Meta Is In Serious Trouble (And Wall Street Is Freaking Out)

Fears about what Meta just reported have spread into bonds, both selling and buying. Facebook's parent reported another enormous increase in future spending commitments tied largely to computing power and AI infrastructure. You know the story by now. Those commitments just for Meta have climbed to roughly $825 billion, up from nearly $700 billion in July, a huge increase in a short amount of time. At the same time, Meta's quarterly free cash flow turned negative, negative $2.12 billion compared with a positive $10.6 billion during the same quarter last year. Meta's shares dropped more than 6%. Its bonds weakened relative to Treasuries, and the price of credit protection on the company jumped to a record high. And it wasn't just Meta. Credit default swaps for Oracle, Microsoft, and Broadcom have also moved higher. SoftBank has lost half its market value since June as investors focus on its leverage and borrowing costs. CoreWeave's debt has already fallen to levels associated with distress, while credit markets have priced a greater than 60% probability of default within five years. Yesterday, we went through why the AI math doesn't work. Today, markets started trading the consequences. Investors sold AI-related stocks and corporate credit and bought treasuries. Treasury yields are down again as the yield curve continues to flatten, with September's technical effects fading and the underlying economic and financial fundamentals regaining their control. There is also an extreme, historically unusual divergence developing between bill rates and term SOFR rates that I'll get to here. One side of the market is looking for another Federal Reserve hike. The other is aggressively hedging against the possibility that deteriorating credit, weaker consumers, and an AI correction will stop the Fed from ever getting to it. This isn't the bond market celebrating AI, it's the market preparing for what happens when a debt-fueled investment boom runs into cash flow reality. A reality where the bond market math just doesn't add up. Start with bond market reaction because there are two very different kinds of bond trading taking place today. Investors are selling or demanding greater compensation to hold corporate bonds connected with the AI build-out. At the same time, they're buying safe and liquid government bonds again now that TGIF is here. Thank God it's October. Meta's five-year credit default swap rose as high as roughly 119 basis points, which is the highest level since those swaps began actively trading about a year ago. Its 5.25% bonds due in 2036 weakened with their spread over Treasuries widening approximately six basis points to around 120. Now, six basis points by itself doesn't sound catastrophic. Meta remains enormously profitable with substantial resources, but the direction matters, as does how that fits into the wider context. This is a company that had been treated as one of the safest participants in the AI investment cycle, yet its credit risk gauge surged to a record immediately after investors learned future commitments had grown by another $125 billion to a ridiculous $825 billion. That's the market asking an extremely important question. Where does all this money come from? And the treasury market's answer is equally revealing. The two-year yield fell to 4.69%, while the 10-year was below 5%. This is what risk migration looks like, at least in the short run. Investors aren't merely changing their opinions about whether meta shares deserve a higher or lower earnings multiple. They're reconsidering the financing structure, supporting the entire AI boom. Corporate risk premiums go up, treasury yields go down. Debt markets are taking control of the narrative. The fundamental problem is that the AI investment cycle always starts out like any other cycle, assuming the adoption process is going to go smoothly, when it never, ever does. In fact, it follows the same pattern as we went over yesterday. The optimistic model says companies spend hundreds of billions of dollars building data centers, customers rapidly adopt AI services, productivity immediately surges, and new revenue arrives just in time to pay off everything but the costs as we know are immediate the revenue on the other hand is hypothetical and in the future data centers have to be built now chips have to be purchased today there's always a collection energy contracts leases network capacity they all have to be secured right away engineers have got to be paid meanwhile customers have to determine whether these products even work whether they can be integrated into existing systems and whether they generate savings large enough to justify their huge costs. Adoption is going to happen, but it will not proceed in a straight line because nothing ever goes in a straight line. There will be implementation failures, budget constraints, disappointing productivity, other kinds of delays, and customers willing or unable to pay what AI providers need to charge. That's what Ken Griffin was talking about. It's especially dangerous because so much of the cycle depends on debt markets remaining cooperative. So CoreWeave represents the more vulnerable end of that structure. Its data center expansion has been financed by issuing debt. Its bonds have fallen to distressed levels, its credit protection has surged, and its shares have been hammered. SoftBank is exposed through a highly leveraged network of investments spanning AI infrastructure, energy, robotics, and technology companies. Its stock has fallen 50% since June, while the cost of insuring its debt has reached the highest level in nearly four years. And yes, there's always a correction. There's always a correction. Meta is far stronger than either company, but when even Meta produces negative free cash flow, while reporting $825 billion in future commitments, investors naturally begin to question the entire chain. The technology can be real while the financial model still fails. history is filled with transformative technologies that were accompanied by terrible investments excessive leverage and companies that just don't survive long enough to reach the promised future before we go on with our video today let me ask you a question what if your gold could actually pay you every single month in more gold that's what today's video sponsor monetary metals lets you do and here's how it works you still own the gold it's still in your name it's It's fully insured, but instead of just sitting there doing nothing, it actually starts working for you. Monetary Metals has built a system that turns idle gold into productive gold, earning real yield, paid in physical gold, not dollars. No selling, no trading, no converting. You keep your gold and every month you get more of it. It's a simple idea, but it solves a big problem. If gold is money, it should act like