Economy & Markets
Sep 23, 2025
The Blob: Capital, China, Chips, Chicago and Chilliwack
Michael Cembalest
Chairman of Market and Investment Strategy for J.P. Morgan Asset & Wealth Management
In this piece, we look at the AI and data center takeover, and the OpenAI-Oracle deal; the US government equity investments in Intel and MP Materials, the origins of TSMC and how many countries support national champions via industrial policy; efforts in China to reduce excess capacity and consequences for equity investors; crime and municipal solvency in Chicago and Illinois; how tight net new equity supply has been supporting US equity markets since 2011; and pictures from Chilliwack, Canada.
Descriptive audio
Good afternoon, everybody. This is Michael Cembalest with the almost October 2025 Eye on the Market podcast. This one’s called “The Blob,” and I’m gonna cover a few different topics: AI data centers, the Oracle OpenAI deal, China, chips, Chicago, Chilliwack, bunch of different things.
So I made a fake movie poster of a movie called The Data Center Blob, starring Amazon Web Services, Microsoft Azure, Google Cloud, etc. And it also shows who the producers are, the directors and who financed it. And it’s a, it’s a recall of the 1950s Blob movie with Steve McQueen. And what I, what I was trying to capture is the extent to which AI and data centers are kind of just taking over and gobbling up all sorts of different things.
We put a table together and some of the numbers here are kind of amazing. There are, we have a basket of AI-related stocks, and they’re either, most of them, about 30 of them, are direct AI plays, and the other 10 are either utilities or capital equipment stocks. So those 40 AI stocks make up some staggering number, whether it’s 70 to 75 to 80% of all the returns and the earnings growth in the market since ChatGPT was launched in November 2022, and make up over 90% of the capex and R&D growth since that same period.
So this is by far the highest concentration I’ve ever seen. There were high concentrations of tech stocks at the end of the 90s to price returns, but they certainly didn’t have that high share either of capital spending or earnings. So this is the first time you’ve seen this kind of concentrated earnings growth, price returns, capex, concentration in, you know, 40 stocks out of the S&P 500.
And it’s, it’s a, it’s, it’s an amazing thing to see. And this is happening just as data centers are eclipsing office construction spending, and are also coming underneath, under increased scrutiny for their impact on power grids and rising electricity prices. I’m starting to work on the 2026 energy paper, and some of the research that we are doing is showing that the specialized power rates that some of the utilities are, are, applying to data centers, even though they’re higher than traditional rates, are, aren’t, aren’t nearly enough to pay for the cost of a new combined cycle natural gas power plant, which costs have doubled to about $2,500 a kilowatt over the last couple of years. And in the PJM region, the, the watchdog that pays attention to that particular grid did some analysis and found that 70% of the increase in electricity prices in 2024 was due to data center demand. So the, the AI Data Center Blob is taking, not just taking over equity markets, it’s also taking over power markets. And you can see in the same chart what’s happened to construction spending on electric power more or less has been rising at the same pace as construction spending on data centers.
But now the other big news in the AI space was this Oracle OpenAI partnership, and Oracle stop, uh stock jumped by 25 to 30%. And it was kind of an interesting sign at the time because they were, they were, OpenAI is promising to pay $60 billion a year, which is money they don’t earn, for Oracle to provide facilities for them, which they haven’t built. So there’s a lot of stuff that’s on the gun here. These facilities will eventually require Oracle to build out four-and-a-half gigawatts of power, right, so for people that don’t live in the world of megawatts and gigawatts, that’s over two Hoover Dams or forge, or four large nuclear plants. That’s how much four-and-a-half gigawatts is. And it will, also there’s no way that Oracle is going to earn this money out of its current cash flow. They’re going to have to borrow.
And there’s a chart we have in here that’s kind of a staggering one. Uh, the debt-to-equity ratios at Amazon are 50%, 30% at Microsoft, even less at Meta and Google. At Oracle, they’re 500%. And Oracle is by far and away the highest debt-to-equity ratio of any of the direct AI stocks in our universe.
So we’re up until now, the hyperscalers have been funding this AI boom through cash flow. If you’re now going to get companies joining in who are gonna, who are gonna amplify this even further through borrowing, you have the risk of an even greater AI arms race with, with, with another crescendo in capital spending and related earnings activity next year. So this thing looks like it’s going to keep running, assuming that Oracle is able to borrow the money to start building this stuff. So that is, that’s kind of a remarkable development in the AI space, is to see this, just how Oracle at its debt level is gonna take on even more debt.
So the other thing that happened was the, the U.S. government equity investment in Intel. You know, Intel is one of the stocks that’s in our direct AI basket, but it is most certainly not part of the reason that stocks have performed well since ChatGPT was launched. Intel earnings and market cap have actually declined since November 2022. There are a lot of critics of this deal where the U.S. government is taking a 10% stake in Intel, but I’m not one of them. The horse has left the barn in a year. It’s way too late in the game to assume that somehow market forces are going to help support U.S. domestic chip fabrication. And you know, the, the, the, the CHIPS bill is, is, it is not, it’s more than a drop in the bucket, but it doesn’t get anywhere more than 10 to 15% domestic production self-sufficiency for chips in the U.S.
So Intel needs an anchor. It’s, it needs an anchor demand to make it, the concept of a chip foundry viable. But what customer is going to go with Intel instead of one of the other established vendors? If there’s a chance the foundry won’t make it. And so, this is a perfect example of, of where an anchor investment from the government makes plays to make sense.
And Nvidia, by the way, followed last week with a $5 billion partnership with Intel. SoftBank then provided $2 billion, and other fabulous chip companies may follow. And so, you know, I, there have been so many successful entrepreneur stories in the U.S. The U.S .has an amazing track record generating new companies just in the private sector. But the U.S. lives in, in a very mercantile world. And there are times in places when these kind of equity investments from the government are going to make sense.
