AI Summary. The author argues markets are volatile because interest rates are moving higher than investors thought possible while AI demand comes in lower, with US growth sustained by competition for capital between AI investment and government debt. With inflation stuck above 3%, bond markets price a Federal Reserve rate of 4.75%, versus closer to 6% in a typical tightening cycle.

Date Posted:

Podolsky warns the inflationary impulse is not isolated to the US. Core PCE inflation is stuck just above 3%, and Japanese inflation is rising, though at a lower 2.1%. He argues that markets, which price the FFR rising from 4% to 4.75%, are underestimating how far the Fed may need to go: a typical tightening cycle with inflation this high would take the rate closer to 6%, ~200bp above today’s level.

Why are markets volatile as rates rise and AI demand disappoints?

Core argument: U.S. inflation is stuck above 3%, and the author argues the Fed is well behind the curve, with what first looked like the removal of insurance cuts now amounting to a full tightening cycle.

Markets are moving in a skittish, volatile manner because key anchor points are shifting, and trillions of dollars of capital are being forced to adjust. Specifically, interest rates are moving above what investors thought was possible, and AI demand is moving below what investors thought was possible. Here are the salient points: U.S. growth is not slowing (yet) because what I have referred to as the “cage fight for capital” (AI capex and U.S. government debt). This is producing fast, concentrated growth and slow, broad-based inflation. The Fed is now well behind the curve. What initially seemed like the removal of “insurance cuts” now looks like a full-fledged tightening. Inflation is stuck above 3 percent. Fed Governor Waller floated raising rates 50 bps at once, which is appropriate in my view. This inflation impulse is not isolated. The same thing is happening in Japan (though not so much in other countries). This means bond yields need to go higher. The consensus as priced into bond markets is that inflation will settle and fall. The Fed rate is currently at 4% and priced to rise to 4.75%, but a typical tightening cycle with inflation this high would move it closer to 6%. Similarly, in Japan, rates are supposed to top out at 2%, whereas before Japan hit their deflationary depression, a policy rate of 4% was more normal.

Takeaways by Macro Roundup® AI

  1. U.S. inflation is stuck above 3%, and the author argues the Fed is well behind the curve, with what first looked like the removal of insurance cuts now amounting to a full tightening cycle.
  2. The Fed’s policy rate stands at 4% and is priced to rise to 4.75%, but the author contends a typical tightening cycle with inflation this high would push it closer to 6%, implying bond yields must go higher.
  3. The author attributes fast, concentrated U.S. growth and slow, broad-based inflation to a “cage fight for capital” between AI capital expenditure and U.S. government debt, with a similar inflation impulse appearing in Japan.

AI Summary. The U.S. accounts for about 70% of developed-market equity capitalization but roughly 50% of GDP, and the authors attribute 64% of that gap to U.S. listed companies generating more earnings per dollar of GDP. A further 18% reflects the higher-multiple industry mix of U.S. indexes, leaving 17% from same-industry valuation differences.

Date Posted:

The US has ~70% of developed-market equity value and about half of developed market GDP. D.E. Shaw finds higher listed earnings per dollar of GDP explain 64% of the market-cap-to-GDP gap and suggests that listed firms earn a larger share of US corporate profits.

Why does the US equity share exceed its GDP share?

Core argument: The U.S. accounts for 70% of developed-market equity capitalization but only about 50% of GDP, a gap the authors attribute mainly to structural and compositional factors rather than U.S. stocks being pricier than global industry peers.

In developed markets, the U.S. represents 70% of equity capitalization but only about 50% of GDP. Our findings suggest that the gap stems mainly from structural and compositional factors, not from U.S. stocks being more expensive than their global industry peers. US publicly listed companies generate more earnings per dollar of GDP than developed market peers, which explains nearly two-thirds (64%) of the gap. Our analysis suggest that this likely reflects a larger share of the U.S.’s economic activity being driven by publicly listed companies, not corporate profits representing a greater proportion of GDP. The remaining 36% is attributable to U.S. equities trading at a higher aggregate multiple. However, half of that differential (18% of the total) simply reflects industry composition: U.S. indexes comprise higher-multiple industries in greater proportion. In other words, less than one-fifth (the remaining 17%) of the gap reflects disparity in same-industry valuations.

