourself. You do it by realigning everyone else first, then China is left standing alone.

As we discussed in February of last year, two weeks into Trump 2.0, the tariffs were never about revenue. They were about realignment – away from China’s influence, and back toward the U.S.

The U.S. has the one thing every export economy on earth needs, the American consumer. Country after country came back to the table on those terms. China didn’t, not surprisingly. China retaliated, built workarounds, and stalled.

The Western Hemisphere: Trump has taken back the shipping lanes (Panama and Greenland), removed China’s operational partners (Venezuela and Cuba), and realigned much of the rest of the hemisphere with economic leverage (tariffs). This was mostly done inside a year.

Then Middle East: He’s removed a hostile regime, and taken its oil out of the hands of the people who used it to fund chaos. And Iran’s oil goes to China through a strait which is now administered by the U.S. Navy. Control the chokepoint and you control who gets energy and at what price. That’s leverage over China.

And Europe: European leaders weren’t interested in realigning with a new American administration that was reversing the climate and social agenda. But realignment can come when economic, defense and energy security is leveraged against them. With that, it’s now showing up for European voters in the gas bill. It’s showing up in the bond market. And it’s now showing up in the ballot box (political change).

Then China: Isolation. The model is on display against Iran right now: every bank, shipper, insurer and buyer that finances the target is told to choose between China and having access to the dollar system.

Where are we on the Western Hemisphere map?

The day after a rate hike, the 10-year yield traded down seven basis points, to 4.93% (more curve flattening). Stocks rallied. Technology led.

We’ve talked often about the parallels between the current environment and the late 90s boom.

And this rate hike yesterday gives us another point of comparison.

As we know, a technology revolution was underway in the late 90s, with the rapid adoption of the internet. Productivity was high. Growth was hot. Inflation was tame (relatively low). And the Fed juiced it with rate cuts, starting in 1995. 

The stock market boomed in 95, up 34%.  And up another 20% in 1996. The economy boomed, growing 3.8% in 1996, up from 2.7% in 1995.

Then, in March of 1997, Alan Greenspan raised the Fed Funds rate a quarter point, to 5.5%.

The reason the committee gave, in its own words: “persisting strength in demand, which is progressively increasing the risk of inflation imbalances developing in the economy that would eventually undermine the long expansion.”

They hiked because the economy was strong and they were worried about what that strength might eventually do.

Compare that to what Warsh said yesterday.

He was asked how a quarter point helps when it can’t reopen the Strait of Hormuz (i.e. a quarter point hike can’t fix the oil supply disruption). Warsh said, the Fed can’t affect any individual price, but it can “ensure that any change in relative prices don’t broaden out, don’t have second and third order effects.” This is another, Greenspan-like hike based on what prices might eventually do. 

Same argument, twenty-nine years apart.

So, what came next after the hike in 1997?

For eighteen months the committee pushed Greenspan to hike again as unemployment fell toward 4% and growth boomed. The Fed’s models said that had to produce inflation. He refused, and told them why: information technology was raising productivity, raising the economy’s potential growth rate, and unemployment could fall further than anyone thought without prices rising.

He was right. Inflation actually fell. 

Still, the Fed held rates steady at 5.5%. But not just steady, at real rates (Fed Funds rate minus inflation) between 3% and nearly 5% — very restrictive policy.

None of it stopped the boom.

The economy averaged 4.6% quarterly annualized growth through the end of the decade. Stocks put up five consecutive double-digit years, averaging 26%. And inflation moved lower (not higher). 

It turns out, the productivity gains from the tech revolution were more powerful than the Fed’s restrictive policy. 

Through those years, American productivity growth averaged about 2.7%.

Fast forward to today: since the release of ChatGPT in late 2022, it has averaged 2.5%.

The Fed Funds is now at 3.75% to 4%, against 5.5% then. And the real rate, after yesterday’s hike, is positive, but at just 0.18%, against roughly 3%+ then.

So, we have a productivity rate that rhymes with the late 90s.

A policy rate nowhere near as restrictive as the one that the 90s boom absorbed without breaking stride.

And the AI-driven tech revolution is bigger than the internet. 

The Fed raised rates today.

So, the tightening that the bond market already delivered, as we discussed, didn’t deter them.

And the hike was a unanimous 12-0. Warsh even led the statement with that signal of unanimity (after July’s 9-3 hold). 

How did the bond market respond?

