In this week’s AI Update, we dive into the rapidly shifting landscape behind today’s sky-high AI valuations and explore why macroeconomic forces may be creating unseen systemic risks across the tech ecosystem. Pavan Agarwal breaks down how the unwinding of the dollar-yen trade, leveraged Bitcoin positions, and hyperscalers’ overreliance on overpriced NVIDIA hardware are converging to put pressure on companies like OpenAI—potentially triggering a broader contagion that could impact banks, data-center REITs, and long-term AI investment flows. This episode uncovers how tightly interconnected global markets, digital assets, and AI infrastructure really are—and what leaders in mortgage and fintech should be watching next.
[David] Listeners, it’s time for another AI update. We have joining us again, Pavan Agarwal. Pavan, good to have you back.
[Pavan] It’s good to be here. What an exciting time we’re at.
[David] Oh my gosh. And look what’s happening in the world of AI. There’s
all these reports coming out, which I want to dive into. I sent you a video the other day about an analysis about Open AI, the stock, the valuation of it. And you came out with some great perspective. And I was thinking, is this time to be shorting that stock and play that side of the market? The short end of the market there. You had some really good markets. And I know everyone’s looking at valuations. You had an amazing valuation that came in. $119 billion came in on Celligence for the Angel AI product. So I mean you understand valuations. You hired one of the top firms. What’s going on in the world of open AI, this LLM, Large Language Models, when it comes to valuations? There seems to be some shifting and going on there. Please share your thoughts.
[Pavan] Well, I mean, at this point, what is not going on in the world? I mean, it’s like it’s almost like you don’t even know what’s going to happen next. This is very, very interesting times. and I think the thi ng that I’m watching the most closely is the dollar yen trade.
[David] And I’m gonna interject because it’s so interesting you bring that up. When we’re recording this, we’re recording this on Tuesday, December 2nd. We’ll get this released this week. But what happened was, is Japan was the reason the bonds you know deteriorated, the tenure deteriorated. They blamed it on that. So it’s interesting that you’re bringing up the Japanese yen because of their recent decision, their central bank recently increased their tenure treasury or ten years, and that kind of gave the market a reason to go up. Now, there’s a lot to that, so let’s go ahead and jump over there. I was going to wanted to talk about valuations in a bit.
[Pavan] Well, it’s all related, it’s all interconnected.
[David] That’s interesting.
[Pavan] The dollar yen trade is tied to Bitcoin and is tied to AI and tech stocks, this is such an interconnected world, right? And we have like policymakers and politicians do something and they have no idea like the ripple effect that they send through the whole system, right? So there’s nobody there’s nobody smart enough to figure this stuff out, and they just do stuff.
[David] Not knowing the cause and effect of what they’re doing.
[Pavan] Not knowing the consequences, and we all know that none of our leaders on either side of the aisle are that smart. Yeah, no.
[David] So you were starting to talk about the yen against the dollar.
[Pavan] The new prime minister has been raising the rates. I actually agree with her because you know, Japan’s been doing basically for 30 years, they keep, you know, definition of sanity is doing the same thing over and over again, expecting different results. Right? Thirty years they kept the interest rates artificially low, they kept their yen high, right? And interest rates low. So what’s been happening is basically wealth from Japan has been going to higher interest rates markets, which is the US, right? So a lot of money from Japan has moved into the U.S. and has been driving up U.S. equities and so forth.
[David] That makes sense. And then let’s and think about that for a minute, listeners. You have the ability in Japan to borrow money at zero interest. That’s where it’s very close to. Now it just jumped up to 1.61. So it’s still way below our 10-year treasury, but that’s now taking away some of the advantage for people to go in and borrow on the Yen and then go and invest it in dollars. In the U.S. market and spread right there. You could just buy treasuries and be making a nice spread. So she’s fixing that. That’s what you’re referencing.
