The Next Phase of the Market: What Iran, AI, and Crypto Are Signaling and What to Buy

Hello and welcome to Being Exponential. All right, so Luke, we spent the past few weeks trying to figure out what's going to wake this market up, but suddenly we've got a war pushing oil higher, big tech borrowing at unprecedented amounts of money to build AI, Amazon making a huge move into custom silicon, and Bitcoin coming back above 80,000. Briefly, are these separate stories or are they all telling us something about where the next phase of the market is headed? I'd love to start the topic with the Iran war. Uh, are we thinking it's gonna oil is going to rise above 90 to $100 a barrel? Do we think this is going to continue even further or uh are we expecting the taco to finally come in? >> Oh man, I mean the taco's got to come in, right? Like this is this is an ugly situation. Saudi Arabia has now shut down some energy facilities in the south because of attacks. Um, this is starting to feel like March, right? Like this is not what we had in May, June, July where it was pretty calm and missiles weren't flying and diplomacy talks were happening and we were flying diplomats out there and we're meeting and you know like that's not what's happening right now. What's happening right now is the ante keeps getting up again and we're getting active strikes and and that changes the game. I mean, look at oil. Oil, you know, this is what happened when we were we were calming May, June, July, and then all of a sudden, whoa, stuff's picking up. Stuff's picking up. We retook the 200 day moving average on oil. We retook the 50-day moving average on oil. We retook the 100 day moving average on oil. We're getting close to taking out this big spike from July 23rd, 9219. We're at 9213. We actually did take it out this afternoon. We're up towards 93. So, we're going to take that out. So now we're back where we were when the missiles were flying. We're in the March April regime again. 92 oil. That's a I mean, excuse my language. That's a situation for the economy. Okay, that that's a bad situation for what's going for where we are in the economy right now. We just got the NFIB small business survey. And you know what that survey said? Inflation is sticky. Inflation is hot. It's hurting small businesses. ISM manufacturing, same damn thing. ISM services, same damn thing. every economic data point you look at right now is and these are August prints which is captures some of this spike but not the big you know kind of final portion here where we went from you know 83 up to 93 and potentially kind of touching that $100 level pretty soon. um inflation is sticky, the economy is slowing uh and it's it it's a tough situation for the market to really do well. So that that is a major overhang. But to your point, that's this is where taco happens, right? This is where we do back down. This is where the escalation is one-sided here. Okay. Iran doesn't want to escalate. They're retaliating. We're the escalatory force. America's escalatory force here. As soon as America deescalates, this this is over. Like this this goes bye-bye and we go from, you know, this big collapse 110 down to 70 on oil. That that could happen over the next few months if we pull out. Now, I think we're going to and I think that's exactly what is going to happen. I mean, this is political suicide going into the midterms to continue to push the agenda here. And so, you know, with midterms around the corner, maybe this is kind of like the last few weeks of all right, push, push, push, push, push. Let's see how far we can go. Let's see how much we can extract. Let's see how favorable of a deal we can get. Really try and break their spirits and break their soul. And then boom, art of the deal. 11th hour, bam, we're here. We got a deal. It's the best deal that we could possibly get and we're out of there. So, I think that's what's happening. That's why despite me acknowledging that the current conditions are quite frankly crap, they're not going to persist for all that much longer. This is not a durable headwind against the markets. And that's why I think the markets are are pretty much hanging in there. Um, and it's why the 10-year Treasury yield is not breaking out. You know, the 10ear Treasury has been a very good barometer for kind of like how this war is quote unquote going, if you will. Now we're obviously moving higher. Look at that. Since the Iran war started, we were what, down here around the, you know, fours basically. Now we're at 4.8. So that that's a pretty big uptrend. Look at the MAS. There's there's a clear trend there that we're going up and to the right. But um we're kind of we're kind of basing here. We're kind of basing at 4.8, which is a major resistance right here. 4.8. That was the high from uh January of 2025, right before the liberation day stuff happened and you know, things went differently. Um that that was that was the level that we stopped at. 4.8 was kind of the level that we stopped at here in October of 23 until we had this big spike up to five and then got stopped out at five. But the 4.8 to 5.0 5.0 level on the 10-year. That's resistance. That's major resistance there. The fact that we are again being resisted at that resistance tells me the market's not all