China’s “Government-Backed Companies Are Missing Payments. How Bad Is China’s Debt?

China's real estate crisis has just crossed another line. Zhuhai Huafa Group, one of the country's largest local government-owned conglomerates, has missed payments on two trust products. It's also asked at least three banks to extend existing loans. Now, Huafa isn't some obscure private developer that borrowed too much during the housing bubble. It's owned by the local government in Guangdong. It's developed more than 200 residential projects and helped build landmarks including the Zhuhai Tower. And as recently as March, The company reported approximately 413 billion yuan, or nearly 62 billion U.S. dollar equivalent, in interest-bearing liabilities. This is the largest property-dependent company to seek delayed payments since China Banking began doing the same thing about a year ago. It is a major escalation. China's real estate bust began with heavily indebted private developers. It spread to homebuyers, construction companies, trust firms, and local governments. Now it's reaching large enterprises whose state ownership was supposed to make them safe. China's banks and regulators continue publishing remarkably reassuring statistics about loan performance. Their behavior tells us they don't believe those statistics. Banks over there are making fewer loans, accumulating government bonds, and driving sovereign yields to record lows. Yes, record low rates in China. That's not the behavior of a banking system preparing for recovery. It's the behavior of banks protecting themselves from a crisis. And yes, this is a banking crisis, not necessarily one characterized by deposit runs and banks closing their doors, but one in which lenders distrust borrowers, avoid risk at all costs and redirect money away from the local economy. Wafa shows us why that's the case. If even state-owned firms are missing payments, Chinese banks have no choice but to become defensive. And the banking crisis over there also helps explain, if not entirely explain, what's going on in the Chinese stock market. The AI bubble turned AI bust. These are not different stories. They're all the same story from separate angles. And it starts with, what is it? Yes, definitely a Chinese banking crisis. For years, China's property crisis was described as a manageable cleanup of reckless private developers. Evergrande, supposedly different. So was Country Garden. Each failure could be blamed on a particular company's aggressive borrowing, poor management, or speculative land purchases. The state sector would remain stable. That distinction has steadily collapsed. Vanke, once China's largest developer and a company with substantial state backing, began seeking help with its debts. Now, Huafa has missed payments and requested loan extensions. Whatever legal or regulatory label officials eventually apply, this is a technical default in every economic sense. Money was due. It was not paid on time. Creditors are being asked to wait. And Huafa's scale makes that impossible to dismiss. Its liabilities include bank loans, trust financing, and bonds, so any restructuring therefore reaches across several parts of China's financial system. The market understood the significance immediately. One of Huafa's offshore yuan-denominated bonds fell by a record 17%, bringing it to approximately 74 cents on the yuan. Shares of its listed property arm dropped as much as 8.9%. What makes the episode especially revealing is that Wafa had priced a new 1 billion yuan bond at only 2.07% just days earlier. Investors were still treating a local government-owned borrower as as relatively safe. Then missed payment reports forced them to reconsider that assumption almost overnight. Large part of Chinese credit has been priced around implied state support. Investors assume strategically important enterprises would receive assistance before creditors suffered meaningful losses. And Huafa may still receive that assistance. Loans could be extended. Assets could be transferred. Banks might be pressured to refinance. And local government officials, they're still going to provide some guarantees. But Rescue doesn't eliminate the underlying loss, or more importantly, The very reasonable fears there really is an avalanche of bad debts sitting unpublished on bank balance sheets all over the country. Instead, WAFA's technical defaults are just more proof that is indeed the case. Most AI presentation tools have the same problem. Everything they make looks exactly the same. But Gamma just launched Gamma 5, and I gotta say, this is pretty darn impressive. I can give it a rough idea for a presentation and rather than immediately spitting out a bunch of slides, its agent will ask me questions and then give me an outline that I can actually edit. Then it turns that outline into a fully designed presentation. But here's what's crazy. I can take that exact same content, change the style, And it will give me something that looks completely different. I can even import an old PowerPoint, redesign it, and export it with the formatting intact. So whether you're pitching investors or presenting to clients or just running a business in a company, you can create something that you'd actually be proud to put your name on. So why don't you give Gamma a try? And you can do that. You can try Gamma 5 at gamma.app. And as all this is going on, China's stock market has been