money, and now it finally can. if you're a long-term gold holder looking for a smarter way to store value check them out at monetary-metals.com slash snyder again that's monetary-metals.com slash snyder i highly recommend booking a call with them today to see if they're the right fit for you we thank monetary metals for their sponsorship but now let's get back to the rest of our video and this brings us maybe unexpectedly to an unusual and historically significant divergence between treasury bills and termed SOFR. This is something that started developing earlier in the energy shock a couple months ago, but the proportions it's gotten to now, never seen anything like this before. At the very front of the market, bill rates continue to reflect the possibility that the Federal Reserve will react to recent reflation reports with another rate hike, another tighten, either next week or perhaps the December meeting. Term SOFR rates and the hedging taking place around them are sending a very different message. That side of the market is heavily protecting against the possibility that it won't be able to hike because financial and economic conditions are deteriorating too rapidly. The divergence, as I said, has reached an extreme level by historical standards. This isn't a schizophrenic market. It's one market pricing two stages of the same policy mistake. Stage one is the Fed looking backward at the CPI, the PPI, and elevated energy prices, then concluding it needs more tighten. Stage two is credit markets looking forward to consumers, leveraged borrowers, and debt-dependent AI companies, then concluding the economy won't be able to withstand all of that. Too many headwinds. Bills are closely connected to immediate policy expectations, cash requirements, and near-term financing conditions. It's a real cash market. The term SOFR reflects expectations and hedging from somewhat different investor classes. They don't have to send identical signals, but typically they trade very close together. So when the gap becomes this large, it tells us that the range of expected outcomes is widening dramatically and that the market expects the Fed to behave irrationally, but any justification for them to do so is rapidly disappearing. That same distinction is visible in the treasury curve. The Fed may still control the overnight rate, but it does not control the markets or the credit cycle. Which brings us right back to Meta. Meta's purported negative free cash flow is something that's impossible to ignore here, especially given these circumstances and especially given the math, the bond market math, the financing math, the credit cycle math that we're all trying to face up to. A year ago, third quarter free cash flow was positive $10.625 billion. this quarter negative 2.12 billion its future leases and non-cancellable contracts increased to approximately 825 billion those commitments aren't the same thing as 825 billion of debt due tomorrow they stretch across time and include obligations to support meta's broader business but again the direction the context are both unmistakable cash generation is falling just as future spending obligations are rising dramatically mark zuckerberg remains confident the investment will produce enormous returns and he may eventually be right but credit markets don't have the luxury of relying entirely on eventually debt has interest payments refinancing dates and contractual obligations it has to be serviced regardless of whether the monetization schedule slips by six months two years or five years and meta is only one participant we keep seeing this over and over again the ai ecosystem and it includes hyperscalers we've got chip manufacturers, data center operators, plus utilities, private credit funds, other leveraged investment vehicles, and on and on and on. Each depends on spending by all the others. One company borrows to construct a data center because another has committed to renting capacity based on demand projections from a third business. That structure works beautifully while funding remains cheap, asset values keep rising, and projected demand is treated as guaranteed. It becomes dangerous When bond investors demand higher yields, collateral values fall, or a major customer reduces its capital spending plans, and the debt market itself becomes more and more inelastic. Stocks can ignore the arithmetic for a while. Credit markets, they just can't. That's why today's reaction is so important. The pessimism isn't confined to equity traders arguing about future valuation multiples. It's entered the market responsible for funding the boom itself. And the Treasury rally that's going along with it gives us even more evidence and more confidence in our interpretation of what happened, especially in the second half of September. And I'm talking about the September effects, all three of them. TGIO, thank God it's October, and the Treasury market as fundamentals and curve dynamics reassert themselves without the interference of the mechanical processes that took place in September. The same technical factors nobody ever talks about or have even noticed and there were three september effects the first was the recurring tendency for global government bonds not just treasuries to drift higher beginning at some point in august continuing through september and sometimes into early october the second of those was the quarter end balance sheet bottleneck dealer capacity repo market elasticity and collateral movement become more constrained around quarter end two weeks before the quarter end to be specific and september's bottleneck tends to be particularly pronounced the third of those involved leveraged treasury trades and self-reinforcing positioning when repo becomes inelastic or more expensive basis traders for one may have to reduce their positions that can require treasury selling or just remove the marginal buyer from the marketplace which accelerates the rise in long-term yields those forces push long-term rates upward particularly during the second half of September and steepen the yield curve artificially in the short run. Now, the mainstream interpretation was that the bond market had suddenly discovered uncontrollable inflation or excessive government debt. Here comes treasury rejection or dramatically stronger growth. But now September is fading and so is the sell off. Rates are declining again with the yield curve itself, right back to flattening to the point of threatening inversion. Even supposedly hot cpi and ppi reports last week were unable to sustain higher treasury yields today ai and credit anxiety adding even more to the safe haven buying that's why it's so important to stay on top of these market signals and interpret them properly to be able to interpret them properly because the yield going up doesn't automatically mean growth or inflation on its own a yield going down doesn't automatically mean