Subscribe
Share
Share
Embed
The AI Supercycle Podcast from QF Markets provides independent capital intelligence for the AI industrial economy. From semiconductors, AI factories and data centres to power grids, energy markets, critical materials, orbital compute and embodied intelligence, we follow where capital is being deployed, where the binding constraints are emerging, and what it costs to build through them.
Each weekly episode connects the latest developments to the multi-year forces behind the AI build-out, and sets out a clear base case along with the evidence that would change it. For traders, investors and anyone trying to understand where the capital in this cycle is really going. Published weekly by QF Markets.
[0:10] Hello and welcome to the AI Supercycle Podcast, where we follow the capital
[0:14] flows and physical constraints shaping the largest industrial build-out of our generation. From semiconductors, AI factories and hyperscale data centers, to energy markets and power grids, critical materials, orbital compute, and embodied intelligence. We track where the capital is being deployed, where the binding constraints are emerging, and what it means for traders, investors, and the broader economy. Today is Wednesday the 29th of July. I'm Tim Hardwick, founder of Quantumfields Market Intelligence.
Capital Formation Emerges
[0:45] We are now six episodes into this series, and across those six episodes, we have worked through what I think of as the physical stack of the AI supercycle. The chips that do the computing, the power that runs them, the materials that build them, and last time, the photonics that increasingly connect them, along with the possibility of taking some of this infrastructure into orbit. Each episode has been about a different layer of what I call the AI continuum, from the mine to the antenna, from geological time to orbital mechanics.
[1:18] This week, we add something that connects all of those layers together. It's not really a new layer. It's more like the nervous system running through every layer we've already discussed. Because before a single chip gets ordered, before a single watt of electricity gets contracted, before a single ton of copper gets mined, somebody has to raise the money. And right now, how that money gets raised is starting to tell us something the rest of the stack hasn't shown us yet. I flagged this one a long way back, right in episode 1. A constraint sitting above all of the physical ones we've covered since. Capital formation. The money has to arrive before any of this gets built. And I said back then that this deserved its own episode, eventually.
The Financing Layer
[2:05] And that episode is this one. And it's good timing, because over the past few weeks, the capital formation story has stopped being a background consideration, and started generating its own price action. So let's start with where the number stands. It has moved even since I first raised this on the show six weeks ago. The largest hyperscalers such as Microsoft, Alphabet, Amazon, Meta, and Oracle are now guiding toward a combined capital expenditure approaching $805 billion in 2026 alone, and climbing towards $1.1 trillion in 2027 as the AI data center build-out accelerates.
[2:47] And this isn't a static number, it's moving in real time, as earnings season plays out. Alphabet reported last week and raised its own 2026 capex guidance again, to somewhere between 195 and 205 billion, and that's up 15 billion from where it stood just three months ago. Google cloud revenue grew 82% in the quarter, comfortably ahead of what analysts were expecting. But Alphabet also reported negative free cash flow for the first time in history, burning around $6 billion in the quarter, even as it kept raising the amount it plans to spend. Microsoft, Amazon and Meta report this week, and the market is watching for exactly the same signal. Strong demand and even larger capex forward guidance. So let's think about that for a moment. One year of spending from five companies is larger than the annual infrastructure investment of the entire European Union. And this is not a forecast. It is disclosed guidance from management teams who have, so far, largely delivered on what they said they would spend. So here's the question this episode is built around. Can the world's capital markets actually absorb this? And if they cannot absorb it comfortably, what happens to the constraint architecture we have spent six episodes building. Before we get into today's numbers, it's worth taking a look at a comparison.
[4:16] Because we've actually seen a spending program like this before. Not at this scale, but close enough in shape to be really useful.
Telecom Bubble Comparison
[4:24] At the peak of the fiber optic telecoms boom back in 2000, US telecoms companies were investing around 150 billion a year building out fiber infrastructure. That cycle is the closest thing history offers to what we're watching now. Long duration, physical infrastructure, built ahead of confirmed demand, funded increasingly through debt. 150 billion a year. That was widely described afterwards as unsustainable, and it was about a fifth of the scale of what today's AI build-out is running at in a single year, from just five companies. For context on where this goes if the current trajectory holds, JP Morgan's own estimate for cumulative AI-related infrastructure spending through the end of the decade runs at 5.5 trillion dollars. This kind of comparison is exactly the sort of thing our research looks at systematically. We maintain a database of historical analogues, spending cycles, credit cycles, infrastructure booms going back a couple of hundred years that we use to stress test exactly this kind of question. Now I'm going to be a little bit careful here because the comparison cuts both ways. The demand behind the AI build-out looks considerably more real than speculative fibre capacity ever was.
