A Global Monetary Crisis Is Coming & AI Could Make It Worse | Stanley Druckenmiller

November 3rd, 2010. I'm sitting in my office at Duquesne Capital staring at a number on a screen that made my stomach drop. Not because it was catastrophic, because it was a warning sign that virtually nobody else was paying attention to. The Federal Reserve had just crossed a threshold that no central bank in modern history had ever successfully crossed back from. Their balance sheet had hit two thousand trillion. Two trillion dollars of printed money used to paper over the damage from the 2008 financial crisis. And I remember picking up the phone, calling my head of research, and saying two words. I said, this ends badly. He said, when? I said, I don't know when, but I know how, and I want to be positioned accordingly. Fifteen years later, sitting here in February 2026, I want to tell you that we have crossed not one threshold, but three of them simultaneously. And this time there is a fourth variable in the equation that didn't exist in any previous crisis. A variable that most economists haven't fully processed. A variable that changes the math, changes the speed, and changes the severity of what is coming in ways that I find genuinely alarming. That variable is artificial intelligence. And by the end of this video, you will understand exactly why the combination of a collapsing monetary system and an accelerating AI revolution doesn't just create a financial crisis. It creates something that has no historical precedent. Something that every serious investor needs to understand and position for right now. I'm Stanley Druckenmiller. I ran Duquesne Capital Management for 30 years without a single losing year. 30 years, not one. I generated an average annual return of 30% across every kind of market condition you can imagine. Raging bull markets, brutal bear markets, currency crises, tech bubbles, housing crashes. I've seen everything this system can throw at an investor. And I'm telling you with complete seriousness that the setup I'm looking at right now in February 2026 is the most complex and potentially most dangerous convergence of risks I have seen in nearly five decades of doing this professionally. I'm not saying this to scare you. I'm not saying this to get clicks. I'm saying this because I spent 49 years building a track record based on telling people the truth even when the truth is uncomfortable. Especially when it's uncomfortable. So let's get into it. Let me start with the monetary crisis because you need to understand this foundation before I explain how AI makes it worse. And I'm going to use numbers that a high school student could follow because this isn't complicated math. It's just math that most people haven't looked at honestly. Right now today, February 17th, 2026, the United States government carries 36 trillion in debt. But that headline number doesn't even tell you the real story. When you add corporate debt, household debt, and financial sector debt together, total US debt across the entire economy hit the 98.8 trillion in the second quarter of 2025. That is 324 percent of our entire gross domestic product. 324 percent. For every single dollar of economic output this country produces in a year, we collectively owe 3.24 in debt. And it's growing faster than the economy is. The interest payment on just the federal portion of that debt, just the government's share, has now crossed one trillion dollars a year for the first time in American history. One trillion dollars. Every single year. Just an interest, not paying down what we owe. Just servicing the cost of existing debt. And here's the number that I keep coming back to because it captures something important about where we are. That one trillion dollar annual interest payment is now larger than the entire United States defense budget. Think about that for five full seconds. The most powerful military force in the history of human civilization costs about 80-60 billion a year. The interest on money, we already borrowed costs more than that. We spend more keeping our creditors happy than we spend keeping our enemies at bay. But here is where it gets truly serious. That one trillion dollar interest payment was calculated at an average interest rate of roughly 3.2 percent across all existing U.S. debt. The problem is that the debt being issued right now, the new borrowing happening today, is going at rates of 4.5 percent to 5 percent. As the old cheap debt that was issued during the zero interest rate era matures and gets replaced with expensive new debt, that interest payment is not going to stay at one trillion. The Congressional Budget Office projects it will consume 27 percent of all federal tax revenues within a decade. Half of all revenues within three decades. The arithmetic isn't complicated, but the implications are staggering. Now here is the event that is unfolding in real time as I speak to you. In 2026, approximately nine to ten trillion dollars worth of U.S. Treasury debt is