Let me just remind you, where TSMC, which is one of the most admired companies in the world, where did they come from? Well, in the 1970s, Taiwan was facing a triple whammy—the end of military and economic aid. And, and we have a chart in here that goes back to 1946 and shows the military and economic aid to Taiwan petering out by 1977. They were thrown out of the UN, and there was an energy crisis. And so Taiwan decided that they were going to use semiconductors as a, as a lifeline to get out of this mess.
And RCA, which was a U.S. company, was the only one that was willing to license its semiconductor design to Taiwan. And TSMC was founded by Taiwanese American Morris Chang with, with what, $100 million in private sector money? No; $100 million seed money from the Taiwanese government, which took a 49% stake. And even though the Taiwanese government’s share has declined since the 70s, its National Development Fund is still the largest shareholder in TSMC. It owns 6 to 8% of the company, and it has a seat on the company’s board. The Taiwanese government continues to, to support TSMC in spite of how successful it now is, with subsidized prices for electricity, tax credits. They build industrial parks and research centers to provide the labor that TSMC needs, and TSMC’s market cap is now greater than Taiwanese GDP.
And when you look at the history of South Korea and Samsung, and the Netherlands and ASML, which is the most valuable European tech company, you get a similar story. So it’s a very, very mercantile world out there. And I consider it a little bit naive to think that the U.S. is going to be able to resolve all of its supply chain issues by just relying on market forces, when most other large developed and developing countries don’t do that.
So I have a few charts in here that, that you might not have seen before that were really interesting to me. The first one looks at government expenditures on subsidies from the year 2000 to around 2020. France, Sweden. Germany, much, much higher than the U.S. Then a really interesting analysis from CSIS, which is one of the best think tanks in DC, that looks at industrial policy spending in key economies. And they look at state investment funds below market credit, government support for R&D, tax incentives, direct subsidies. And, of course, China blows everybody away. China spends almost 2% of GDP on this kind of thing. But South Korea, France, Germany and Taiwan are ahead of the United States. So another example where the United States is, is not playing the same industrial policy game as other countries.
We have a chart in here that looks at the state-owned bank share of banking system assets. The U.S. bar is so small it barely shows up, but something like 1%. And you got much higher bars on here, not just in emerging markets, but in advanced economies like Germany, the Netherlands, South Korea, Switzerland, the U.K. and France. So another example of when the other countries are playing a different game.
And as for China, their direct subsidies at, for some industries account for, for anywhere from 10 to 30% of the profits in that entire industry. And for that reason, unsurprisingly, an analysis that was done a couple of years ago found that China was the most mercantilist country in the world, when looking at forced local production. You know, if you do business there from another country, you have to produce locally. IP protections, market access limits, domestic benefits preferences, currency manipulation. China blows everybody away. But then you got India, Brazil, Indonesia, Argentina, Thailand, Russia. And so it’s, it’s again, it’s a very mercantile world out there. And so, I felt the same way about the United States Department of Defense and Energy investment in materials for critical minerals.MP Materials is having to compete with Chinese critical mineral companies, only around half of which are even regulated by China. So it’s, it’s obvious that if we really want to secure some of those supply chains, these kind of deals are going to be a thing of the future.
Okay. As it relates to China, the buzzword of the day, which has been helping its equity markets, is “involution.” And I hadn’t heard of this word before. The financial services industry is great at coming up with new words. It makes it seem like something interesting is happening. So what’s involution? Well that refers to excess capacity and oversupply in China, which has led to a really weak period of industrial profits and deflation. China is trying to stimulate enough demand to consume all of this extra production is going on there, but it hasn’t been very successful. So now instead of trying to stimulate demand to absorb this extra supply, they’re gonna allegedly push banks to stop lending to these companies that are producing the oversupply solely for meeting GDP targets. This is a tough ask for the Chinese corporate sector, which has been exporting a lot of its excess production to the rest of the world.
And here, we have a chart you can see that China’s now in the longest period of deflation, certainly since, like, the late 90s. And when you decompose the things that are working in, in China, everything having to do with exports looks great. Everything having to do with the domestic economy looks very weak. And I think there’s two charts in here which really make that point. Exports versus corporate earnings, that there’s definitely a divergence taking place there. And then obviously as everybody in Europe is now aware, Chinese domestic vehicle sales are kind of flattish, but Chinese vehicle production is soaring because of how much of that they’re exporting to the rest of the world.
So Chinese equities have done pretty well since this involution campaign was announced. And there is evidence that China is cutting back on some fixed asset investment, whether it’s in manufacturing or infrastructure or real estate. So that’s so, it does look like there’s some follow-through by the government to reduce the amount of extra production, but it’s not clear how long this will last. And it’s also not clear who the winners and losers are inside China, because some companies that you might want to buy are going to have their production forcibly cut. I think the safer play is this: Focus on the OECD industrial companies, which would finally benefit from some reduction in global Chinese oversupply because they’d be the ones to finally get some pricing power back. And the MSI Developed World Industrialists Index has started to pick up a little bit. I think it makes a lot more sense to focus on that than trying to pick the winners and losers inside China.