[5:42] Token usage keeps climbing. Enterprise adoption keeps climbing too. And the companies funding this cycle, such as Microsoft, Amazon, Alphabet, Meta and Oracle, have balance sheets that most telecom companies of that era simply didn't have. So this is not a prediction that the ending looks the same. But scale is scale. And when a spending program runs at multiple times the size of the last comparable cycle, in inflation-adjusted terms, the financing structure supporting it stops being a footnote. It becomes one of those things that we need to really pay attention to and do more analysis on. That's really the whole argument of this episode.
[6:25] Here's the shift that matters the most, and it's easiest to see as three stages. Stage one is internally funded expansion. For the first couple of years of this build-out, the hyperscalers largely paid for it themselves. Free cash flow, retained earnings, balance sheets so strong that external financing was optional rather than necessary. Low cost of capital, high risk appetite and a market narrative built entirely around technological promise rather than how any of it got paid for. That was a really unusual position for an industrial build out of this size to be sitting in, and it didn't last.
Debt Markets Take Over
[7:04] Stage two is external financing acceleration, and this is where the story has been sitting through most of this year. Internal cash flow stops being sufficient to fund the pace of investment the management team wants. So companies start going to the debt markets, issuing equity, and drawing in sovereign and infrastructure capital at scale. That shift on its own isn't a warning sign. It can just as easily mean management sees an opportunity large enough to justify borrowing for it. But it is a structural change worth noting, because it changes how sensitive the whole build-out becomes to the price and availability of external capital. And the evidence that we're deep into that second stage and starting to see the edges of a third is everywhere in the data right now.
[7:52] In the first week of July alone, Amazon went to the bond market seeking at least $25 billion.
[7:59] NVIDIA priced its own $25 billion bond offering in June. That was its first return to the debt market in five years, with tranches stretching as far out as 2056. SpaceX followed its record initial public offering with a $25 billion bond sale of its home within days. And Meta, Oracle and Alphabet have been issuing at a pace that would have been unthinkable for this group of companies just three years ago. Add it all up and the picture is stark. The US high-yield bond market alone absorbed nearly $32 billion of new AI-related debt in just the first eight days of July.
[8:39] Investment-grade issuance from hyperscalers and AI-adjacent names has already blown past $200 billion this year. That's more than double last year's full-year total. One estimate now puts total AI-related debt issuance for this year at around $570 billion globally. And JPMorgan's own number for debt financing through the rest of the decade runs to 4.1 trillion. This is the top of the capital stack we've been referencing throughout this series. The tier that sits above semiconductors, above power and above materials. Before any of the physical layer gets built, it has to be funded. And the funding mix has just changed character. From internal cash generation to external credit markets in the space of about a year.
[9:27] Now this is where it gets really interesting and where I think most coverage of the story is missing the point. The headline number, 500 billion, 700 billion, whatever figure you want to use, is less important than what is happening to demand for that debt. A quick explain for anyone who doesn't live in bond markets day to day, when a company sells a bond, its bank asks investors how much they'd like to buy. If a company wants to raise 10 billion dollars and investors put in orders for, say, 50 billion, that deal is covered five times over. It's the same idea as a house going up for sale and getting five offers above the asking price. Lots of demand, and the seller can afford to be picky or push the price up.
[10:12] That number, how many times a deal is covered, is what people in this world call the coverage ratio. And it's one of the simplest, fastest ways to see whether investors are still enthusiastic about a particular kind of debt well before that enthusiasm or the lack of it shows up anywhere else.
Demand Weakens for Bonds
[10:29] Back in February, orders for hyperscaler bonds were covering the amount on offer by close to five times, the equivalent of that house getting five strong offers. Investors were queuing up. By July, that coverage ratio had fallen to below two times, more like one modest offer above asking, maybe two. That is a really large move in a short period, and it is happening specifically in paper tied to this build-out, while broader investment-grade coverage, meaning the same measure applied to other highly-rated corporate borrowers generally, has barely moved. What this tells you is not that the market is closed, 570 billion of debt is still finding buyers.