[Pavan] That’s what she’s fixing, right?, because that’s what they’ve did since the currency crash in the late 80s and basically the last decade of Japan that become the last three decades in Japan. And they’ve kept this policy forever, they’ve got to change something. I agree one hundred percent. I mean, you know, the something’s gotta change. Right. So the whole idea of the central bank was let’s drop interest rates to zero so that we can spur economic activity, but instead of doing that, okay, people with money did said, hey, it’s a lot easier. Why would I take risks and invest it locally when I could just go ahead and buy US equities and US bonds and make a lot more money that way and not have to work as hard. So clearly that’s not working. So all that did was impoverish the country, right? Is essentially wealth left the country to the US and other places around the world and so that means, well, if this isn’t working, and essentially the government is subsidizing that, right? By beneficial interest rates. Basically, the government’s losing money. Yeah, right. The Japanese central bank is losing money by lending out yen for nothing. And so and so that all so that was basically this this dollar yen trade is what they call it, and it was a it was a it was a risk on trade, right? So if I could borrow money at zero, right, then I’m and I’m incentivized to put it in riskier assets, okay, like Bitcoin, like AI shares and things like that, because then I could say, hey, you know what, I could bet some over here and I I’m not paying any coupon, right? I’m borrowing it at zero, and I can get five hundred percent returns, right? Or maybe five thousand percent returns, right? Um and it’s like it that made so much sense, and that’s why all so I saw a headline the other day, some six hundred billion dollars of Bitcoin, um or is it six hundred billion or six hundred million? I don’t remember the exact number, but a very large amount of Bitcoin was sold uh off in Japan. Okay, so this is the dollar yen trade unwinding. I think it was six hundred billion, so it was it was dollar is this is the dollar yen trade unwinding. Right. So it’s like okay, well if borrowing money gets more expensive in Japan, then I can’t put it into Bitcoin. Okay, because I need cash to pay the coupon. If I could borrow a zero, I don’t need any cash, but now I can’t put it in Bitcoin because Bitcoin doesn’t pay me any current coupon.
[David] That’s right. No there’s it’s not a bond. It’s not paying out dividends, yep?
[Pavan] Yes, exactly. So the Japanese have been selling off Bitcoin, so that’s what’s causing the Bitcoin to fall. And they’ve also been selling off their like you saw Softbank unwound their position on NVIDIA.
[David] Exactly right. That was significant.
[Pavan] Yeah, right? Completely unwound their position on NVIDIA. Um and and so there’s a lot of Japanese money that has been moved into the AI um AI world, also, right? So because it’s like okay, I could borrow money at zero and and put it in in AI uh private equity, right? And yeah, I know as as I gotta wait. You know, if I’m the investor, I’m thinking, yeah, I know I gotta wait until these things go public before I get a pot, but I’m not paying any interest anyways, so let’s let it sit there. Right? And and just let it sit there and watch the equity value of these things grow and rise, and then as they rise and then they they liquidate, then I I’ll get my returns. Right. Right. Right? But because there’s no cost to carry. When you borrow money at zero, there’s no cost to carry. Right. So you can do trades where you can just buy and hold. Okay, because there’s no negative carry on those trades. Okay, but now that there’s a negative carry, all of those trades are unwinding. Okay. So and that’s pulling money out of um anything that’s a long hold before there’s a return. So we’re seeing thatt effect uh RUDI is affecting the Bitcoin market in a big way, and we’re seeing it also uh affecting the AI markets, right? Uh with Softbank selling some other position, that’s that’s an early indicator that this this stuff is happening. Yeah.
[David] Which goes to the whole thing, I mean, I mean, but on the in the long-term picture, still Bitcoin is still a great investment. This has just created a buy opportunity for the same.
[Pavan] Right. This has created a buy opportunity, and there’s other things happening with Bitcoin because um, you know, yesterday, you know, Peter Schiff uh he’s put out he put out a very funny, funny post as you know, you know, he lives down the street from me here in in Puerto Rico and he’s been right about being long gold, right? He’s been talking about you got to go long gold for a long time, and he’s been right. I’m looking at gold is I don’t know what the latest is four or five thousand dollars an hour or something like that. It’s something crazy, right? So yeah, I got ahead of him. He was one hundred percent right, and he saw the train wreck coming. And he’s also been saying, hey, Bitcoin is way too high. Right. That’s been his position. I don’t I don’t agree with him on the long run. Okay, full disclosure, I think Bitcoin in the long run is still heading towards astronomical heights. I think this is a short-term um trading technical or is this short-term market correction, but long run I think this is gonna go go really well for Bitcoin. So a lot of drove up Bitcoin was MicroStrategies buying up uh the market. And then yesterday MicroStrategies basically is like saying uh they’re having to instead of buying Bitcoin, they’re putting cash out to attract right, they’re paying a 10% preferred return to get money to come in so they can meet their margins.