Okay, so now I’m going to switch from China to Chicago. And the president, as many people were aware, was at one point considering sending federal troops or the National Guard into Chicago. He, he decided not to do that, and is reportedly going to send them to Memphis instead. But it made me wonder, just how bad the crime situation is in Chicago. We hear a lot about it, and I thought I would look at the numbers now because a lot of different sources for crime data. There are issues that crime is often underreported. So, but my best source, the one I prefer to look at, is this thing called the Council on Criminal Justice. They do a lot of cleanup of other FBI and other sources, they clean up that data. And they have a report that comes out twice a year, and they just released their mid-year 2025 update. Unfortunately, a lot of the cities you’d like to compare them to don’t always report on every category. But we have a grid in here that shows where Chicago ranks, and they rank pretty high on certain things like aggravated assault, carjacking, gun assaults, robbery and certain kinds of burglaries. They’re at the bottom of the top third on homicides. And so if you were just going to use this kind of data to think about federal intervention, cities like Memphis, Baltimore or Saint Louis seem like better candidates than Chicago because they seem as dangerous as Chicago, if not more so. But Chicago’s homicide rate is five times higher than New York’s. And if we narrowed the universe just to the few mega cities with at least 750,000 people, China ranked number one in aggravated assaults, gun assaults, homicide and robbery. So I think it’s fair to assume that Chicago is grappling with a crime level problem, even though, as the governor and mayor of Chicago like to point out, those levels are down from where they were a couple of years ago.
And, you know, the mayors also brag. The mayor of Chicago has also bragged recently about improved clearance rates, like in other words, the resolution of certain crimes are being resolved faster. You know, not so fast. So what is a clearance rate? It doesn’t always mean someone was arrested. It just means that the police now figure out they think they know, they think they know who did the crime. So maybe prosecutors don’t move to trial, but they think they know who did it, so it’s cleared. Or someone that we’re going to charge and never got around to it died of old age, cleared. And of the, of the actual 140 Chicago murders that occurred in 2023 and were cleared, charges were only filed in 80% of those cases. And if we zoom out to the whole decade of 2013 to 2022, the percentage of murders in Chicago that were cleared by an arrest, rather than just somebody saying, I think I know who did this, was only 20%. And that compares to over 50% in New York and a national average of 45%. So the gap in Chicago between perception and reality of crime solving is so great that Illinois legislators have introduced a bill that would require law enforcement to disclose the number of people arrested and charged, rather than just cases cleared.
So kind of an example of when, of when, you know, a local government doesn’t really telling you the facts you’d want to know. And then since 2018, of all the 19,000 people wounded in a non-fatal shooting in Chicago, arrests were only made in 6% of cases, and 80% of non-fatal shooting investigations are suspended each year so that officers are no longer investigating. And then there’s some other stats here. Over 70% of arrested and charged individuals are released before trial. Four hundred people over the last couple of years were arrested for murder or attempted murder were out on bail for a prior felony. And then, you know, a sign of the times, there was a photo earlier this month that showed Governor Pritzker posing with someone that was described as a community violence intervention worker, like, you know, to prevent crime. And a couple weeks later, that person was charged with murder after a smash and grab burglary, and car escape went awry when they robbed a Louis Vuitton store on the Magnificent Mile.
So now you would think that Chicago would be inclined to try to spend some money to solve or address some of these issues, but that’s what brings us to the municipal solvency question. Does Chicago, and Illinois more broadly, do they have the money to spend?
And that gets us back to our Ark and the Covenants project. So for those of you that have been longtime readers of the Eye on the Market, you’ll remember that in 2014, I started this project to measure the total liabilities of U.S. states because the traditional rating agencies don’t do that. But they certainly did do it then, and they’ve gotten a little bit better at it. But so we built these models and we pour through thousands of pages of consolidated annual financial reports because we’re trying to get at this question of how much debt do they have, and what are their unfunded pensions, and what are their unfunded retiree healthcare obligations as a percentage of city or state revenues?
And when we last did our city analysis a couple of years ago, Chicago came in as the most indebted city. And little, unsurprisingly, little has changed since then. And so we show this chart here that shows the cost of pension and retiree healthcare plans, including the average amortization of the unfunded amounts, and we’re, we’re generous. I amortize those unfunded amounts over 30 years, okay. And we also look at government debt in defined contribution plans as a percentage of city revenues. So, and, you know, so Chicago’s in pretty bad shape. And the other thing, because they have the highest ratio, something, somewhere between 35 to 40% of all of the revenue collections of the city of Chicago would be required to deal with the interest on a debt and the amortization of its unfunded pensions, retiree healthcare.
The other problem Chicago faces is that if you, if you’re looking at a dollar of an unfunded pension, that’s much worse than the dollar of an unfunded retiree healthcare obligation because the first one’s contractual and immutable, whereas the second one, you can change co-pays, you can change deductibles, you can change coverage rules. And a lot of states and cities and counties have been adjusting those retiree healthcare rules over time.
So, and here’s the odd part. The Chicago Police Department, which would be responsible for any changes in law enforcement policies, I wonder how they feel. They have a woefully underfunded pension. The Chicago Police Annuity Benefits Fund has a 24% funding ratio. It’s the second lowest of all the pension plans in our universe. And the only one that’s worse is, is a certain California Judges Retirement Fund.
So, and then the picture is similar when you look at the state of Illinois. So we did the analysis at the state level as well. Illinois comes in a close number two behind Connecticut as the most indebted state. But again, Illinois has a much greater share than Connecticut in pensions relative to healthcare. So what I would argue that Illinois is actually in worse shape. Illinois also continues to show up either as number one, two or three on the list of states experiencing the highest outflow of people, which is not going to make resolving these issues any easier. So there you go. My wife’s family is from Chicago, and this is just goes to show that I don’t play any favorites.
Okay. Then just a couple more quick things. Thank you for listening this long. I actually remember there was a, there was a Stephen Hawking book years ago that they sold, and they did an experiment where they, they put a voucher for $100 on, like three-quarters of the way through the book to see if anybody that bought it actually read it that far. And only like 10% of the people that bought the book ever redeemed the voucher. So if you’ve gotten to this far in the podcast, you know, maybe I owe you a voucher of some kind.