[11:11] What it tells you is that the market is starting to ask for a higher price to keep buying it. Now that price shows up as something called the spread, which is really just the extra interest a company has to pay above the safest possible benchmark, normally government bonds, to compensate investors for the risk of lending to them instead. A wider spread means investors are demanding more compensation. A narrower spread means they're comfortable lending cheaply. So when I talk about spreads widening through the rest of this episode, that's what it means in practice. Borrowing gets a bit more expensive for whoever's doing it. Wider spreads, more scrutiny. The easy phase of financing this build-out, where every deal was oversubscribed and pricing kept getting tighter, appears to be behind us. Think of it as a third stage layered on top of the two I've just described. Stage two was simply more issuers coming to the market. The third stage is what happens once enough issuers and enough capital converge on the same theme at the same time. Sovereign wealth, infrastructure funds, private credit, public bond and equity markets, all leaning into the same story at the same time, all leaning into the same story simultaneously. When that happens, the question quietly shifts. It stops being, can this get financed? And starts being, is too much money chasing the same trade at once?
[12:36] A falling coverage ratio, even from a very high starting point, is one of the clearest early tells that a market has moved from the first question to the second. Okay, so let's be specific about what this is and is not. It's not a credit freeze. It's not the debt markets closing the door on the AI build-out. It is the earliest form of price discovery in the financing itself, and it's worth watching closely because it moves faster than anything else in this story. A grid connection queue takes years to clear. A bond coverage ratio can shift in a single quarter. There is a second layer to this that deserves attention, because it is less visible than a public bond deal, and that is exactly why it matters.
Off-Balance-Sheet Leverage
[13:20] A meaningful and growing share of this build-out is now being financed through structures that keep the debt off the sponsoring company's own balance sheet. The mechanism is a special-purpose vehicle. A separate legal entity is set up to own and build a specific data centre campus. Private credit firms and institutional investors put in the debt and often some of the equity. The hyperscaler then leases the finished facility back and records the payment as rent rather than as interest on debt. Here's a simple way to picture it. Imagine you want a new extension on your house, but you'd rather your own mortgage statement didn't show that extra debt. So instead, you set up a little company that borrows the money, builds the extension, and then you personally just pay that company rent to use it. Your own mortgage looks exactly the same as before, but you're still on the hook, indirectly, for the extension either way. That's essentially what's happening here, just at a scale of tens of billions of dollars, and with the data centres instead of house extensions. On the balance sheet, the leverage looks like it has vanished, The obligation has not.
[14:31] META's financing for its Louisiana data center complex, arranged with Morgan Stanley and Blue Owl Capital, ran to roughly $30 billion structured this way. A separate META vehicle for a Texas facility in El Paso raised a further $13 billion on similar terms. And this is now a pattern across the industry, not a one-off. Morgan Stanley itself estimates the technology sector will need something in the region of $800 billion of this kind of off-balance sheet financing by 2028. Now, who is actually holding this risk matters a great deal, and it's worth naming names. Private credit funds such as Blackstone, Blue Owl, Apollo and PIMCO have gone from near zero exposure to AI infrastructure lending to more than $200 billion outstanding in the space of just a few years. This is a different investor base from the one holding a Microsoft or an NVIDIA bond. It's less regulated than a bank, less liquid than a public bond, and increasingly sits underneath the retail-adjacent capital that has been flowing into private credit funds more broadly.
[15:39] That's not automatically dangerous, but it is a different risk profile, and a problem in that layer would not necessarily show up first in the places most investors are watching. The Bank for International Settlements flagged exactly this in its annual report, published in June, and they didn't mince their words. Their central warning is that the AI build-out now ranks amongst the largest investment booms in US history, and that its heavy reliance on debt, plus something called circular financing, raises real questions about how sustainable it is.
Circular Financing Risks
[16:12] So what is circular financing? There's a simple version of the idea, and then there's the messier version the BIS is actually worried about. The simple version first. A chipmaker invests in a customer, and that customer uses the money to buy the chipmaker's own chips. Money goes out, and a large chunk of it comes straight back in as revenue. That much has been true for a while, and it doesn't necessarily mean anything's wrong. The messier version is what the BIS is really flagging. These deals increasingly blend equity stakes, debt and long-term supply contracts all together. And the terms are often never made public. That opacity creates a specific concrete risk. The same underlying asset, say a claim on future chip supply or data center capacity, can end up being pledged as collateral more than once in more than one of these deals, without anyone outside the companies involved being able to see what is happening. So if demand ever disappoints, the BIS's concern isn't just that one company takes a loss, it's that nobody quite knows how many of these deals that loss would ripple through because the web connecting them was never fully visible in the first place.