that worried about this turning into something economically catastrophic. And that my theory of this is kind of like that last minute 11th hour push before the 11th hour deal before the midterms is the prevailing consensus view on Wall Street and the most likely path forward. So I think things will be okay over there. Now having said that the aftershocks of this are such that the AI bifurcation will continue right because oil and high inflation and the Fed and all that stuff it's hitting consumer businesses but it's not impacting hyperscaler spending. It's not impacting development of new AI models. that's not impacting the purchasing of memory, of storage, of networking, of optics. That stuff has continued. And so I think even after the Iran war stuff kind of fades a little bit, you're still going to see this AI bifurcation trade that has been the trade for three plus years. I think it'll rem actually almost four years now, right? Chat launched almost four years ago. Wow. Um I think that's going to continue for uh for the foreseeable future. >> Yeah, understood. It it really feels like this left hand right hand thing. You got Trump pushing uh Wars to cut rates and then Bentons buying uh buying back the bonds, but we're we're still at war. Do you think there could be any uh potential uh geopolitical blowback because of how long this has endured that could affect the market in the future? >> Um the the political blowback is the the anti-data center legislation push that we've talked about. That is the blowback, right? that that is a direct response to um what's happening right now. Whether we talk about the and it's all it's all related whether we talk about the Iran war, we talk what's going on in Japan or we talk about uh just the AI buildout in general. The anti-data center push is a direct consequence of direct response to current economic conditions propping up the 1% and dragging down the 99%. Right? That's that's what that is a response to. So that is the backlash and again that backlash is not strong enough or widespread enough to do much of anything by the midterms but it will probably be big enough widespread enough to do a lot by the 2028 election cycle which again that I keep coming back to that that is the the real end date that I'd be worried about as a bull not November 26. All right, Luke, moving on to the next topic here. We want to talk about Chad GPT Astra. Uh earlier uh earlier this week, you know, you mentioned Jensen Wong saying this is probably the first real step into AGI. Uh Chad GPT themselves scored uh said Astra scored 99.9% on ARC AGI 3. Is this the true next step in the Frontier Lab Evolution? Is that what we're looking at? I watched the essentially trailer commercial for it and it it looked crazy. Yeah, I mean they they're trying to make it look like you're talking to Jarvis, [laughter] right? That's what they're all trying to do. Um, and yeah, it's cool. It's awesome. I'm going to push back against the AGI claims. I don't think the data supports it's a broad leap in general intelligence, but it does support that it is a genuine meaningful significant two-step change um in two specific areas and that's cyber security and long context reasoning. Um when you kind of look at the data on it, Astra hit 100% 78.5% from uh its predecessor and became the first model to cross OpenAI's critical cyber capability threshold which is good enough to independently find and exploit zero day vulnerabilities. It actually found two real ones during testing. So like that that to me is that's big. Those numbers are big. Those are big numbers. They're real numbers. um they've been backed by uh independent third-party tests. The cyber jump here is very real. Meanwhile, on long document retrieval, it hit 96 to 100% accuracy at up to 100 up to 1 million tokens, sorry, which is a really big jump for anything that's document heavy. Uh coding is also meaningfully more efficient by by a few metrics. comparable output using a third of the tokens of it of its predecessors is sort of what the most models out there or most tests on it have come up with. So the data supports that this is a really good model that is significantly better in a few areas but generally speaking is pretty much as capable as like you know Fable and some of the other top models that are out there right now. Now what's the read through here? Well, Jensen calls it AGI. We're saying it's not AGI. I don't think that really matters. What matters here is that we have seen new models shipped every few weeks. It feels like maybe the cadence has slowed to about every month now. And each model that is shipped whether it's from Anthropic, whether it's from uh XAI, SpaceX, whether it's from um Open AAI, Google, each model is a noticeable improvement from the previous leading model. Sometimes it's in one specific area, sometimes it's in two, sometimes it's in general intelligence, but each time there is there is a step forward. We are continuing on this journey of the models at a very quick rate getting more capable, getting better and producing more value. And that is a huge positive for the AI trade because it means all this spending that is going on is worth it. And we're seeing that in the token pricing. So if you look at the H100, actually I'll pull