So the last thing I want to show is the U.S. equity markets have been remarkably resilient to a lot of shocks since 2011. And I’m not a technician, so to speak, a market technician, but I do think there’s a technical answer for part of that, and it has to do with global net equity supply. If you, going back to 1999, generally every year, the U.S. and the non-U.S. regions had an increase in global net equity supply, meaning primary and secondary issuance was greater than, than buybacks and other thing, and, and mergers and things that took companies private.
Since 2011 with, with the temporary exception of that idiotic SPAC boom, since 2011, U.S. net equity supply has been negative. It’s been, it’s been declining in dollar terms. And at the end of the day, there’s $1.5 trillion of annual, not to the stock, but annual defined benefit and defined contribution payments into qualified plans by households and employers. That’s got to get it, end up invested someplace. And so, they’re still running it. Those contributions are still running at about 3 to 4% of S&P market cap every year, while the supply of U.S. equities continues to shrink. So when we think about how the U.S. equity markets have been so resilient to shocks, I think this is part of the answer.
And then just to close out, here are some pictures. And I also included in the Eye on the Market a link to an Instagram reel. I had another, another great trip to the Fraser River in Vancouver, where I have my, I have a kayak stashed there, and I fish for giant sturgeon. That first one that you’re seeing there is about seven feet and 200 pounds. And a great time was had all around. And thanks to Rachel for going with me on this trip. But I don’t have any pictures of her with fish.
Thank you very much for listening, and we will see you soon. We’re working among other things on our biennial alternative investments review, which I expect to have out before the end of the year. Bye.
(DESCRIPTION)
Title card: JP Morgan, Eye on the market. JP Morgan. September 2025. Image: A large, angry building with cables and plugs bursting from its mouth looms over frightened people and bent power lines. Yellow and red text surrounds the building with phrases like STRAINED POWER GRID!, HIGHER POWER PRICES!, and RAPID DEPRECIATION OF CHIPS AND INTERCONNECT! A black section at the bottom lists companies like AMAZON WEB SERVICES, MICROSOFT AZURE, GOOGLE CLOUD, and credits a parody film titled THE DATA CENTER BLOB! in red letters. Title: The Blob, Capital, China, Chips, Chicago, and Chilliwack. Video box on the right. Michael Cembalest has short hair, glasses, and wears a dark zip-up jacket with light blue trim. He sits in front of a virtual modern shelf that holds vases, gold decorative items, books, and geometric objects.
(SPEECH)
Good afternoon, everybody. This is Michael Cembalest with the almost-October 2025 Eye on the Market podcast. This one's called "The Blob." And I'm going to cover a few different topics, AI data centers, the Oracle-OpenAI deal, China, chips, Chicago, Chilliwack, a bunch of different things.
So
(DESCRIPTION)
Slide: The Blob: the A.I. and data center takeover.
(SPEECH)
I made a fake movie poster of a movie called "The Data Center Blob," starring Amazon Web Services, Microsoft Azure, Google Cloud, et cetera. And it also shows who the producers are, the directors, and who financed it. And it's a recall of the 1950s Blob movie with Steve McQueen. And what I was trying to capture is the extent to which AI and data centers are kind of just taking over and gobbling up all sorts of different things.
We put a table together. And some of the numbers here are kind of amazing.
(DESCRIPTION)
Slide. A table compares AI-related stock performance and contributions to the S&P 500 since November 2022. AI Direct shows the highest returns, earnings, and growth, while AI CapEquip reports negative capex + R&D growth. Data source: Bloomberg, JPMAM.
(SPEECH)
We have a basket of AI-related stocks and they're either-- most of them, about 30 of them, are direct AI plays, and the other 10 are either utilities or capital equipment stocks. So those 40 AI stocks make up some staggering number, whether it's 70% to 75% to 80%, of all the returns and the earnings growth in the market since ChatGPT was launched in November 2022 and make up over 90% of the CapEx and R&D growth since that same period.
So this is by far the highest concentration I've ever seen. There were high concentrations of tech stocks at the end of the 90s to price returns, but they certainly didn't have that high share, either of capital spending or of earnings. So this is the first time you've seen this kind of concentrated earnings growth, price returns, CapEx concentration in 40 stocks out of the S&P 500. And it's an amazing thing to see.
And this is happening just as data centers are eclipsing office construction spending and are also coming underneath-- under increased scrutiny from their impact on power grids and rising electricity prices. I'm
(DESCRIPTION)
Slide. A line chart shows US construction spending from 2014 to 2025. Electric power leads, rising sharply after 2020. Office buildings peak in 2022, then decline. Data centers show steady growth, surpassing $40 billion by 2025. Spending is measured in billions, seasonally adjusted annual rate. Source: Census Bureau, JPMAM, July 2025.
(SPEECH)
starting to work on the 2026 energy paper and some of the research that we are doing is showing that the specialized power rates that some of the utilities are applying to data centers, even though they're higher than traditional rates, aren't nearly enough to pay for the costs of a combined cycle natural gas power plant, which costs have doubled to about $2,500 a kilowatt over the last couple of years.
And in the PJM region, the watchdog that pays attention to that particular grid did some analysis and found that 70% of the increase in electricity prices in 2024 was due to data center demand. So the AI data center blob is not just taking over equity markets, it's also taking over power markets. And you can see in this same chart what's happened to construction spending on electric power more or less has been rising at the same pace as construction spending on data centers.
Now the other big news in the AI space was this Oracle-OpenAI partnership. And Oracle stock jumped by 25% to 30%. And
(DESCRIPTION)
Slide. A bar chart compares debt-to-equity ratios of Direct AI stocks, with most companies under 100%. Oracle stands out at nearly 500%, far above others like AMD, Apple, and Broadcom. A sidebar explains Oracle’s rise due to a $60B OpenAI deal, increased borrowing, and future power demands. Source: Bloomberg, JPMAM, September 16, 2025.