suffering a substantial AI-driven rout. After share prices soared along with everyone else early part of this year, the AI bubble really taking hold, in China, the rout, maybe even the bubble going bust, is far more advanced. October produced the market's worst monthly decline in more than a decade. The blue-chip CSI 300 fell into bear market territory from its late June peak, while the semiconductor-heavy Star 50 plunged as much as 48%. Beijing is responding with the same emergency measures it used during previous rescues. The People's Bank of China is providing cheap funding for corporate share buybacks and liquidity for institutional investors. The so-called national team appears to be buying shares. Regulators are reportedly using informal guidance to discourage brokerages and large institutions from selling equities. Those measures tell us more than official statements ever could. Governments don't instruct institutions not to sell when markets are healthy. Central banks don't subsidize share purchases when private investors are confident. Officials are attempting to interrupt a liquidation. The selling has been concentrated in the same technology and AI companies that previously led the market higher. State-owned insurers had increased their exposure to those stocks, partly in response to official pressure to supply the market with long-term capital. China Life, for example, raised equities to 19.2% of its portfolio by the end of June, up from 16.9% in 2025, while expanding AI-related investments. That left insurers exposed near the peak because, of course. But blaming insurers for the crash confuses the sellers with the reason they're selling in the first place. The deeper problem is that China's AI boom depends on enormous amounts of capital, as we know only too well. Chipmakers, data centers, cloud operators, and semiconductor manufacturers need years of financing before the promised returns become available. It isn't any different there from here. That financing was supposed to come from the same banking system that is now trying to avoid property developers, local government borrowers, and even state-owned conglomerates. If Chinese banks don't want those risks, how in the hell are they supposed to fund what might be the biggest risky crapshoot since subprime mortgages? China's stock crash isn't merely a change in sentiment. It is the equity market discovering that the credit needed to deliver the AI payoff just doesn't exist. And of course, as all that's going on, the published statistics from Chinese regulators and even the banks themselves show no problem whatsoever in terms of loan performance. But we all know that's not actually the case. It can't possibly be the case. The Chinese blew the biggest asset bubble humanity has ever seen. And that asset bubble has been busting for several years now, which means there are bad debts and there may be bad debts all over the system as we're getting here with Zhuhai Huafa. Whatever they're publishing out there, Chinese banks are acting as though the risks are indeed much worse. New yuan loans totaled 1.08 trillion in September, down from 1.29 trillion a year earlier. Outstanding loan growth, as I've been pointing out repeatedly, slowed to a record low 4.8%. During the first nine months of the year, new bank lending fell to 11.5 trillion yuan from 14.75 trillion during the same period a year earlier. And that same period a year earlier Earlier at $14.75 trillion, Was not good. That's a decline of more than 3 trillion yuan despite repeated government efforts to encourage lending. Household borrowing in China remains especially weak and that's downplaying it. Household loans were merely flat in September after contracting by more than 200 billion yuan in August. And for the year, they are down, yes, contracting by a trillion yuan after rising by a trillion in the first nine months of 2025. And that was an alarmingly low loan total. Lower interest rates haven't restored mortgage demand because households don't want to borrow against homes they expect could lose even more value. Businesses, they don't want to invest when sales are weak and not expected to get any better. And banks, they don't want to extend credit when collateral values and anticipated cash flows are obviously deteriorating. The official loan statistics also obscure what happens after a borrower gets into trouble. If a bank extends the maturities we're seeing here with WAFA, rolls over the loan, or even avoids formally recognizing the impairment, The debt might not appear immediately as non-performing. The accounting label remains reassuring, even though the economic quality asset has clearly declined. And now WAFA asking for banks to extend their loans. And if those extensions are approved, the immediate default may disappear from the headline figures. The bank, however, still has the exposure to a borrower. They couldn't meet the original schedule. This is how reported loan performance can remain rosy in public. While bank behavior becomes increasingly defensive. The banks know what is buried on their balance sheets. They don't need to wait for regulators to officially recognize it or the outside world to realize what's really going on over there. And it really is spilling over into something like artificial intelligence. You need the money to finance all those massive expenses. It's not just happening in the United States. The Chinese in a