up the chart. The Orin H100 compute price index. Now this is the cost per hour to rent an H100 Nvidia GPU. It's climbing. So this and this is an H100 which is three years old at this point. Three-year-old architectures on AI computing. the rental prices for them are continuing to climb and are close to hitting a new record high, which they briefly had here during that massive May supply constraint. That tells you that all these new models are not just getting better and better and better and better and better and more capable, but companies users of the AI models are finding more use cases for them and using them more. So, demand for the compute is going up. So, price is going up. Now this price is going up as supply of H100s is also going up dramatically. So that means demand is really really going up. And so all of this means that yes this multi- trillion dollar infrastructure buildout cycle, super cycle, whatever you want to call it, is absolutely worth it. Every dollar they are spending there is value coming from that. There is ROI on that spend. There is demand on the other side meeting it. And therefore the spending will continue and the core trade that benefits from that spending socks SMH AIS those things will continue to work really really well. When we look at the charts on those things they're gorgeous. AIS has become my new favorite. The Vista shares artificial intelligence super cycle ETF. Maybe because it has the name super cycle in it, but mostly because the constituents of the CTF are very very very attractive in terms of recipients of all that AI capex. This chart is as pretty as it gets. This is so damn pretty. You got a big winner that had a big crash below the 100 V recovery right above the 100. retook the 50, came back down to the 100, turned resistance into support, bounced back above it, retook the 50 again. This is higher lows and higher highs out of a big sell-off, continuing a long-term trend, uptrend. This is textbook pretty. This is a gorgeous rebound here. This is a rebound I would buy. This is the trade that works because of Astra. This is the trade that works because H100 prices are going up. This is the trade that works because the spending will continue. So that's why the Iran war stuff is important to to monitor. What's going on in Japan is important to monitor. Election cycle is important to monitor. But at the end of the day, the core fundamentals here remain intact. And that's why this trade will persist through that noise. And that's why on a day like today, AIS is up two bucks.50 50 cents while oil is is going to 9293. This this persists through it. It's durable beyond geopolitical headwinds. >> Excellent. So, just kind of diving into that topic deeper talking about the next wave or next phase of the AI trade. I want to dig back into a topic we talked in our stocks episode that Qualcomm plus Amazon deal. So, Amazon is set to receive about $4 billion of Qualcomm stock and that can grow much larger over the next 10 years or so. So, just continuing on AI becoming profitable. Uh, dive into uh Qualcomm for us, Luke. I like I said in the last episode, I love Qualcomm. I the custom silicon thing is is such a big movement within. So, here's the thing. When you have the AI super cycle, you can just buy AIS and chill and that's totally fine. Or you can dig into the sub themes within the super cycle where the money is rotating into like there's this big slosh trillions of dollars slloshing around going towards the capex super cycle where spec you could either just invest in the super cycle in one general broad ETF AIS or you can ask the question which specific sub the subverticals of the super cycle are winning the lion share of that spending which ones are having accelerating share of that market growing their their piece of the pie if you will custom Silicon is is one of those narratives. It's one of those themes. It's one of those trends. And I think it's absolutely where you want to be. Marll, Qualcomm, Broadcom, the networking names benefit from this because again, Nvidia kind of has a lot of in-house networking. And so when you do custom stuff, that means more Lumenum networking, that means more Broadcom switches, that means more Arisa networks, that means more lums all that stuff. So applied opto electronics, AOI on a nice rebound. talk about that stock a lot on this podcast. So I think that's that's where that's that's a key theme here that you want to be participating in and the Qualcomm Amazon news is just further validation of that. Now again like I said in the last podcast, Qualcomm I think is one of the more attractively valued plays here because it's like it's 17 times forward earnings for an a really attractive accelerating revenue growth narrative over the next few years. I that that's just a great great setup. I like the chart too. Um, but I think Marll looks really good. I think Broadcom is really good on the dip and the networking plays have been stuck but hot and so I think that they get even hotter and I that's also where I would be. AOI again I love AOI. I think it's a great play here. >> All right, love it. Okay, Luke. So uh another topic that is becoming uh something of a potentially bearish argument for the eye trade is how much the hyperscalers are becoming debt