Takeaways by Macro Roundup® AI

  1. The U.S. accounts for 70% of developed-market equity capitalization but only about 50% of GDP, a gap the authors attribute mainly to structural and compositional factors rather than U.S. stocks being pricier than global industry peers.
  2. U.S. listed companies generate more earnings per dollar of GDP than developed-market peers, explaining 64% of the gap, which the analysis suggests reflects a larger share of U.S. economic activity run by publicly listed firms rather than higher corporate profits as a share of GDP.
  3. Higher aggregate multiples account for the remaining 36% of the gap, half of which (18% of the total) reflects U.S. indexes' greater weighting toward higher-multiple industries, leaving 17% attributable to same-industry valuation differences.

AI Summary. The Economist estimates artificial intelligence needs roughly $3tn in revenue by 2030 for the planned data-centre buildout to earn a conventional return, a figure other analysts have also reached. That equals almost a tenth of the roughly $35tn in total worker compensation across rich OECD countries.

Date Posted:
Is Database:
Database

The Economist calculates that the US data center buildout needs ~$3T in revenue by 2030 to offer a competitive ROI. That amounts to ~18% of US annual labor compensation inclusive of health care or ~10% of labor compensation across the OECD.

How much revenue does the AI buildout need to pay off?

Core argument: The Economist calculates that AI needs roughly $3trn in revenue by 2030 for the planned data-centre buildout to deliver a conventional rate of return.

The Economist calculates that AI needs around $3trn in revenue by 2030 for the planned data-centre buildout to offer a conventional rate of return. Other analysts, including Stijn Van Nieuwerburgh of Columbia Business School and a duo at the Stanford Institute for Economic Policy Research, have reached similar figures. To put that number into perspective, all compensation for all workers—wages, plus perks like health insurance—adds up to around $16.5trn in America, or $35trn across the OECD, a club mostly of rich countries. So AI would need to lay claim to almost a tenth of today’s rich-world wages for the data-centre boom to pencil out. Of course, that could happen in a more benign fashion if the entire economy grows swiftly—and AI evangelists do also expect dramatic growth from here.

Takeaways by Macro Roundup® AI

  1. The Economist calculates that AI needs roughly $3trn in revenue by 2030 for the planned data-centre buildout to deliver a conventional rate of return.
  2. The $3trn figure amounts to almost a tenth of the roughly $35trn in total worker compensation across the OECD, which compares with about $16.5trn in America alone.
  3. Stijn Van Nieuwerburgh of Columbia Business School and a pair of researchers at the Stanford Institute for Economic Policy Research have reached similar revenue figures, while AI evangelists expect dramatic economic growth from here.

AI Summary. Inflation-adjusted stock market recoveries have typically taken longer than nominal ones: after the 1929 crash, nominal share prices took about 22 years to regain their prior peak from the 1932 trough versus nearly 27 years in real terms, and after 2008, four years versus five and a half.

Date Posted:

Using Robert Shiller’s monthly-average data, the inflation-adjusted S&P 500 price index, excluding dividends, took ~12 years from its February 2003 trough to regain its August 2000 peak in November 2014. Following the financial crisis, it took four years and eight months from the March 2009 trough to regain its October 2007 peak in November 2013.

How long do stock markets take to recover after crashes?

Core argument: After the 1929 crash, nominal U.S. share prices took roughly 22 years to regain their prior peak from the June 1932 trough, while in real terms recovery took close to 27 years.

The duration of recoveries in real share prices in the 20th and 21st centuries has typically been longer than the duration of recoveries in nominal share prices. For instance, for the 1929 crash it took around 22 years before nominal share prices reached their previous peak after bottoming out at the trough in June 1932. In real terms, it took close to 27 years. When there is inflation and prices in society rise, the stock market takes even longer to recoup its purchasing power than it takes simply to recover in nominal terms.

Takeaways by Macro Roundup® AI

  1. After the 1929 crash, nominal U.S. share prices took roughly 22 years to regain their prior peak from the June 1932 trough, while in real terms recovery took close to 27 years.
  2. Following the 2008 global financial crisis, the S&P 500 recovered in four years from its March 2009 trough in nominal terms but required five and a half years in real terms.
  3. Recoveries in real share prices in the 20th and 21st centuries have typically taken longer than nominal recoveries, since inflation means the stock market needs more time to recoup its purchasing power.