[17:26] Their own words on this were blunt. The race to commit early through debt and circular finances makes a downturn, if one comes, more likely than less. And just this week, we got the clearest live example yet of exactly what the BIS means by that.
[17:43] NVIDIA is now lining up something in the region of $750 billion in commitments across two relationships that fit exactly that description. Roughly $500 billion of that tied to a partnership with SK Group, the parent of memory chipmaker SK Hynix, covering NVIDIA's own purchases of memory alongside SK's purchases of NVIDIA hardware. And separately, NVIDIA is in discussions to guarantee as much as $250 billion of OpenAI's data center lease payments, on top of financing OpenAI's purchases of NVIDIA chips. None of that means the underlying demand for chips is fake, but it does mean less risk diversification is happening than the headline figures suggest, because the same balance sheet keeps sitting on both sides of these deals. And this is exactly the story that's been rattling markets over the past few days. It's a live, developing narrative right now. We've seen real position unwinding in semiconductor stocks this week, running alongside the investor anxiety about the sheer scale of hyperscalar capex we covered earlier in this episode. Put simply, when investors start worrying about circular financing, at the same time they're already nervous about capex levels, sentiment in the financing layer can shift very quickly indeed, well before anything actually changes in the underlying demand for AI compute.
[19:05] Now I'm raising this not to be alarmist. The physical demand behind this build-out, and we've made this point repeatedly in the previous episodes, is very real and in a way that previous speculative infrastructure cycles were not. But the financing structure is a very distinct risk from the demand question, and the two need to be assessed separately. You can be right that the world needs more compute and still be exposed if the vehicle funding a specific data centre runs into trouble. Bond markets are one half of this equation. The private credit and SPV structures we've just covered are the second half.
Private Funding Boom
[19:39] There's a third piece worth separating out clearly because it is a different kind of capital again.
[19:47] Private equity and venture funding, the money going directly into the AI model companies themselves, rather than into the data centres and the infrastructure around them. And this has been moving even faster than either of the other two. OpenAI's most recent funding round took the company to a valuation of roughly $750 billion, raising around $110 billion in the process. Now Anthropic closed a series H in the tens of billions that pushed its own valuation toward the region of 1 trillion, briefly making it the most valuable private company anywhere on earth. XAI raised $20 billion in a single round. And the pace of these rounds is itself a signal worth analysing in more detail.
[20:30] When a company moves from one funding round to the next in a matter of months, each time at a higher valuation, that is not primarily a statement that the business has gotten dramatically better in that window. What it is is a statement that its capital needs are compounding faster than even its last round anticipated. Almost always because compute costs are scaling with model ambition faster than revenue can keep up.
[20:57] Now, layer the IPO story on top of this. Going public is really just a company moving from this private funding world into a public equity market, letting ordinary investors buy shares for the first time. SpaceX completed its initial public offering earlier this year, raising in the region of $86 billion in what became the largest public listing in history, before promptly returning to the bond market for a further $25 billion just days later. And the market is watching closely for what comes next. Both OpenAI and Anthropic are now understood to be weighing public listings of their own before the year is out. If either proceeds, it would represent the single largest test yet of whether public equity markets can absorb a company of this scale and this capital intensity in one listing. Every one of these events, the funding round, the IPO, the bond sale that follows days later, is what I've referred to in this framework as a corporate action. It is the ecosystem reorganizing itself in real time, and each one tells you something about where the capital believes the value of this build-out actually sits, and often this is well before that belief shows up in anyone's revenue figures.
Capital as the Constraint
[22:09] So where does this leave the constraint framework we've built across this series? Memory, power and materials are constraints because physical supply cannot expand fast enough to meet demand. Capital formation is a different kind of constraint because in absolute terms, global capital markets are not short of money. What can run short is willingness, the price at which that capital is available and the discipline with which it gets allocated. This is precisely the divergence the bond coverage ratios are starting to show you. Not, the money is gone, but rather, the money now wants a better price and is starting to ask harder questions about where it's going. When investment grade spreads for this specific cohort of issuers widens meaningfully beyond their recent range, that is the earliest point at which discretionary capital expenditure gets delayed. Not cancelled, just delayed. And a delay at the very top of the stack works its way down through every physical layer beneath it on a lag of somewhere between six months and two years depending on how far along the project already is. That is why I've argued that capital formation deserves to sit alongside memory, power and materials in this framework even though it behaves differently from all three.