[David] I could there’s that there’s a lot of margins in calls that are happening and people are having to sell other things or selling Bitcoin to get to the point where exactly.
[Pavan] And and MicroStrategies can’t sell Bitcoin, right? That would be converse or in conflict to their business plan. Right. So instead they’re paying a high preferred return to get cash in the door to then turn around and meet their margins because they bought a lot of the Bitcoin they bought is on leverage. So if Bitcoin falls, I forget the magic number, if it falls to 75,000, then MicroStrategy gets into really, really big trouble. So and I think and I think the hedge funds are out there, they can look at this. It’s just as you know just like just like how the hedge funds showed it showed it Lehman brothers out of existence. And I think they’re looking at the uh play on hey, micro strategy is on the verge here. We have Bitcoin is falling, right? We have the all the stars of line in the world just right, you know, with the dollar yen trade and a couple other things happening. They’re like, hey, maybe if we press it, we can cause Bitcoin to drop, pop micro strategies, and pick up pick up all this all these gains, right? So the shorts are now out and creating this feedback.
[David] Yeah. It’s really interesting. And how is this so how is this affecting the Open AIs of the world? Tie it together.
[Pavan] I wouldn’t want to be open AI.
[David] This article and I’ll by the way, listeners, I’ll put a link to the video that I watched, the YouTube video I watched that really brought this conversation about today because there’s some major corrections happening.
[Pavan] I think the biggest challenge I see with the OpenAI plan is so much of the money is spent on building these mega, what they call hyperscalers, these mega data centers. Yep. Right. And those data centers are buying NVIDIA chips and for fifty thousand dollar Nvidia chips that are probably worth really two thousand dollars, right? Wow. Okay. And so all that money is effectively going back to NVIDIA, and that’s causing this loop of looping up. That’s what he brought out in the video. Looping valuations, right?
[David] And looping up. I mean and to unrealistic.
[Pavan] So two weeks ago, SoftBank sold their NVIDIA position, right? and what two days ago, Google announced basically their AI architecture, the AI data centers are Nvidia free with their own Google TPUs, right? Tensor processing units, right? Which do the same thing, you know, specialized software obviously don’t run the TPUs, but they’ve built the whole ecosystem around it, they do the same thing at a fraction of the cost. Right. So then it starts to question wait a minute, all this money, um, which by the way, I think over um I think it’s close to a trillion dollars has been invested in building these mega data centers powered by Nvidia chips, right? And a lot of those data centers, by the way, are levered up by various major banks. Okay.
[David] So the failure when as this starts this is to unwind,
[Pavan]: There’s could be a b major contagion. Now if you’re saying these chips are two thousand dollars selling for fifty thousand dollars, right? That’s like, you know, a twenty twenty-five X overvalued, right? And Google just demonstrated, hey, I’m I I’m doing it better and cheaper. Yeah. So they don’t really care about the noise, and they’re just they’re just moving, moving forward and building a better AI solution. Okay. And so you so you got Google doing this, you have AMD coming out with their chipset that competes with uh Nvidia and an and an open architecture, right?
[David] So that is a huge that’s a big key.
[Pavan] Right. So Google has the Google released an open architecture, AMD released an open architecture, and then there’s the open source community, Apache and a few others that have come up with completely open architectures for GPUs. So and whereas what’s been keeping everyone hooked on to NVIDIA is NVIDIA software stack, which they call CUDA, C U D A, right, and that’s keeping everyone which is a very good ecosystem. Right, yeah, right.
[David] But they when you have a limited commodity controlled by one person, they start setting the prices, and that’s why you get 20, 25x of the value of the chip.
[Pavan] Exactly, exactly. So and this and these days, look, you can create like if you’re Apple with the iPhone, that’s a closed architecture. Yeah. Okay, but and they succeeded at keeping that closed because you know the it’s sold to consumers. The consumers are not going to care about getting inside and reprogramming and rewiring it, right?
[David] Yeah, and it’s gonna close is better than it works, and that’s why we both own one.
[Pavan] Right. Well, and I actually am an Android guy.
[David] Well, that’s right, you’re an Android guy. That’s why I didn’t mean I’m an Apple guy. I’m in the Apple system.
[Pavan] Right. So if you have a closed architecture, that’ll work if you’re working with consumers. Consumers don’t want options.
[David] Right.