The 2-year Treasury yield opened at 4.65% and closed at 4.73%. The 30-year closed a touch lower, around 5.36%.

That’s short end up, long end down.

That’s a flattening yield curve. That’s the Fed putting pressure on future growth.

Let’s talk about Warsh’s press conference.

The first question in the Q&A was well placed. A journalist in the room pointed out that a quarter point rate hike does not reopen the Strait of Hormuz. That addresses directly the point we made going in. You don’t hike rates into a supply shock. In this case, it does nothing to bring down the price of oil. 

Warsh agreed. And this is where the rate hike was framed.

He said the Fed cannot affect any individual price, but what it can do is “ensure that any change in relative prices don’t broaden out, don’t have second and third order effects.”

So he conceded the tool doesn’t fix the problem, and hiked anyway to prevent the problem from spreading.

What would make it spread? A strong economy.

On that note, Warsh called the economy strong and strengthening. 

So, this was a hike to slow growth. This, from a Fed that Warsh said has “an attitude of optimism.” Presumably, that means the Fed is confident that the economy can absorb a quarter point hike.  

That said, the old Fed takes a growing economy out back and shoots it.

The Warsh “mental model,” as we’ve been told, is that you don’t push demand down to respond to a supply shock. And you don’t hike rates to slow an economy running hot on productivity gains. Hot productivity is the cure for elevated inflation. 

The Fed decides tomorrow.

The 10-year Treasury yield closed at exactly 5% today. As we discussed yesterday, that’s the level that stopped the Fed in October of 2023, when Powell looked at a 5% long bond and told the New York Economic Club that financial conditions had “tightened significantly.”

The bond market had done the work. He stopped the tightening cycle, and within two months the 10-year was under 4%.

Kevin Warsh made the same observation himself in July, in his own press conference: “we haven’t done much in 42 days. The markets have done quite a bit.” He’s talking about the rise in longer term bond yields. 

That said, the market says he hikes tomorrow. The bond market has given him reason not to.

Now let’s talk about something bigger than the Fed meeting — the AI doom campaign.

Remember, it was just twelve days ago, OpenAI released its GPT-6 model.

Three days later, the man running the most valuable, and most important company in the world, said this about it …

AGI is “Artificial General Intelligence.” GPT-6 itself defines it as “AI that can learn, reason, and solve problems across a broad range of tasks at roughly human level or better.“

And this is the stage where the AI begins to autonomously improve itself, which creates a path to “superintelligence” — AI far beyond human capabilities.

Now, after this marinated a few days, one of the great hedge fund traders of our time, Paul Tudor Jones penned an op-ed in the Wall Street Journal sounding the alarm on AI safety. 

Was he talking his book? Very likely. 

By Saturday, the two leading frontier model CEOs were calling for a slow down on AI model development.

The media and politicians have since amplified the case.

But keep in mind, AGI has been declared by the man who sells the chips that AGI will need in unlimited supply. Meanwhile, the margin of performance between the frontier models, as measured by the third party Artificial Analysis Intelligence Index, is extremely tight.

Notice the cluster of the top models circled in the graphic below. This has GPT-6 leading, but at a level well below the performance reported from OpenAI’s internal tests.

So, the American labs continue to lead: OpenAI, Anthropic, Google. But the Chinese labs, DeepSeek, Alibaba, Kimi, and others, have continued to compress the gap.

Until proof of AGI, the frontier is still converging.

And the consequences of who ends up on top are massive.

As we’ve discussed over the past eighteen months, whoever gets to human-level intelligence first sets the standard. They attract the talent. They decide what gets embedded in governments, in banks, in power grids, in weapons. It’s the difference between AI that serves humanity and AI that serves the Chinese Communist Party.

You cannot slow down in that race.

We get the Fed decision this week.

The ECB raised rates last week, into an energy supply shock-driven inflation number, and European yields have since risen (benchmark German 10-year yields). The euro has fallen. And risk premiums in Europe widened on the move.

Why? Because increasing borrowing costs will do nothing to lower the cost of energy in Europe.

Still, the ECB is sending signals to markets that more rate hikes are coming. 

As for the Fed, the market is now pricing in a hike this week, and as many as two more by March.

As we’ve discussed, the new Fed, under Warsh, is not about giving the market signals. As Warsh has said, he wants the markets to “play the ball, not the referee.”  