(SPEECH)
it was kind of an interesting sign of the times because OpenAI is promising to pay 60 billion a year, which is money they don't earn, for Oracle to provide facilities for them, which they haven't built. So there's a lot of stuff that's on the come here.
These facilities will eventually require Oracle to build out 4.5 gigawatts of power. So for people that don't live in the world of megawatts and gigawatts, that's over two Hoover Dams or four large nuclear plants. That's how much 4.5 gigawatts is.
And it will also-- there's no way that Oracle is going to earn this money out of current cash flow. They're going to have to borrow. And there's a chart we have in here that's kind of a staggering one. The debt-to-equity ratios at Amazon are 50%, 30% at Microsoft, even less at Meta and Google. At Oracle, they're 500%. And Oracle is by far and away the highest debt-to-equity ratio of any of the direct AI stocks in our universe.
So up until now, the hyperscalers have been funding this AI boom through cash flow. If you're now going to get companies joining in who are going to amplify this even further through borrowing, you have the risk of an even greater AI arms race with another crescendo in capital spending and related earnings activity next year. So this thing looks like it's going to keep running, assuming that Oracle is able to borrow the money to start building this stuff. So that's kind of a remarkable development in the AI space, is to see this, just how Oracle, at its debt level, is going to take on even more debt.
So the other thing that happened was the US government equity investment in Intel.
(DESCRIPTION)
Slide: Chips: the US government equity investment in Intel, TSMC and a world of mercantilist industrial policy. A line chart compares TSMC and Intel market caps from 1995 to 2025. TSMC rises sharply after 2020, nearing $1,000 billion by 2025, while Intel declines below $200 billion. A sidebar notes Intel’s market cap and earnings have dropped since GPT’s launch. Source: Bloomberg, JPMAM, September 12, 2025.
(SPEECH)
So Intel is one of the stocks that's in our direct AI basket, but it is most certainly not part of the reason that AI stocks have performed well since ChatGPT was launched. Intel earnings and market cap have actually declined since November 2022.
There are a lot of critics of this deal, where the US government's taking a 10% stake in Intel, but I'm not one of them. The horse has left the barn here. It's way too late in the game to assume that somehow market forces are going to help support US domestic chip fabrication. And the CHIPs bill, it's more than a drop in a bucket, but it doesn't get anywhere more than 10% to 15% domestic production self-sufficiency for chips in the US.
So Intel needs an anchor-- it needs an anchor demand to make the concept of its chip foundry viable. But what customer is going to go with Intel instead of one of the other established vendors if there's a chance the foundry won't make it? And so this is a perfect example of where an anchor investment from the government makes plays to make sense. And NVIDIA, by the way, followed last week with a $5 billion partnership with Intel. SoftBank then provided 2 billion and other fabless chip companies may follow.
And so there have been so many successful entrepreneur stories in the US. The US has an amazing track record generating new companies, just with the private sector, but the US lives in a very mercantile world. And there are times and places where these kind of equity investments from the government are going to make sense.
Let me just remind you where TSMC, which is one of the most admired companies in the world, where did they come from. Well,
(DESCRIPTION)
Slide. A bar chart shows US military (gold) and economic (blue) aid to Taiwan from 1946 to 2025. Aid peaked in the 1950s, declined sharply by the 1980s, and reappeared slightly after 2000, with a notable military aid spike near 2025. Text on the right reads: End of US military and economic aid, UN expulsion, Energy crisis. Source: Council on Foreign Relations, September 2024.
(SPEECH)
in the 1970s, Taiwan was facing a triple whammy, the end of military and economic aid-- and we have a chart in here that goes back to 1946 and shows the military and economic aid to Taiwan petering out by 1977-- they were thrown out of the UN, and there was an energy crisis. And so Taiwan decided that they were going to use semiconductors as a lifeline to get out of this mess. And
(DESCRIPTION)
Slide. A bullet point list.
(SPEECH)
RCA, which was the US company, was the only one that was willing to license its semiconductor design to Taiwan.
And TSMC was founded by Taiwanese-American Morris Chang with what, $100 million in private sector money? No, $100 million seed money from the Taiwanese government, which took a 49% stake. And even though the Taiwanese government's share has declined since the 70s, its national development fund is still the largest shareholder in TSMC. It owns 6% to 8% of the company and it has a seat on the company's board. The Taiwanese government continues to support TSMC, in spite of how successful it now is, with subsidized prices for electricity, tax credits. They build industrial parks and research centers to provide the labor that TSMC needs and TSMC's market cap is now greater than Taiwanese GDP.
And when you look at the history of South Korea and Samsung and the Netherlands and ASML, which is the most valuable European tech company, you get a similar story. So it's a very, very mercantile world out there. And I consider it a little bit naive to think that the US is going to be able to resolve all of its supply chain issues by just relying on market forces when most other large developed and developing countries don't do that.
So I have a few charts in here that you might not have seen before that were really interesting to me. The first one looks at government expenditures on subsidies from the year 2000 to around 2020.
(DESCRIPTION)
Slide. A line chart shows government subsidies as a share of GDP from 2000 to 2020. France and Sweden trend upward, with France rising sharply after 2010. Germany declines steadily, while the US remains flat below 0.5%. Text says that other countries offer more subsidies to manufacturing than the US. Source: Sweden Board of Trade, 2019.
(SPEECH)
France, Sweden, Germany, much, much higher than the US.
Then a really interesting analysis from CSIS, which is one of the best think tanks in DC, that looks at industrial policy spending in key economies. And
(DESCRIPTION)
Slide. A stacked bar chart shows industrial policy spending as a percent of GDP across key economies. China leads with nearly 2%, driven by China-specific factors, state funds, and R&D support. South Korea, France, and Japan follow with moderate spending. Germany, Taiwan, the US, and Brazil spend less than 0.5%. Source: C.S.I.S., 2022.