race with the U.S. need the capital to do so, but they can't get it. And they're not likely to get it because, as we see with WAFA, there really is a credit crisis still ongoing in China. And it isn't getting better, as we know from the bond market. It's actually getting worse. China needs all the same power generation, transmission equipment, data centers, cooling systems, enormous quantities of advanced hardware. And those all require... Credit, not eventually, today. Large technology companies, they can use their own internal cash flow. They can issue some bonds. They can do some equity financing. But the wider AI supply chain consists of smaller manufacturers, local projects, and a whole bunch more that depend heavily, especially in China, on the banking sector. And those borrowers now face a banking system attempting, not just attempting, but actually and actively and heavily reducing their risk. That creates a fundamental contradiction. Beijing wants banks to support strategic technology, local governments, and to finish some of these unfinished housing projects, not to mention troubled state-owned companies and support them, all at the same time. But a bank's balance sheet can't absorb unlimited politically directed lending without consequences. Every yuan used to extend a loan to a struggling property conglomerate is a yuan that can't be deployed elsewhere without increasing leverage, leverage that they increasingly don't want. Every loss hidden through restructuring reduces the bank's capacity to finance a genuinely new project. And every new AI loan forces the bank to ask whether the proposed collateral will retain its value. Advanced chips and computer equipment can become obsolete very quickly. Data centers, they're valuable only if customers use them and eventually pay for them. Expected future demand is not the same thing as current cash flow. There's tremendous risk in AI at a time when Chinese banks are actively de-risking. Now, the distinction didn't matter much while technology stocks were rising. Investors could extrapolate extraordinary future growth and assume financing would remain available. Xi Jinping would figure it out somehow. It matters enormously when banks continue to retreat and get more and more reminders about why they're doing so. The Chinese AI bubble was built on the expectation the state could mobilize. Effectively unlimited credit behind the industry, that's what everybody assumes. Huafa is evidence that state-directed credit system is already overloaded and unprepared. If governments have to be ready to jump in for developers who built years ago, they can't fill in the funding gap created by banks who have no room to build tomorrow's tech boom. The money just isn't there and China's stocks have figured this out. And the bond market confirms everything that we just went over, including the behavior of banks and what's going on in lending. Risk aversion. They don't want to extend risky loans, so therefore buy safe and liquid. Depression economics and then some, to the point that we're seeing Chinese bond yields, central government bond yields, hit record lows, record low rates in China, consistent with banking crisis. We've seen lows just recently at the five-year maturity and the 10-year maturity. Those yields are often explained as evidence of expected monetary easing. That's complete nonsense. Central bank policies follow the market, not lead the system out of the downturn that low market rates were warning about. Bond prices are also rising because Chinese banks are in the market for safe assets. When lenders trust the economy, they make private loans. Those loans, they offer better returns than government securities. And expanding credit creates deposits, it creates investment, and it creates economic activity, which we know in China isn't taking place. When lenders distrust the economy, the calculation changes. As I've been telling you, everything is about risk-adjusted returns. A private loan may offer a much higher rate of return, but the repayment of that loan, if it's in question, a lower yielding, even much lower yielding government bond is the much better bet. Sovereign bonds, they offer liquidity, they offer regulatory advantages and protection from the uncertainty surrounding property collateral and now corporate cash flows, even among the largest state-owned firms. Record low yields, therefore, they reveal a powerful defensive depression bid. Banks are accepting lower returns in exchange for lower perceived risk because they perceive tremendous and growing risk everywhere in China. That is the defining mechanism of this, yes, banking crisis. Again, banking crisis doesn't require crowds outside of branches demanding their cash back. It can operate quietly through lending committees, collateral haircuts, maturity extensions, and the gradual substitution of government bonds for private credit or the accelerating Subtitling government bonds for private credit. The result is still just as destructive. Viable companies find financing harder to obtain. Troubled companies consume more resources through repeated extensions. Economic growth weakens, by collateral value falls further, and banks become SUBS even more defensive. In China, they want to win the AI race for the future, but Chinese banks are Hamburg riddled with massive mistakes of the past. And that feedback loop is now expanding beyond Huafa. Another