the fed has become dovish you have to understand the curve the funding markets collateral mechanics swaps credit and the economic fundamentals beneath all of them and this is again why we're doing the webinar next Thursday October 29th at 6 p.m. Eastern Time we're going to connect these signals and go through how to distinguish temporary market mechanics from genuine changes in the economic and credit cycles that you really need to to pay close attention to how do you sort signal from noise in September perfect example of how badly markets can be misinterpreted when the fundamentals are ignored so join me next Thursday October 29th 6 p.m. Eastern Time there's a link in the description to sign up it is free to attend it will be well worth your time to do so and part of that is recognizing how this AI anxiety isn't developing in a benign environment the context matters we're talking about and have been talking about the credit cycle downturn along with macroeconomic hardship basically everywhere you look outside the mainstream media and central bank discussions. Triple C rated junk bond spreads blew through the first real critical threshold of 1300 basis points, meaning investors are now demanding roughly 1300 percentage points of additional yield over government debt to hold some of the weakest corporate borrowers. That is getting to be an enormous risk premium, one that you need to pay attention to because the triple hooks, the triple C borrowers, they tend to have the highest leverage, the weakest cash flow, poor interest coverage, and limited refinancing options. They're often smaller businesses exposed directly to consumer demand and financing conditions there's always a correction and elevated energy costs make their situation worse from both directions consumers have to spend more on gasoline and utilities and transportation which leaves them less for discretionary purchases like at the convenience store small companies then face higher operating costs at the same time their customers are becoming let's use the term more uncertain the company's expenses go up its revenue becomes less reliable, its existing floating rate debt remains expensive, refinancing becomes more difficult, and that increases the default risk and liquidity risk across the entire credit system. And it also threatens the AI adoption assumption. Businesses struggling with higher energy bills, weak sales, and expensive debt aren't going to adopt every new AI service simply because a technology company built that capacity. They're going to delay projects, they're going to demand lower prices, and cut their experimental budgets. The infrastructure may exist without enough profitable demand to use it across the entire AI spectrum. And that is how macroeconomic weakness feeds back into the AI cycle. Slower adoption weakens cash flow. Weaker cash flow raises credit risk. Higher credit risk increases financing costs, higher financing costs, forced spending cuts, which slows the adoption even further. Not only that, but making it more expensive for that adoption to take place. In the first place, the weakest borrowers showed the problem first, which is why we pay attention to the triple hooks first. Private credit exposed the first stage. Junk bonds and leveraged loans made it public. Now the concerns are reaching the supposedly safest companies at the center of the AI boom. And then you've got the rest of the treasury markets, something like tips as well as swaps, which just reinforce all the signals that we're getting from the treasury curve itself. The compelling message isn't just that the 10-year treasury is below 5% today. The compelling message is the flattening yield curve set alongside negative swap spreads that are becoming more negative again, as well as tips break-evens that don't signal the hint of accelerating inflation anywhere. If markets believed higher energy prices were about to produce a sustained generalized inflation cycle, inflation break-evens should be surging. And they just aren't. The market is treating energy as a tax on consumers and businesses rather than the beginning of a 1970 style wage price spiral that by the way as i've shown you repeatedly is literally impossible in these conditions the difference in income growth and potential is just night and day different and that's exactly what break-evens are telling you swap markets are making the same distinction that we see between bills and terms so far there is room for another fed mistake in the short run officials might hike again or threaten to keep policy elevated because backward-looking inflation reports remain uncomfortable. But farther out, swaps continue to point toward lower rates. The market hasn't abandoned the expectation of economic deterioration. It merely inserted the possibility of one more policy detour before the credit cycle forces the ultimate destination. And that explains the renewed treasury buying. Investors aren't buying long-term government bonds because they believe the economy is strong. They're buying them because AI credit risk, triple C deterioration, weak consumers, flat beverage, downgrade in retail sales and what that showed, all of those things increased the probability of a more serious downturn. The inevitable AI correction was always going to hit once the marketplace started doubting the assumptions that fueled the initial boom to begin with. Nothing different here. It's the same as we see in every technology cycle. Equity investors can debate total addressable markets decades into the future. Bond investors, they need to know who will make the next interest payment, who will be able to make the next interest payment. That's why the debt markets are in charge. So Meta's earnings reaction wasn't just another bad day for technology stocks. It goes much deeper than that, especially given the broader context or context. Negative free cash flow, $825 billion in future commitments and record credit protection changed the conversation from AI excitement to AI financing, the nuts and bolts. And that concern spread across hyperscaler credit, SoftBank, CoreWeave, even global stock markets. At the same time, money moved toward treasuries again, pushing the two-year yield to around 4.69 and a 10-year below 5%. September's treasury market technical pressure is fading. The curve is flattening. Tips aren't validating sustained inflation. Swaps in terms so far are hedging the economic consequences of another possible Fed mistake. And triple C spreads tell us the broader credit downturn is already well underway. So the question is really not about artificial intelligence and the technology. It's about whether the debt finance companies that are building it will be able to remain liquid enough long enough into the future for that promised revenue to arrive. And today, the bond markets increasingly question those prospects. There's always a correction.