(SPEECH)
they look at state investment funds-- below-market credit, government support for R&D, tax incentives, direct subsidies. And, of course, China blows everybody away. China spends almost 2% of GDP on this kind of thing. But South Korea, France, Germany, and Taiwan are ahead of the United States, so another example where the United States is not playing the same industrial policy game as other countries.
We have a chart in here that looks at the state-owned bank share of banking system assets. The
(DESCRIPTION)
Slide. A bar chart compares the share of banking system assets held by state-owned banks across countries. India, China, and Russia lead emerging markets with over 50%. In advanced economies, Taiwan and Germany show the highest shares around 30–40%, while countries like the US, Japan, and Italy have near-zero levels. Source: C.S.I.S., 2022.
(SPEECH)
US bar is so small it barely shows up. It's something like 1%. And you've got much higher bars on here, not just in emerging markets but in advanced economies like Germany, the Netherlands, South Korea, Switzerland, the UK, and France, so another example of when the other countries are playing a different game.
And as for China, their direct subsidies for some industries account for anywhere from 10% to 30% of the profits in that entire industry. And
(DESCRIPTION)
Slide. A horizontal bar chart shows China’s direct subsidies to listed firms by subsector as a share of profits from 2017 to 2020. Software & services leads at nearly 40%, followed by technology hardware and automobiles. Pharmaceuticals receive the lowest share, under 15%. Source: C.S.I.S., 2022.
(SPEECH)
for that reason, unsurprisingly, an analysis that was done a couple of years ago found that China was the most mercantilist country in the world when looking at forced local production-- if you do business there from another country, you have to produce locally. IP protections, market access limits, domestic benefits, preferences, currency manipulation, China blows everybody away. But then you've got India, Brazil, Indonesia, Argentina, Thailand, Russia.
(DESCRIPTION)
Slide. A vertical bar chart ranks the 20 most mercantilist countries by ITIF Global Mercantilist Index score. China leads at nearly 50, followed by India, Brazil, and Indonesia. Scores reflect factors like forced local production, IP protections, and currency manipulation. Canada ranks lowest. Source: Information Technology & Innovation Foundation, 2019.
(SPEECH)
And so again, it's a very mercantile world out there.
And so I felt the same way about the United States Department of Defense and Energy investment in MP Materials for critical minerals. MP Materials is having to compete with Chinese critical mineral companies, only around half of which are even regulated by China. So it's obvious that if we really want to secure some of those supply chains, these kind of deals are going to be a thing of the future.
OK, as it relates to China, the buzzword of the day, which has been helping its equity markets, is involution. And I hadn't heard of this word before. The financial services industry is great at coming up with new words. It makes it seem like something interesting is happening.
So
(DESCRIPTION)
Slide. A line chart tracks year-over-year changes in Chinese bank loans to the industrial sector and real estate from 2013 to 2025. Real estate lending surged until 2021, then fell sharply below zero. Industrial lending rose after 2021, peaked in 2023, and declined by 2025. Source: The People’s Bank of China, Bloomberg, JPMAM, Q2 2025.
(SPEECH)
what's involution? Well, that refers to excess capacity and oversupply in China, which has led to a really weak period of industrial profits and deflation. China's been trying to stimulate enough demand to consume all of this extra production that's going on there, but it hasn't been very successful.
So now, instead of trying to stimulate demand to absorb this extra supply, they're going to allegedly push banks to stop lending to these companies that are producing the oversupply solely for meeting GDP targets. And this is a tough ask for the Chinese corporate sector, which has been exporting a lot of its excess production to the rest of the world.
(DESCRIPTION)
Slide. A line chart shows China’s GDP deflator from 1993 to 2025. It starts near 15% in the early 1990s, then fluctuates sharply, with peaks above 10% around 2008 and 2011. Since 2022, the deflator has remained near or below 0%, indicating deflation. Source: Bloomberg Economics, JPMAM, Q2 2025.
(SPEECH)
And here we have a chart. You can see China's now in the longest period of deflation, certainly since like the late 90s.
And when you decompose the things that are working in China, everything having to do with exports looks great, everything having to do with the domestic economy looks very weak. And
(DESCRIPTION)
Slide. A line chart tracks China’s exports and corporate earnings from 2017 to 2025, indexed to 100 at January 2017. Exports rise steadily after 2020, peaking near 180 by 2025. Corporate earnings increase through 2021, then decline and flatten around 120. Source: Haver, China Customs, Bloomberg, JPMAM, August 2025.
(SPEECH)
I think there's two charts in here which really make that point, exports versus corporate earnings, that there's
(DESCRIPTION)
Slide. A line chart compares China vehicle production and domestic vehicle sales from 2016 to 2025. Production rises sharply after 2020, peaking near 35 million units by 2025. Domestic sales remain lower, fluctuating around 26 million. The gap between production and sales widens after 2022. Source: China Association of Auto Manufacturers, JPMAM, August 2025.
(SPEECH)
definitely a divergence taking place there. And then obviously, as everybody in Europe is now aware, Chinese domestic vehicle sales are kind of flattish, but Chinese vehicle production is soaring because of how much of that they're exporting to the rest of the world.
So Chinese equities have done pretty well since this involution campaign was announced.
(DESCRIPTION)
Slide. A line chart shows total equity returns in US dollars from January to September 2025. MSCI China rises the most, reaching above 135. The Shanghai Shenzhen CSI 300 lags but trends upward, while MSCI World climbs steadily, ending just above 115. Source: Bloomberg, JPMAM, September 12, 2025.
(SPEECH)
And there is evidence that China's cutting back on some fixed asset investment, whether it's in manufacturing or infrastructure or real estate.