scheduled to mature. One quarter of all U.S. government debt, money that was borrowed when rates were near zero during the pandemic, needs to be refinanced this year at dramatically higher interest rates. The government has to go back to the market and say please lend us money to pay back the money we already borrowed. And please do it at the highest rates in a generation. This is what I call the debt wall. And it is not theoretical. It is a scheduled event. It is happening right now. And the question of who buys all that new debt is the single most important financial question of 2026. Because the answer to that question determines everything else. It determines interest rates. It determines the dollar. It determines equity valuations. It determines inflation. It determines your retirement. Everything. So let me tell you who used to buy U.S. Treasury debt and why they're stepping back. For decades the global financial system ran on a simple arrangement. Countries that ran trade surpluses. Countries like China and Japan that sold more goods to America than they bought. They took the dollars they earned and recycled them back into U.S. Treasury bonds. It was a self-reinforcing system. America spent. The world lent. America spent more. And it worked beautifully for 50 years because everyone trusted that the U.S. government would always be good for the money. China at its peak held over 1 trillion dollars in U.S. Treasuries. Today that number is a 731 billion and falling. China sold off over 300 billion worth of U.S. government bonds and replaced them with something else. I'll tell you what they replaced them with in a moment. Japan, which holds roughly 13% of all U.S. Treasury debt outstanding, is under enormous pressure to sell its holdings to defend the Japanese yen, which has been in freefall. Japanese bond markets in January 2026 were described by professional traders as chaotic. Saudi Arabia and OPEC nations are diversifying away from dollar-denominated assets as the geopolitical landscape shifts and as the logic of petrodollar recycling weakens. European central banks are gradually reducing their dollar exposure. And here is the threshold that I believe future historians will mark as one of the great turning points in monetary history. In 2025, for the first time in three decades, global central banks combined gold holdings exceeded their total holdings of U.S. Treasury bonds in their international reserves. Read that sentence again. Central banks collectively now hold more gold than U.S. Treasury bonds. The institutions that set the rules of the global monetary system just voted with their balance sheets. They chose gold over American debt. That is not a minor portfolio adjustment. That is a signal. So if the traditional buyers are walking away from Treasury bonds, who picks up the slack? Who buys the nine trillion dollars of new debt that needs to be rolled over this year? The answer right now is domestic institutions. U.S. banks, U.S. pension funds, U.S. money market funds, domestic insurance companies. The government is essentially borrowing from its own financial system to fund its obligations. And this creates a problem that economists call crowding out. Every single dollar that a U.S. bank puts into a Treasury bond is a dollar that doesn't go into a small business loan, a corporate bond, a mortgage, a startup investment. The government is hoovering money out of the private economy to fund its deficit at the exact same time. The Federal Reserve was until recently running quantitative tightening, removing 60 or a billion per month from the financial system, which is another 720 billion per year drained out. Add the Treasury borrowing pressure to the Federal Reserve's balance sheet contraction and you have roughly 2.7 trillion of liquidity being pulled out of the financial system in 2026 alone. The market capitalization of Apple, the largest company on earth, is roughly three die trillion. We are removing almost one entire Apple worth of money from the system. And people are looking at the S&P 500 near all time highs and wondering why something feels wrong underneath the surface. Here is what's happening underneath the surface and why the AI piece of this story is so critical and so underappreciated. Let me tell you how I first started thinking seriously about artificial intelligence because I'm not someone who jumps on technology trends. I've watched dozens of them. My analyst came to me about three or four years ago and said, Stan, something's different. The best engineering graduates from Stanford and MIT, they're not going into crypto anymore. They're all going into AI. I've seen that movie before in reverse. I watched the same caliber of minds flood into internet companies in 1998 and 1999. And what I've learned is that when the best human capital in the world concentrates on a single technology, the technology delivers. ChatGPT launched in November 2022. I took notice. I bought NVIDIA in the fourth quarter of 2022. I wrote it up 700% before