Zhuhai borrower, Daheng King Group, recently held meetings to reassure creditors after Huafa's missed payments triggered concern across the entire city. Daheng King postponed a private bond sale because investors demanded a higher premium. It told creditors it could repay a 600 million yuan bond maturing in November with cash if market conditions failed to improve. The firm also reportedly lent 2 billion yuan to a joint venture with, yes, Huafa. And guaranteed over 3 billion yuan of project financing. So yes, they're trying desperately to reassure investors because investors are now asking more and more questions. That's how contagion works. Not necessarily through panic, but through these kinds of close connections. The same one Chinese banks know about, even if the public only gets the occasional glimpse into them. And through it all, only one part of the Chinese system is even actively working. And that is, of course, the export boom. Of which AI is a huge part of it. That means that the Chinese economic system anyway is hugely, hugely dependent on a very narrow part of the Chinese system. And that strength can make the wider economy appear healthier than it really is. Industrial production has continued expanding, especially with strong output in computers and communications technology, electrical and machinery. But the factories are producing into a domestic economy where customers and borrowers are actively retreating. And we see that in retail sales that have been unusually weak this entire year across China. Automobile sales are down big time, furniture sales, building material, you basically name it. Consumer spending in China is in real serious trouble. Fixed asset investment continues contracting at substantial rates. Property investment is collapsing again. And weakness is spreading beyond just strictly real estate. So China can manufacture more AI-related equipment for foreign buyers, but exports can't fully repair household balance sheets or restore confidence in housing. They also expose the economy to a dangerous concentration. If the global AI boom slows, China loses one of the few remaining sources of strong external demand. If Chinese banks simultaneously refuse to finance domestic AI expansion, both sides of the industry are going to weaken at the same time. That's what technology stocks have been pricing over there. The promised AI future depends on a long chain of assumptions. Continuing foreign demand, domestic investment, cheap financing, government support, sufficient final revenue to justify all that spending. The property crisis is breaking one of the most important links in that chain, credit. Now, Huafa, they're likely to be rescued in one form or another. That doesn't make this, however, some kind of false alarm. If anything. It further proves and establishes the banking crisis and the reasons behind it. A disorderly failure would threaten a lot more than property projects, it would threaten local employment, banks, trust companies, state-owned borrowers, so officials have every incentive to arrange extensions or move assets and liabilities to stronger institutions. In fact, that's one reason why the Chinese central government has been doing its bank recapitalization and insurance company recapitalization. But rescuing Huafa requires someone else to assume the risk. And it also confirms what the system or those outside the system. Already suspects this is indeed a big problem and it is getting worse, not better, no matter how much the CCP declares stimulus and its unbreakable intent to fix everything. Banks can be ordered to extend loans, but that leaves them with more low-quality assets and less capacity to support actually productive investment. Local governments can provide assistance, but many are already burdened by falling land sale revenue and enormous financing vehicle debts of their own. Other state-owned firms can purchase assets, but that merely transfers exposure within the public sector. The immediate default may be avoided While the system becomes more constrained. That's why banks are buying government bonds. They understand that political support can prevent a visible bankruptcy without restoring the borrower's underlying cash flow, the fundamental proposition behind the loan to begin with. They also understand Huafa may not be the last supposedly safe state company to request an extension or even worse than that. Once state ownership no longer guarantees timely payment, every similar borrower must be reassessed. Risk premiums rise, bond sales become harder, banks demand more collateral, credit becomes scarcer, and the crisis only broadens further. The most important warning for WAFA isn't that one conglomerate could fail. It's that a company with local government ownership, hundreds of projects, and nearly $62 billion US dollar equivalent in interest-bearing liabilities could no longer meet every payment on schedule. If that borrower isn't safe, Chinese banks have to ask, what else isn't safe? They're already answering that question. They're making fewer loans and they are accumulating a whopping amount of government bonds, therefore driving sovereign yields to record lows as the real estate crisis spreads now to the state sector. It is indeed a banking crisis, even if it doesn't look like the one you've been trained to see. Chinese banks see it. Chinese banks are acting on it. And that is what really matters.