[23:26] It is the fastest moving layer in the entire stack. A bond spread can move in an afternoon, a venture round can close in a week, and because it sits at the very top of the chain, above compute, above physical infrastructure, above the energy that powers all of it, it has the capacity to reprice or delay everything underneath it faster than any of the constraints we've covered so far. There is also a feedback loop that we need to understand here, because it is what eventually closes this whole story. Strong returns on the capital already deployed make the next round of financing easier and cheaper to raise. Weak or disappointing returns do the opposite. They raise the risk premium investors' demand, which makes capital more expensive and more selective, which slows down the pace of investment. Nobody can point to a clean, ecosystem-wide answer on returns yet. Most companies still don't disaggregate AI revenue clearly enough for outsiders to judge return on invested capital with any precision. But that question, where the money is already spent, is earning its key. It's the one that ultimately governs whether this becomes a multi-year financing headwind, or just a normal, temporary rise in what it costs this cohort to borrow.
[24:40] It just hasn't started driving the price of capital yet. When it does, it will move faster than anything else in this episode.
Fed Risk Returns
[24:48] And you know, none of this is happening in a vacuum, and I'd be doing a disservice if I didn't place this against what's going on with rates. A quick reminder for anyone less familiar with this part of the picture, when the Federal Reserve sets its policy rate, that rate works its way through to the interest cost on pretty much every other kind of borrowing in the economy, including the bonds we've talked about in this episode.
[25:11] Higher policy rates generally mean more expensive borrowing everywhere else too, so what the Fed does next matters directly to everything in this episode. We now have a new Federal Reserve Chair, Kevin Walsh, who took over in May, replacing Jerome Powell, after a really contentious confirmation process. Going in, many assumed he'd lean towards cutting rates, given the administration's public pressure for lower rates throughout the last year. But Walsh has been explicit that he'll follow the inflation data rather than commit to a path in advance, and the data has not cooperated. Oil prices have pushed back above $100 a barrel amid the tension with Iran, and inflation has proven stickier than expected through the middle of this year. As a result, this week's meeting, which is happening as I record this, has turned into a genuine coin toss. Most economists still expect a hold, but the market is now pricing a rising chance of a hike rather than a cut. With that probability climbing further at the September meeting and beyond. That is close to the opposite of what many expected from this Fed chair just a few months ago. So here's the twist, and it's something we need to pay attention to. Normally a Fed chair would explain their thinking to the market well in advance, but.
[26:26] Through speeches and testimony, so investors and banks can build a clear picture of how they'll react to new data. Walsh has deliberately not done that. He has said plainly that he intends to give less forward guidance than his predecessors, not more. So the market is having to price this meeting almost blind, without the usual roadmap. That matters directly for everything in this episode. If the Fed ends up hiking, or simply signals that hikes are more likely than cuts from here, borrowing costs across the entire financing structure, the bonds, the private credit, the project finance, get more expensive at exactly the moment coverage ratios are already compressing. Two separate pressures on the same financing layer, arriving at the same time, would be a really different picture from the one I've built through the rest of this, what we've been talking about. And because Walsh isn't giving the market much warning either way, that shift, if it comes, is more likely to arrive as a surprise than as something priced in gradually. So that's worth having on your radar alongside everything else that we've been talking about today.
The Skeptics Case
[27:31] So let me put the other side of this story to you, because there is the sceptics case. And the sceptics argument runs like this. Coverage ratios falling from five times to under two times still means the debt is being placed. Yep, absolutely. There is no failed auction here, no deal being pulled. Investment-grade spreads for these issuers remains historically tight in absolute terms, even after the recent widening. The companies doing this borrowing, Microsoft, Alphabet, Amazon, Meta, NVIDIA and so on, are among the most credit-worthy borrowers on the planet, generating operating cash flow that dwarfs almost any other sector. Choosing to borrow at 25-60 basis points over treasuries, and that's 0.25% to 0.6%, which is a tiny premium, when your own cost of capital is far higher, is simply good capital discipline. It's not distress. And the private credit and SPV structures, while less transparent, are not new inventions built for this cycle.