eventually exiting the position. I believe artificial intelligence is as genuinely transformative as the internet was. I said publicly that it could be more transformative. I meant it. But here's the thing about transformative technologies. They don't just create winners. They also create disruption. And the specific kind of disruption that AI is creating right now is interacting with our existing monetary and fiscal crisis in a way that makes the crisis significantly more dangerous. Let me show you exactly what I mean. Microsoft's 2025 analysis of AI's impact on the labor market identified approximately 5 million white collar jobs, specifically management analysts, customer service representatives, sales engineers, paralegals, junior financial analysts that are facing significant automation pressure. Klarna, a Swedish financial technology firm, reduced its workforce by 40% as it implemented AI systems. Salesforce cut its customer support headcount from 9,000 to 5,000. IBM replaced several hundred human resources roles with AI. Amazon has eliminated more than 30,000 positions since late 2025, including 16,000 in the early weeks of 2026, tied directly to AI driven restructuring. Ford's CEO said AI would cut in half the number of white collar jobs at the company. JP Morgan and Goldman Sachs are both harnessing AI specifically to employ fewer people. According to Challenger, Gray and Christmas, the employment research firm nearly 55,000 US job cuts in 2025 were directly attributed to AI automation. That's a 400% increase year over year. The World Economic Forum estimates 92 million jobs will be displaced globally by 2030. These are not fringe predictions from technology pessimists. These are forecasts from the most establishment institutions in global finance and business. Now I want to be precise here because I think about this carefully and I've seen people on both sides of this argument make mistakes. There are credible economists at Oxford and at Yale's budget lab who argue that AI's actual labor market impact so far has been modest and concentrated that the macro data doesn't yet show widespread disruption. And they're right that the high level employment numbers have been relatively stable. But I've been doing this for 49 years and I know something about lag effects, about how structural changes in an economy take time to show up in the aggregate statistics. The subprime mortgage crisis started showing cracks in early 2007. The S&P 500 didn't peak until October of that year. For nine months the foundation was rotting while the roof looked fine. The question isn't whether AI disruption has fully shown up in the employment data yet. The question is whether the direction of travel is clear and the direction of travel is unmistakably clear. When Mustafa Suleyman, the CEO of Microsoft AI, says that all white color work could be automated within 18 months, that's not a fringe view. That's the CEO of the AI division of the world's largest software company describing his own products trajectory. Here is why this matters so much for the monetary crisis and why the two stories are not separate. They are the same story. The United States government's fiscal model depends on a very specific thing. It depends on employed workers paying income taxes and payroll taxes. The federal income tax and the payroll tax together account for roughly 85% of all federal government revenue. When you have five million white collar workers who were earning 80,000, 100,000, 120 thousands a year and paying substantial taxes on those incomes and those positions are replaced by AI subscriptions that cost $20 a month, the government loses an enormous amount of tax revenue and simultaneously it gains enormous amounts of fiscal pressure because those displaced workers need unemployment insurance, retraining programs, potentially health care subsidies, possibly eventually some form of income support. The government's spending goes up at the exact same time that its revenue base is being eroded. This is not theoretical. This is the arithmetic of what happens when the tax base shrinks faster than you can cut spending or raise other taxes and it hits at the worst possible moment. At the exact moment when the government already faces a $9.6 trillion debt refinancing wall and a trillion dollar annual interest payment that is already consuming more than the defense budget. But there's a second dimension to this that is even more unsettling. When companies replace workers with AI, they generate higher profits. Profit margins in big tech increased by more than 20% in the fourth quarter of 2025, while the broader market saw almost no change. So you get this widening split. The companies that own the AI infrastructure get richer. The workers displaced by the AI get poorer. Corporate tax revenues might not collapse because profits are rising, but individual income tax and payroll tax revenues, the fiscal backbone of the government, those face structural pressure. And here is what I find most