issuance companies. So Nvidia is practically one of the biggest lenders of uh capital in the in the market today. So what what is your take on that argument? Okay. So, the part of that argument that has legitimacy is liquidity because with all of these debt issuances, it's kind of crowding out treasuries and it's pushing up 10ear treasuries. There's an argument that because of all of this debt that the hyperscalers are issuing, people or institutional investors rather that were buying US treasuries are now going to are now buying hypers scale debt because it's basically a safe, right? If you think about the US economy and big tech are one and the same pretty much. If you look at the risk profile of the US economy versus the risk profile of the combined, you know, Meta, Amazon, Microsoft, Google, it's a similar risk profile. And so the argument goes institutional buyers of US treasuries are becoming institutional buyers of hyperscal debt at the expense of buying US treasuries. And so there's more selling pressure on US treasuries which pushes treasury yields treasury bond prices lower and treasury yields higher. So that to me is a legitimate piece of that argument because the higher that tenure goes, the more in trouble the market is from a valuation pressure perspective as well as a cost of the debt. Right? If a lot of this is debt financed, then it's based off of that 10-year Treasury yield. And that 10-year Treasury yield continues to go up, then the cost of debt for these companies goes up as well, and that can change the ROI economics of the spending. So, that's the legitimate piece of that argument. One of the reasons I'm not too worried about that is because I don't think that there's enough hypers scale debt out there to really crowd I mean the treasury market we're over 40 trillion in debt like that that's a number that is an order of magnitude larger than what we're talking about with the hyperscalers and the hyperscalers I think are I mean I don't know if they're going to issue that much more debt. I think that they're kind of issuing debt right now and then plan to fund additional capex with cap with cash flow growth because they are seeing, you know, Meta and Amazon and Google. I mean, these companies are growing 20% year-over-year on the top line with margin expansion. Uh, and so there's a lot of room for cash flow growth over the next several years. And so, I think we're we're nearing that trillion dollars per year in annual capex spend. That to me is kind of like a carrying level. And so I don't see much more debt necessary to fund additional spending and cash flow growth should be sufficient. And so I think the current situation we have with the hypers scale debt markets is kind of the situation we're going to have for the next few years. It's not going to get worse. If it doesn't get worse and you know the Treasury market remains an order of magnitude larger than the hypers scale debt market and the hypers scale debt market does not have a lot of growth coming into it which I think is what's going to be the case then the risk here seems I mean the terms we've been using like go back to last week captain contained right the risk here seems captain contained to me and that's why I'm not too I acknowledge the 10-year Treasury yield risk but I don't think that risk is capp and contained that that that's basically what I have to say about that I think it's capped and contained. Got it. So, just kind of two follow-up questions on that topic. Uh, it it looks like OpenAI and Anthropic are reportedly pursuing investment grade credit ratings around their IPOs. Uh, does that change the math on it? And is there going to be a point where that we reach where the biggest constraint on AI isn't, you know, GPUs or data centers, it's the amount of capital flowing through the system? Um the anthropic open AI news is it's a good headline and I don't think it means anything fundamentally. Um in terms of is the constraint here going to turn into capital? [sighs and gasps] I don't think so because based on the math that I've done well let's put it this way. I I think the the super big growth years for capex are over. Again, we're nearing that trillion dollar a year mark. PWC thinks that we're going to go to 1.8 trillion per year by 2050. So that means we go from a trillion in 2027 to 1.8 trillion in 2050, which is adding 800 billion incrementally. Basically, you know, a little bit less than a double over the course of 25ish years. That's not a lot of growth per year. That can be funded by cash flow alone. So that's why I'm saying I do not think the debt I think the debt issuance spike was this year and we will not see a debt issuance spike next year or in 28 or in 29 or in 30. I think the debt issuance spike happened this year. Cash flow growth should be sufficient to fund additional capex from here over the next several years. Assuming capex growth let's say 10 15% per year. I think the cumulative cash flow growth from these capex from these hyperscalers should exceed 10 to 15% per year and therefore it will be enough to fund additional capex. From that perspective I don't think the capital becomes a a big constraint. Now I could be wrong if there's some spending surge from here but such a spending surge would be probably