AI Summary. OpenAI's quarterly revenue grew roughly 50% to $10bn, while its adjusted operating loss, which excludes stock-based compensation, rose from $10.1bn to $10.9bn and its adjusted operating margin improved from -150% to -109%. Anthropic reported quarterly revenue of $17bn and a smaller operating loss of $567m that includes stock-based compensation.

Date Posted:

OpenAI’s quarterly revenue reached ~$10B in Q3, up 50% from Q2. While its margins improved q/q from -150% to -109%, its adjusted operating loss still increased to $10.9B

Are rising AI revenues shrinking losses at leading model developers?

Core argument: OpenAI's revenue rose roughly 50% to $10 billion from the second quarter to the third, up from 18% growth in the second quarter, while the company continued to report heavy operating losses.

OpenAI grew revenue by roughly 50% to $10 billion from the second to the third quarter, but continued to report heavy operating losses.The company shared its latest results with investors on Tuesday, just as Anthropic was preparing to mount its much-anticipated IPO next week. The maker of Claude recently disclosed that its revenue rose to $17 billion in the third quarter while reporting a smaller operating loss of $567 million during that period. The numbers show that OpenAI is accelerating revenue growth but still has a long way to go to match its rival’s sales and operating efficiency. The company grew revenue by only 18% in the second quarter compared with the first. OpenAI told investors Tuesday that its adjusted operating margin improved from minus 150% to minus 109% from the second to the third quarter, according to documents viewed by The Wall Street Journal. Even so, the company’s adjusted operating losses ticked up slightly from $10.1 billion to $10.9 billion, based on those figures. The calculations for the losses exclude stock-based compensation, which ticked up to $2.6 billion in the third quarter, according to a person familiar with the matter. Anthropic’s operating loss includes stock-based compensation.

Takeaways by Macro Roundup® AI

  1. OpenAI's revenue rose roughly 50% to $10 billion from the second quarter to the third, up from 18% growth in the second quarter, while the company continued to report heavy operating losses.
  2. OpenAI's adjusted operating margin improved from minus 150% to minus 109% between the second and third quarters, yet its adjusted operating losses rose from $10.1 billion to $10.9 billion, excluding stock-based compensation of $2.6 billion in the third quarter.
  3. Anthropic disclosed third-quarter revenue of $17 billion and an operating loss of $567 million that includes stock-based compensation, ahead of its planned IPO next week.

AI Summary. Anthropic's bankers are weighing raising the share price above the announced $70-$78 range for its initial public offering after heavy investor demand, which could lift the valuation target from $1.4tn to $1.5tn or more and push total funds raised beyond $100bn.

Date Posted:

Anthropic’s bankers are increasing the firm’s valuation to $1.5T, which would increase the total funds raised in the IPO to over $100B. “Three of the company’s backers said Anthropic was now focused on raising as much as possible via the listing to support its heavy spending on computing power, rather than maximising its valuation.”

Is investor demand pushing an AI company's IPO valuation higher?

Core argument: Anthropic's bankers expect to price its IPO above the announced $70–$78 range, which could lift the valuation target to $1.5tn or more, up from the $1.4tn disclosed in a filing.

Anthropic’s bankers are weighing plans to increase the AI group’s initial public offering price after receiving heavy demand from investors, less than a day after disclosing that it would target a $1.4tn valuation when it lists next week. Bankers have indicated they expect to raise the share price above the $70-$78 range announced in a filing early on Tuesday, which could increase the valuation target to $1.5tn or more in the coming days, according to five people familiar with the matter. A boost in the share price would also increase the total funds raised in the IPO beyond $100bn. Three of the company’s backers said Anthropic was now focused on raising as much as possible via the listing to support its heavy spending on computing power, rather than maximising its valuation.

Takeaways by Macro Roundup® AI

  1. Anthropic's bankers expect to price its IPO above the announced $70–$78 range, which could lift the valuation target to $1.5tn or more, up from the $1.4tn disclosed in a filing.
  2. A higher share price would push the total funds raised in Anthropic's listing beyond $100bn.
  3. Three of Anthropic's backers said the company is focused on raising as much as possible in the IPO to support heavy spending on computing power, rather than maximising its valuation.