disturbing when I sit down and think through the full implications. If AI driven displacement becomes large enough, if unemployment rises significantly in a structural rather than cyclical way, the political pressure on governments to respond with massive fiscal stimulus becomes overwhelming. I've seen this movie too. When economies are hurting and people are scared, governments spend money. They don't cut spending. They don't raise taxes enough to close the gap. They borrow and they spend and they pray that growth comes back fast enough to make it manageable. But what if the growth doesn't come back? What if the profits go to AI owning companies while the tax base of ordinary working people erodes? Then you get a debt spiral that feeds on itself. Government borrows more. Interest payments rise. Fiscal position deteriorates. More borrowing needed. More pressure on the bond market. Higher yields. Higher interest costs. More borrowing. This is what I mean when I say these two stories are the same story. And the monetary crisis and the AI disruption are reinforcing each other. And the Federal Reserve, the institution that is supposed to manage all of this, is trapped in a way that has no historical precedent. Let me explain why by comparing this moment to every previous crisis I have navigated. In 1987, when markets crashed 22% in a single day, the Fed cut rates and flooded the system with liquidity. Crisis over in months. In 2000, when the dot-com bubble burst, the Fed cut rates from 6% to 1%. Problem solved, more or less. In 2008, when the financial system almost ceased to exist, the Fed cut rates to zero and printed $4,000 trillion in new money. Massive pain, but the system survived. In 2020, when COVID shut down the global economy, the Fed printed $5 trillion in six months. Markets recovered to all, time highs within a year. Every single crisis had the same solution. The Fed had room to cut rates. The Fed had credibility to print money. Inflation was low enough that they could act aggressively without stoking a bigger problem. That escape hatch is now closed. Inflation right now is still running around 3%, well above the Fed's 2% target. The Fed cut rates in December 2025 to a range of 3.5% to 3.75%. But cutting more aggressively risks reigniting inflation back towards 7 or 8 or 9%. And I've said publicly that I'm nervous about exactly that scenario. The Fed may have declared victory over inflation too early. Gold at 5,000 ounces is telling me the market agrees. Gold doesn't lie. It's been money for 5,000 years and it doesn't have a PR department. When gold goes vertical, it's telling you something about how the smartest money in the world views the credibility of paper currencies and the institutions that manage them. And on top of all of this, we have the political pressure on the Federal Reserve itself. Kevin Warsh has been nominated to succeed Jerome Powell as Fed chair when Powell's term ends. I've watched Fed chairs come and go for 49 years. And the nomination of a new chair at this particular moment, when the government faces $9.6 trillion in debt that needs to be refinanced at the highest rates in a generation, creates an almost irresistible pressure toward what economists call financial repression. That means keeping interest rates artificially low relative to inflation so the government can borrow cheaply even as inflation erodes the real value of the debt it's paying back. It's a hidden default. Not an explicit one. Not the government saying we won't pay you back. Just a slow grinding erosion of purchasing power that transfers wealth from bondholders and savers to the government. If that's the path we go down. And I think the fiscal math makes it nearly inevitable. It means every dollar sitting in a savings account, every pension fund loaded with bonds, every fixed income portfolio loses real value year after year. Not dramatically. Not all at once. But relentlessly. Now let me tell you what I'm watching. What I believe the signals are. And how I'm thinking about positioning. Because analysis without positioning is just conversation. And I'm not in the conversation business. I'm in the money business. The first signal I watch every single day is the 10-year Treasury yield. It's currently trading in the neighborhood of 4.5 to 4.7 percent. If it breaks above 5 percent and holds there for three consecutive trading days, that is my red alert. That means the bond market is beginning to price in genuine fiscal risk on US government debt. It means international investors are demanding a real premium to hold Treasury bonds because they're no longer confident the real return justifies the risk. When that happens, the interest cost spiral I described accelerates dramatically. Nine trillion dollars of debt refinancing at five percent instead of three percent costs hundreds of billions more per year in perpetuity. That's not manageable within the existing fiscal framework without either massive tax increases, massive spending cuts, or inflation. None of