really beneficial for the capex takers since the only reason they would surge spending from here is if they saw some big spike in demand because the you know maybe maybe a model does get to AGI and then we're like boom now everybody's doing it now there's this huge demand surge that seems like a good problem to have as opposed to a bad problem to have. So, I think it's possible, but if we do get there, um I think it's probably bullish. >> Excellent. Love to hear it. All right, so let's pivot to our last topic of the episode. Uh let's talk about crypto and Bitcoin. So, it briefly tested 82,000. Uh we've saw the highest inflow into Bitcoin ETF since January. Uh my question here, Luke, what is the number that we want to see before we kind of confirm that we're back on the crypto bull run? >> We are we are very close, man. We are super close to that number. To me, like I've said multiple times before. We took the 200 day moving average. That was bullish. The cleaner signal historically speaking when a bus cycle turns into a boom cycle again is the 50we retake. We are almost there. We are We are so darn close. Um right now we're at 786 to 50 weeks at 795. I know it looks like on the chart here we may have retaken it, but we didn't get an actual weekly close above where we needed the weekly close. Last week um the close was 79918 and the 50 week was 8228. Week before was 786 and 80.9 or call 81. So last week was the closest we got basically 80 and 82. So 250 bucks off, but we were still 250 bucks off. We did not retake the 50. And now the gap has widened just a little bit. 786 versus 796. So to me, I want a weekly close above the 50 week because every single time that has happened, that is the inflection point we got. Let's just look at the chart. Let's go all the way back to the beginning of time, aka the beginning of Bitcoin. Lost the 50WE here in 2011. Retook the 50WE here in January of 2012. Had some turbulence, but ultimately that was bust cycle turning into boom cycle. Okay, next one. Lost the 50 here in August of 2014. Loss, loss, loss, loss, loss, retake here in 2015, October 2015. And boom. That was the beginning of the next boom cycle in Bitcoin. Okay, next one. Lost the 50 here in June of 2018. Failed to retake. Boom. Lost, lost, lost, lost, retake here in May of 2019. And boom, that was the beginning of the next boom cycle. Okay, let's move on to the next one. Lost it here in January 2022. Tried to retake, not a convincing retake. Lost, lost, lost, lost. Boom. Retake here March 23. That was the boom cycle. So, this is just the pattern. We lose the 50 bus cycle. We retake the 50 boom cycle. That's why I want to see the 50we retake. 50 week right now is 796. If we can close this week above 796, then I would say more confidently. I am bullish on Bitcoin right now. The 200 day retake for me was a lot of really good bullish evidence. I like what's going on with stable coins. I like what's going on uh on the legislative side of of things. I'm positive. But a real clean signal to kind of be like the all clear, if you will, would be a 50we retake. I am cautiously optimistic. We can finally do it this week and I can finally wave the green flag and say, "Let's go." >> Excellent. So, uh, one thing we talked about last week is that liquidity has dried up. So if Bitcoin retakes its 50, uh can that be a leading indicator that liquidity is coming back to the markets and that we'll see another spike in the AI market also? >> Yeah. Yeah. Absolutely. Like spot on. Bitcoin is weirdly good at leading stocks. Weirdly good at it. Like when Bitcoin starts to, you know, really come into its own, stocks tend to come into their own, you know, a few weeks to a few months later. It tends to happen. And so yes, I do think that if Bitcoin really does have a big rally here, that big rally is essentially front running a liquidity reversal, um return of excess liquidity in the markets and a a big rally in in stocks. So I do think that is 100% true. Bullish on Bitcoin and stocks is something you want to do at the same time. >> All right. All right. Bye bye. Bye. Uh any any closing thoughts here, Luke, before we uh wrap the episode? I just think that um the AI bifocation theme has been fundamentally true for a while but now it is it's become again it was not a financial truth throughout June July and August but it is becoming a financial truth here again in September and I like what I'm seeing in the charts out of whether it's socks or SMH or AIS or individual names like Marll or Qualcomm I like what I'm seeing I'm seeing technical evidence to suggest that the financial reality is making a rebound to match the fundamental reality, which gives me confidence in saying that these are rebounds that you want to be buying and that the AI trade is coming back to life. So stick with it. If you own the stock, stick with them. If you want to buy more, buy more. This feels like a lock and load time to me. >> Love to hear it. Yep. Uh these episodes have been a bit longer this week, but there's just so much information [music] to uh recap. Again, uh we hope you like, comment, and subscribe. Ask those questions. We will have a questions episode in the future. But that is all for our macro episode. We hope you enjoyed. Take care. [music] [music] >> [music]