[28:32] Project finances funded power plants, pipelines, and toll roads for decades using exactly this logic.
[28:39] Lend against the contracted cash flow of the specific asset, not the sponsor's entire balance sheet. Institutional investors, pension funds, and insurers managing tens of trillions of dollars globally are actively looking for exactly this kind of long-duration, contracted, essential service asset. Data centres leased to investment-grade hyperscalers on 10-15 year terms fit that description about as well as anything currently available in the market. And this is a fair and reasonable case, and I don't dismiss it. But here's why I still put more weight on the coverage ratio than on the spread level. Spreads are typically the last thing to move in a financing cycle, not the first. Issuers and their banks generally hold the price stable for as long as they can, even as demand cools, because a widely publicised spread jump is itself a signal that damages the next deal down the line.
[29:33] Coverage ratios is what moves first, quietly, in the order book, well before anyone repriced anything. So the fact that spreads are still tight doesn't tell you the pressure isn't building. It mostly tells you it hasn't been forced into the open yet.
Base Case Outlook
[29:49] My own read on this is narrower than either extreme. I'm not calling this as a bubble and I'm not calling it a crisis. What I am saying is that the financing layer of this build-out has quietly become the fastest-moving and least-monitored part of the entire stack and that a practitioner who is only watching semiconductor earnings and power purchase agreements is missing the signal that, historically, moves first.
[30:15] My base case is that we are currently sitting midway through the second stage I described earlier, external financing acceleration. The coverage ratio compression and the SPV growth are both early evidence of a third stage, capital crowding, and that's starting to build underneath it. I don't think we've crossed fully into that third stage yet, but the direction of travel, from a market where every deal was oversubscribed by five times to one struggling to clear two times in the space of five months, is not ambiguous. On that basis, I expect capital formation to become a genuine, if intermittent, constraint on the pace of this build-out over the next 12 to 18 months, without becoming the kind of constraint that halts it.
[30:58] I expect bond coverage ratios for this cohort of issuers to remain choppy, rather than snapping back cleanly to where they were in February. I expect spreads to drift wider before they stabilise, and I expect at least one further round of hyperscaler issuance to be read by the market as opportunistic rather than urgent, which will itself become a signal worth paying attention to. I expect the off-balance sheet and private credit share of total financing to keep growing, not shrinking, because the appetite from institutional investors for this kind of asset class remains genuine, even as the structures used to access it draw more scrutiny from regulators and bodies like BIS. And I expect the OpenAI and Anthropic listing decisions whenever they land to be treated by the market as the definitive test of whether public equity investors will extend the same patience to this build-out that private and sovereign capital has shown so far.
[31:53] The stage after this one, where the market's focus shifts from what is being built to what is being earned, is not here yet, but it is one to watch for, because that transition has historically been sudden rather than gradual, and it is the point at which the feedback loop I described earlier actually starts setting the price of capital in real time.
[32:13] The principal risk to this view is a sharper repricing that I'm currently assuming. A really disappointing quarter from one of the large issuers, combined with a broader at risk of moving the credit markets, could turn a gradual widening into something closer to a proper financing window closing. For anyone tracking this alongside me, here's the practical takeaway, purely from an analytical standpoint. The framework itself does tell you where to look. If that sharper repricing happens, the lag between a formation stage shock and its effect on the physical layers beneath it, memory, power, materials, is the thing that we need to watch closely. That lag is where this constraint's real signal shows up first, well before it reaches the physical layers of the build-out that we've spent all of this series covering.
Nervous System of AI
[33:00] So that's the nervous system running through everything we've covered so far in this series. Memory, power, materials, photonics, and even orbital. And now capital formation sitting above all of them, moving faster than any of them, and arguably the least well understood by the market as a whole. You can find all of our work at qfmi.substack.com and that will be moving to a wordpress site very very soon.
[33:24] The Market Pulse and In The Spotlight articles are free and always will be. The Weekly Outlook, the Weekend Debrief, and the Monthly Strategic Report sit behind the paid subscription. Subscribe and you get the full picture. But before I let you go, one important point. Everything in this podcast, including everything we've heard today, is provided for educational and informational purposes only. It is analysis, not personal financial advice, and nothing in this episode should be taken as a recommendation to buy or sell any security. Always do your own research and speak to a qualified financial advisor before making any investment decisions thank you for listening to the ai super cycle podcast see you next week.