those are politically easy. All of them are economically painful. The second signal is what I call the AI employment shock indicator, which I'm constructing from a combination of data sources. I watch the monthly layoff announcements from the Challenger Gray and Christmas reports specifically for AI. Attributed job cuts. I watch initial jobless claims for any signs of acceleration. And I watch wage growth data in the professional and business services sector, which is the white collar segment most exposed to AI automation. Right now the Challenger data shows a 400 percent year over year increase in AI attributed layoffs. That number is going to keep rising. When it crosses 200 Zalausred attributed job cuts in a single quarter, when you start seeing that show up in the headline unemployment rate, that's when the fiscal pressure I described begins to show up in the government's revenue numbers. And that's when the bond market has to price and not just the current debt load but an accelerating deficit path. The third signal is central bank gold buying. I watch the World Gold Council data obsessively. In 2025 central banks collectively crossed that extraordinary threshold of holding more gold than US Treasury bonds in their international reserves. For the first time in three decades that signal cannot be overstated. These are the institutions that created the dollar based financial system. When they start quietly, methodically, consistently choosing gold over that system's anchor asset, they are telling you something about where they think this is going. If central bank gold buying stays above 600 tons per quarter through 2026, as JP Morgan is projecting, that tells me the world's most sophisticated institutional investors are in structural agreement with my analysis. When smart money votes with its balance sheet rather than its words, I listen. Now let me tell you how I'm thinking about the portfolio in this environment. And I want to be more specific than I usually am in public. Because I think the specificity matters. I want meaningful exposure to gold and precious metals. Not a tiny allocation. Not the 5% that most advisors suggest as an afterthought. Genuine, meaningful exposure. Through physical gold or physically backed funds like GLD or the Sprott Physical Gold Trust. Through gold mining majors like Barrick and Newmont and Agnico Eagle. Because when gold rises 20% these companies profits rise 60% or 80% thanks to operating leverage. The companies pull gold out of the ground at $4,200 to $1,100 per ounce and sell it at whatever the market will bear. Right now the market will bear close to $5,000. That's a $3,500 profit per ounce. As gold rises further those profits compound extraordinarily. On the AI side of this, I'm not throwing the baby out with the bathwater. I sold NVIDIA in late 2024 when I felt the valuation had gotten ahead of the near term reality. But I bought Amazon, Meta and Alphabet in the third quarter of 2025 and I remain long those positions. These are the companies that own the infrastructure of the AI economy. The data centers, the cloud platforms, the advertising ecosystems that AI will run on. If AI creates enormous deflationary pressure on human labor costs and enormous profits for the companies deploying it, you want to own the companies deploying it. But you want to hedge that long exposure with hard assets. Because the monetary and fiscal consequences of AI driven labor disruption feed directly into the case for gold. The two positions actually reinforce each other rather than contradict each other. Long AI infrastructure. Long gold is a hedge against the monetary consequences of AI disruption. It's the same trade expressed two different ways. I want to be cash heavy. Not because cash is a great long term store of value. I know it's not. But because cash is optionality. Cash is power. Cash is the ability to act decisively when dislocations create opportunities. When markets panic. When credit seizes up. When everyone needs to sell and there are no buyers, the person with cash wins. I keep a significant portion in Treasury bills earning four to five percent while I wait. People call that stupid. They say I'm missing the rally. What they don't understand is that the cost of missing a 10% rally by sitting in cash earning 5% is about 5% of opportunity cost. But the cost of being fully invested when something breaks in this environment is 30, 40, 50 percent of your net worth. That's not a symmetric risk. That's a completely asymmetric setup. And I'll take the optionality of cash every single time when the fundamentals are this uncertain. I want to reduce exposure to the most interest. Rate sensitive parts of the market, specifically long duration bonds and highly leveraged real estate. Commercial real estate in particular is a sector that is not yet fully processed. The reality of higher rates, lower occupancy from remote work and tighter credit availability from regional banks that are themselves sitting on unrealized losses. The