(DESCRIPTION)
Slide. A line chart tracks year-over-year percentage change in China’s monthly fixed asset investment by sector from 2015 to 2025. All aspects plunge below -15% after 2021. Total fixed asset investment trends downward. Text notes policy shifts on overcapacity and production cuts. Source: Bloomberg, National Bureau of Statistics of China, August 2025.
(SPEECH)
So it does look like there's some follow-through by the government to reduce the amount of extra production.
But it's not clear how long this will last. And it's also not clear who the winners and losers are inside China, because some companies that you might want to buy are going to have their production forcibly cut. I think the safer play is this, focus on the OECD industrial companies, which would finally benefit from some reduction in global Chinese oversupply because they'd be the ones to finally get some pricing power back. And the MSCI developed world industrial index has started to pick up a little bit. I think it makes a lot more sense to focus on that than trying to pick the winners and losers inside China.
So now I'm going to switch from China to Chicago.
(DESCRIPTION)
Slide: Chicago Crime and Municipal Solvency. A table ranks crime rates per capita across major US cities for 2024–2025 by offense type. Chicago ranks #1 in aggravated assault, gun assaults, homicide, and robbery. Memphis, Baltimore, Philadelphia, and St. Louis also rank high across multiple categories. NYC reports fewer categories. Text notes Chicago’s homicide rate is five times NYC’s. Source: Council on Criminal Justice, mid-2025 update.
(SPEECH)
And the president, as many people are aware, was at one point considering sending federal troops or the National Guard into Chicago. He decided not to do that and is reportedly going to send them to Memphis instead. But it made me wonder just how bad the crime situation is in Chicago. You hear a lot about it. And I thought I would look at the numbers.
Now, there's a lot of different sources for crime data. There are issues that crime is often underreported. But my best source, the one I prefer to look at, is this thing called the Council on Criminal Justice. They do a lot of cleanup of other-- FBI and other sources, they clean up that data and they have a report that comes out twice a year. And they just released their mid-year 2025 update.
Unfortunately, a lot of the cities you'd like to compare them to don't always report on every category, but we have a grid in here that shows where Chicago ranks. And they rank pretty high on certain things like aggravated assault, carjacking, gun assaults, robbery, and certain kinds of burglaries. They're at the bottom of the top third on homicides. And so if you were just going to use this of data to think about federal intervention, cities like Memphis, Baltimore, St. Louis seem like better candidates than Chicago because they seem as dangerous as Chicago, if not more so.
But Chicago's homicide rate is five times higher than New York's. And if we narrow the universe just to the few mega cities with at least 750,000 people, Chicago ranks number one in aggravated assault, gun assaults, homicide, and robbery. So I think it's fair to assume that Chicago is grappling with a crime level problem, even though, as the governor and mayor of Chicago like to point out, those levels are down from where they were a couple of years ago.
And the mayor has also bragged-- the mayor of Chicago has also bragged recently about improved clearance rates, in other words, the resolution of certain crimes are been resolved faster. Not so fast.
(DESCRIPTION)
Slide. A bullet point list.
(SPEECH)
So what is a clearance rate? It doesn't always mean someone was arrested, it just means that the police now figure out-- they think they know who did the crime. So maybe prosecutors don't move to trial, but they think they who did it, so it's cleared. Or someone that we're going to charge and never got around to it died of old age, cleared.
And of the actual 140 Chicago murders that occurred in 2023 and were cleared, charges were only filed in 18% of those cases. And if we zoom out to the whole decade of 2013 to 2022, the percentage of murders in Chicago that were cleared by an arrest rather than just somebody saying, I think I who did this, was only 20%. And that compares to over 50% in New York and a national average of 45%. So the gap in Chicago between perception and reality of crime solving is so great that Illinois legislators have introduced a bill that would require law enforcement to disclose the number of people arrested and charged rather than just cases cleared. So kind of an example of when a local government isn't really telling you the facts you'd want to know.
And then, since 2018, of all the 19,000 people wounded in a non-fatal shooting in Chicago, arrests were only made in 6% of cases, and 80% of non-fatal shooting investigations are suspended each year so that officers are no longer investigating. And then there's some other stats here. Over 70% of arrested and charged individuals are released before trial. 400 people over the last couple of years were arrested for murder or attempted murder were out on bail for a prior felony.
And then, a sign of the times, there was a photo earlier this month that showed Governor Pritzker posing with someone that was described as a community violence intervention worker, like to prevent crime. And a couple of weeks later, that person was charged with murder after a smash-and-grab burglary and car escape went awry when they robbed a Louis Vuitton store on the Magnificent Mile.
So now you would think that Chicago would be inclined to try to spend some money to solve or address some of these issues. But that's what brings us to the municipal solvency question, does Chicago-- and Illinois more broadly, do they have the money to spend? And that gets us back to our "Ark and the Covenants Project."
So for those of you that have been longtime readers of the Eye on the Market, you'll remember that in 2014, I started this project to measure the total liabilities of US states because the traditional rating agencies don't do that. They certainly didn't do it then. They've gotten a little bit better at it. So we built these models and we pore through thousands of pages of consolidated annual financial reports because we're trying to get at this question of how much debt do they have, and what are their unfunded pensions, and what are their unfunded retiree healthcare obligations as a percentage of city or state revenues.
And when we last did our city analysis a couple of years ago, Chicago came in as the most indebted city. And little unsurprisingly, little has changed since then. And so we show this chart here that shows the cost of pension and retiree healthcare plans, including the amortization of the unfunded amounts. And we're generous. I
(DESCRIPTION)
Slide. A bar chart compares the cost of pension, retiree healthcare, government debt, and defined contribution plans as a percentage of city revenues across major US cities. Chicago leads at over 35%, followed by Salt Lake City and Dallas. Red dots show lower reported values than recomputed figures. Source: JPMAM, FY 2024.