AI Summary. Anthropic plans to seek a valuation of about $1.33tn in its initial public offering, below the roughly $1.5tn investors had anticipated and the approximately $2tn figure floated earlier. Some major investors pushed for the lower figure following a steep selloff in artificial-intelligence companies.

Date Posted:

Anthropic is seeking to raise $61B at a valuation of ~ $1.33T in their upcoming IPO. Alongside the IPO, the firm is raising $26B from large investors, including Google and Advanced Micro Devices.

Is an AI selloff lowering the valuation sought in a major IPO?

Core argument: Anthropic will seek a valuation of roughly $1.33 trillion in its initial public offering, proceeding with its stock-market debut despite concerns that the AI boom is on shaky ground.

Anthropic will seek a valuation of around $1.33 trillion in its initial public offering, forging ahead with its stock-market debut despite concerns that the artificial-intelligence boom is on shaky ground. A week ago, investors anticipated the company to target a valuation around $1.5 trillion, but after the steep selloff in AI companies Thursday, some major investors pushed for it to be lower. Anthropic’s targeted valuation is far below the roughly $2 trillion figure that had been floated months ago.

Takeaways by Macro Roundup® AI

  1. Anthropic will seek a valuation of roughly $1.33 trillion in its initial public offering, proceeding with its stock-market debut despite concerns that the AI boom is on shaky ground.
  2. Investors had expected Anthropic to target a valuation near $1.5 trillion, but some major investors pushed for a lower figure after a steep selloff in AI companies.
  3. Anthropic's targeted $1.33 trillion valuation falls far below the roughly $2 trillion figure floated months earlier.

AI Summary. Long-term U.S. Treasury yields did not fall during an artificial-intelligence equity sell-off, which the author argues contradicts the view that AI capital spending and hyperscaler debt issuance are the main force pushing yields higher. The author attributes elevated long-term yields instead to a rising fiscal risk premium tied to U.S. public debt growth.

Date Posted:

Brooks notes that while the 2-year and 10-year Treasury yields fell during last week’s AI sell-off, the 10y10y forward yield did not. He argues its failure to do so “further cements the case for the fiscal risk premium,” concluding that long yields are up “because of mounting fiscal stress, not because of AI capex.”

Are elevated long-term yields driven by fiscal strain rather than AI borrowing?

Core argument: Robin Brooks argues that long-term Treasury yields failed to fall during the late-2026 AI equity sell-off, contradicting the view that AI capital expenditure and hyperscaler debt issuance have been the primary force pushing yields higher.

The black line shows daily data from the start of 2026 for the two-year Treasury yield.The blue line is the 10-year yield and the purple line is the 10y10y forward yield. If the AI bubble were a material driver of the rise in long-term yields, the purple line should have fallen this week alongside the blue and black ones. It didn’t and that in my mind further cements the case for the fiscal risk premium. The midterm elections and the AI sell-off that’s ensued have given us the perfect natural experiment to disentangle what’s going on with long-term yields. They’re up because of mounting fiscal stress, not because of AI. Once the war ends - and it will - we’ll be left with all the structural challenges the US and others in the G10 face. The fact that we’re rapidly running out of fiscal space is top of that list.

Takeaways by Macro Roundup® AI

  1. Robin Brooks argues that long-term Treasury yields failed to fall during the late-2026 AI equity sell-off, contradicting the view that AI capital expenditure and hyperscaler debt issuance have been the primary force pushing yields higher.
  2. The November payrolls print of 165,000 jobs beat the Bloomberg consensus of 70,000 by 95,000, or 1.2 standard deviations against a historical standard deviation of 76,000, and the two-year yield rose five basis points, in line with or above the historical regression of yield responses to payrolls surprises.
  3. Brooks concludes that elevated long-term yields reflect mounting fiscal stress rather than AI, noting the 10-year/10-year forward yield did not decline alongside shorter maturities during the sell-off week and that corporate dissaving is “not remotely near where it was in the IT bubble of the early 2000s.”.