pain in commercial real estate is not over. The restructuring of office buildings, the refinancing stress on retail properties, the regional banks that hold the loans against all of this. I want very limited exposure there and I want to be selective and careful about the zombie company problem. There are hundreds of publicly traded companies that operated perfectly well in a zero interest rate world and are slowly suffocating at 4 to 5 percent rates. Companies with high debt, weak cash flow and business models that only really worked when money was free. When you mix a tightening credit environment with AI driven competition that can replace their expensive human workforces in ways they never anticipated. Some of these companies are going to fail and they're going to fail faster than the market currently. I want to also say something about the one mistake I watch investors make over and over again in environments like this. Because I've made this mistake myself and it cost me real money before I learned better. The mistake is being right about the direction and wrong about the timing and letting the frustration of being early cause you to abandon a correct position or worse to start second. Guessing the underlying analysis. In 1999 I was convinced that the tech bubble would burst. I was right about the destination. I was a year early on the timing and during that year the market went up another 40 percent and I looked like an old man who didn't understand the new economy. You know what? I didn't abandon the analysis. I adjusted my position to reduce the carrying cost of being early and I waited. When March 2000 came the market had its worst two years to year stretch since the Great Depression and I was positioned to profit from it. Being early is not the same thing as being wrong. In markets the distinction between those two things is the difference between wealth and bankruptcy. The investors who built generational wealth from the 2008 crisis weren't the ones who called the exact bottom. They were the ones who understood the fundamental thesis clearly enough and positioned conservatively enough that when the crisis arrived they had capital. They had conviction and they had the ability to buy at fire sale prices when everyone else was in panic mode. I want to finish by returning to the two things I started with because everything in this video connects back to them. The monetary system is under more structural stress than at any point in my 49 years of professional investing. Not because of any single event. Not because of any single policy mistake. But because of the accumulated consequence of 50 years of decisions that prioritized short-term stability over long-term sustainability. Every crisis was solved by borrowing more, printing more, pushing the problem forward. And here we are in 2026 with 36 trillion in federal debt, a 9.6 trillion refinancing wall, a 1 trillion dollar annual interest payment that exceeds the defense budget, foreign creditors stepping back, and a Federal Reserve that is trapped between fighting inflation and managing an unsustainable debt load. That is the setup and into that setup comes artificial intelligence, the most powerful labor, displacing technology in history, hitting the white-collar workforce in a way that erodes the income tax base that funds the government's operations while simultaneously concentrating profits and productive capacity in the hands of a relatively small number of technology platforms. The rich economy, poor people dynamic is real. S&P 500 profits at record highs. Unemployment rising. Consumer sentiment near all-time lows. Layoff announcements in 2025 at the highest level since the COVID pandemic. These data points are not contradictions. They are the early stage fingerprints of a structural shift that is going to interact with the monetary crisis in ways that will likely surprise most investors. I've spent 49 years preparing for moments like this, not by predicting exactly what happens or exactly when, but by understanding the forces at work clearly enough to position for the range of outcomes in a way where I survive the bad scenarios and prosper from the opportunities they create. The setup today is the clearest I have seen in my career in terms of direction. The uncertainty is timing, but I'm not going to wait for certainty to act because in markets certainty arrives only after the opportunity has passed. The time to position is before the storm, not during it. You have that time right now, today, in February 2026. The window is open. The signals are clear. The arithmetic is undeniable. What you do with that information is entirely up to you, but I'll tell you what I'm doing. I'm building real asset exposure, holding substantial cash optionality, reducing fragile and rate-sensitive positions, and watching those three signals every single day because when they move, they'll move fast. And in this market, fast is the only speed that matters. Position accordingly. I'll see you on the other side.