That said, the market knows the old Fed playbook. And Warsh sits at the table with old Fed loyalists. So the market reads three dissenters at the last meeting (who favored a July hike) as signal.

And with that, the U.S. 10-year Treasury yield has been marching higher and traded above 5% today, for the first time since 2023.

Let’s talk about that 2023 episode. 

It was October of 2023 when the 10-year yield tested 5%. The Fed, under Powell, was holding real rates high, and they surprised markets by sending a more hawkish signal in their September meeting — even as inflation was falling.

In response, the 10-year yield started a 64 basis point climb toward the 5% level.

Stocks traded down 5% as rates headed toward 5%.

Mortgage rates traded to 23-year highs.

Investment grade corporate bonds were at one-year lows. 

And at the time, oil had made a 40% surge in the third quarter, up to $95, driven by war in the Middle East. And with that rate and energy market dynamic, in October the Bank of Japan was forced to intervene in the currency markets to defend the value of the yen. 

Sound familiar? 

It turns out the 5% level in the benchmark bond yield was financial stability kryptonite — enough to flip the switch at the Fed.

Just weeks after setting the bond market repricing into motion, Jerome Powell delivered a prepared speech to the New York Economic Club and said that financial conditions had “tightened significantly” since their September meeting (i.e. long-term bond yields).

Translation: If the Fed needed to do more, the bond market had done it for them (and maybe too much). It signaled the end of the tightening cycle.

The reprieve in the bond market was immediate. And within two months 10-year yields were trading under 4%.  

And keep in mind, inflation (both headline and core) was in the mid 3% area (similar levels to now). And Powell backed off the tightening policy path as inflation expectations were, at that time, much higher than current levels. 

 

So, three years ago the bond market did the tightening and the Fed took the excuse.

This week the market expects a hike. But the bond market has again done the tightening, and has given Warsh an excuse to hold. 

Moreover, Warsh doesn’t think you raise rates into a supply shock. He has said that pushing down demand (through higher policy rates) until it meets supply is “not my mental model.”

Add to all of this, Warsh said in his Senate confirmation hearing that he prefers the Dallas Fed’s “trimmed averages” to measure inflation. That number was 2.3% in July. That would put real rates around 150 basis points, which is about the level of October 2023 (policy rates – trimmed mean PCE).     

The European Central Bank raised interest rates this morning, to 2.50%.

The decision was unanimous. It was expected.

Let’s go through what Christine Lagarde said in the press conference, because she spent the time making the case against her own decision.

Asked to explain the framework, she described the situation in Europe as “predominantly a supply shock.” Exactly. Europe has an energy problem, not a demand problem.

Then the monetary policy statement said this: wages do not show a material response to the energy shock at this stage. Compensation per employee grew 3.3% in the second quarter, down from 3.5% in the first. Unit labor costs slowed to 2.6% from 3.5%.

And on food prices, steady at 1.2%. 

So, Lagarde says it’s a supply shock. No wage pressures. Falling unit labor costs. No second round effects visible anywhere in the data.

And yet the ECB delivered a unanimous rate hike.

Now remember what Scott Bessent said from Asheville two weeks ago: “traditionally, you don’t raise into a supply shock unless you see second or third order effects.”

So, the ECB ignored that, and ignored its own history of policy mistakes under similar conditions.

Meanwhile, a reporter in the room pointed out that the ECB’s own June adverse scenario for energy prices assumed European gas at €60.

They moved the goalposts on the macroeconomic projections report they released today. They now see €60 as the baseline scenario for European gas. The adverse scenario assumed €77 and the severe scenario is at €130.

Gas is now at €82 — already above the adverse scenario.

That’s growth destructive, and the scenario analysis does not factor in tightening by the ECB (which they did today).

Now, what’s also interesting in this report, they didn’t model a problem in the sovereign debt market.

They talked about risk sensitivity to these scenarios in corporate bond spreads, bank equity, bank bond spreads, lending spreads, but nothing on sovereign debt vulnerability.

So, if the gas price shock becomes severe, the ECB seems to want the market to believe that their standing threat to backstop the fiscally fragile sovereign bond markets in Europe will be sufficient (such, that it’s not even worth discussing in the report).

But as we’ve discussed for much of the past year, the ECB backstop only works when major global central banks are coordinating (namely the Fed is behind you). And the Warsh-led Fed is unlikely to be there, unless political conditions are met (alignment).