(SPEECH)
amortized those unfunded amounts over 30 years. And we also look at government debt and defined contribution plans and the percentage of city revenues.
And so Chicago is in pretty bad shape-- and the other thing-- because they have the highest ratio. Something,
(DESCRIPTION)
A stacked bar chart shows the share of pension and retiree healthcare obligations as a percentage of annual required contributions across states. Kentucky, West Virginia, and Illinois have the largest total burdens. Retiree healthcare makes up a smaller portion in most states, except New Jersey and Pennsylvania where it exceeds 40%. Source: JPMAM, FY 2024.
(SPEECH)
somewhere between 35% to 40% of all of the revenue collections of the City of Chicago would be required to deal with the interest on a debt and the amortization of its unfunded pension to retiree healthcare.
The other problem Chicago faces is that if you're looking at dollar of an unfunded pension, that's much worse than a dollar of an unfunded retiree healthcare obligation because the first one's contractual and immutable, whereas the second one you can change copays, you can change deductibles, you can change coverage rules. And a lot of states and cities and counties have been adjusting those retiree healthcare rules over time.
And here's the odd part, the Chicago Police Department, which would be responsible for any changes in law enforcement policies-- I wonder how they feel-- they have a woefully underfunded pension. The Chicago Police Annuity Benefits Fund has a 24% funding ratio. It's the second-lowest of all the pension plans in our universe. And the only one that's worse is a certain California Judges Retirement Fund.
And then the picture is similar when you look at the State of Illinois. So we did the analysis at the state level as well.
(DESCRIPTION)
A bar chart shows states’ costs for pension, retiree healthcare, debt, and defined contribution plans as a percentage of state revenues. Connecticut, Illinois, and New Jersey have the highest burdens near or above 25%. Red dots indicate lower reported costs compared to recomputed values. Source: JPMAM, FY 2024.
(SPEECH)
Illinois comes in a close number two behind Connecticut as the most indebted state. But again, Illinois has a much greater share than Connecticut in pensions relative to healthcare, so I would argue that Illinois is actually in worse shape.
Illinois
(DESCRIPTION)
A stacked bar chart shows the split between pension and retiree healthcare obligations as a percentage of annual required contributions across states. Most states have a larger pension share (blue), while retiree healthcare (tan) forms a significant portion in Pennsylvania and New Jersey, both exceeding 40%. Source: JPMAM, FY 2024.
(SPEECH)
also continues to show up either as number 1, 2, or 3 on the list of states experiencing the highest outflow of people, which is not going to make resolving these issues any easier. So there you go. My wife's family is from Chicago and this just goes to show that I don't play any favorites.
Just a couple more quick things, thank you for listening this long.
(DESCRIPTION)
Slide: This chart might be the simplest explanation for US equity market resilience since 2011: limited supply. A stacked bar chart shows global net equity supply from 1999 to 2025 in billions of US dollars. US equity supply (blue) turns negative after 2009 and remains mostly below zero. Non-US equity supply (gold) fluctuates, peaking in 2020–2021 with added US and non-US SPACs (red and purple). Supply narrows by 2025. Source: JPM Flows & Liquidity, September 2025.
(SPEECH)
I actually remember there was a Stephen Hawking book years ago that they sold, and they did an experiment where they put a voucher for $100 on, like, 3/4 of the way through the book to see if anybody that bought it would actually read it that far. And only like 10% of the people that bought the book ever redeemed the voucher. So if you've gotten to this far in the podcast, maybe I owe you a voucher of some kind.
So the last thing I want to show is the US equity markets have been remarkably resilient to a lot of shocks since 2011. And I'm not a technician, so to speak, a market technician, but I do think there's a technical answer for part of that, and it has to do with global net equity supply. Going back to 1999, generally every year, the US and the non-US regions had an increase in global net equity supply, meaning primary and secondary issuance was greater than buybacks and mergers and things that took companies private.
Since 2011, with the temporary exception of that idiotic SPAC boom, since 2011, US net equity supply has been negative. It's been declining in dollar terms. And at the end of the day, there's $1.5 trillion of annual-- not the stock, but annual defined benefit and defined contribution payments into qualified plans by households and employers. That's got to get and end up invested someplace.
And so they're still running-- those contributions are still running at about 3% to 4% of S&P market cap every year while the supply of US equities continues to shrink. So when we think about how the US equity markets have been so resilient to shocks, I think this is part of the answer.
And
(DESCRIPTION)
Slide: Chilliwack, British Columbia, August 2025. A man wearing a blue jacket and dark cap fishes from a bright yellow kayak. He holds a large sturgeon while kneeling by the shore, fights the fish while seated in the kayak on open water, and later poses with it across his lap in the kayak. Mountains and calm water surround him.
(SPEECH)
then, just to close out, here are some pictures. And I also included in the Eye on the Market a link to an Instagram reel. I added another one, another great trip to the Fraser River in Vancouver where I have a kayak stashed there. And I fish for giant sturgeon, that first one that you're seeing there is about seven feet and 200 pounds. And a great time had all around. And thanks to Rachel for (LAUGHING) going with me on this trip. (SPEAKING NORMALLY) But I don't have any pictures of her with fish.
Thank
(DESCRIPTION)
Returns to title card. Title: September 2025. The Blob, Capital, China, Chips, Chicago, and Chilliwack.
(SPEECH)
you very much for listening. And we will see you soon. We're working, among other things, on our biennial alternative investments review, which I expect to have out before the end of the year. Bye.
(DESCRIPTION)
Logo: J.P. Morgan.
Click here for important information
Read or listen to The Blob: Capital, China, Chips, Chicago and Chilliwack
About Eye on the Market
Since 2005, Michael has been the author of Eye on the Market, covering a wide range of topics across the markets, investments, economics, politics, energy, municipal finance and more.
Share