[Pavan] Okay. But but in in data center world, that’s being run by engineers. Right. And if you think about how screwed up the market is, it’s like you got one Nvidia, one chip company, and you got maybe six hyperscalers out there, right? You got you know OpenAI, hyperscaler is real quickly. Real quick briefly. Hyperscaler is is is a fancy word of saying mega data centers. Okay. Mega AI data centers. You got XAI, Microsoft, Meta, OpenAI, Amazon, right? Yep. Maybe one more. I’m you know, so you got basically six big mega tech companies buying up Nvidia chips, right? And they’re run by you know the smartest engineers in in the world, right? And so you don’t think they can figure out how to switch chips, right? If somebody comes up with a chip that’s one-tenth the cost, you don’t think they’ll switch it out?
[David] That fast, yeah.
[Pavan] Right, they’ll throw a team of engineers at it and say, you know, rewrite the code. If they have to, they’ll rewrite every line of code and they’ll switch it out.
[David] And now and now with AI you can rewrite the code so fast. So it’s not a huge amount decline.
[Pavan] Exactly and anyways, all these open standards, like the AMD’s open standard that they created, it’s all backward compatible with Nvidia’s CUDA. So there’s like you know, the whole NVIDIA ecosystem, right? And I think Nvidia thought they were Intel, where they can just own the whole AI space, right? Again, it worked for Intel because they were partnered with Microsoft and they had this and it was targeting consumers, right? Consumers cannot go in and recompile and rebuild their all that software onto alternate chip systems, right? Although that’s taken, you know, it took what forty years, it’s happened now, everything’s switched off the Intel’s a big trouble, but that’s a different story. But for NVIDIA, which is working with only six other companies, really if you think about it, six other mega companies, and they can switch on a dime. So there’s no moat, right? You know, we call in the tech industry a moat. Nvidia has no moat. They had some moat, right? I mean they had a they had a first move advantage, they had a great ecosystem, and it was and everyone plugged into it, right? But those some people who plugged into it can easily plug out of it, right? Nvidia is not selling to me you and me. Yeah, right. They’re not selling to Mr. Smith on Main Street, right. So that’s the problem. I wouldn’t want to be long NVIDIA stock right now. I wouldn’t want to be long on an open AI investment because most of that investment in OpenAI has hasn’t really gone into creating like the open AI tech is awesome tech. But is it any better than Google’s Gemini or Anthropic’s Cloud or you know Meta’s Meta stuff? No, it’s all about the same, right? Yeah and in fact the latest version of Gemini, there’s internal memo that just leaked from OpenAI that basically says how much better Google’s Gemini is than OpenAI. And all that money that’s been spent on OpenAI has been spent on OpenAI making these big, huge mega data centers, right? It’s just hardware, right? I mean it’s like you know, if you if you just go break things down to fundamentals, I mean how often, you know, like we all have laptops and whatever, uh right. And you know, your even your iPhone, I mean you in two years you throw it away, you get a new one.
[David] I know.
[Pavan] It’s like like why would you spend so much money on hardware? Because it’s it’s gonna go obsolete, right? You know, call you can call you can amortize it over three years, six years, play whatever games you want, but you know, it’s not something that’s gonna last, right? An investment is something that will last you a long time. Like good money management, right? Is if you go to the mall, anything, almost anything you go to the mall and you buy, right? if you don’t own, if you will not be owning it after the next three to five years, right, that that’s an expense, it’s not an investment.
[David] This is just so fascinating. And I I mean I what I’m wanting to do is keep these these uh updates as tight as we can. You’ve given us a great amount of information. Sum it up. What should we be looking for? as the listeners that are listening to you on this, you’re truly an expert on this on the on the subject matter, expert on all of this. What is your recommendation as we wrap this up?
[Pavan] This is what I think is gonna happen with OpenAI. Unless like we had a sudden shock to the system with the dollar yen trade. Right. And we also have this whole thing happening around micro strategies in Bitcoin. Right. And the other problem here is that a lot of the Bitcoin, as Bitcoin has shot up past $120,000, a lot of Bitcoin holders had levered up their Bitcoin, right? and said, Oh, I could really increase my returns. Let me borrow money against my Bitcoin and invested in AI stock. So a lot of the money feeding the AI stock, AI ecosystem was Bitcoin money. So the two are interconnected.
[David] Yeah, so they’re getting margin called though, is the bottom line, right?
[Pavan] Right, so the Bitcoin guys are liquidating, so the money flow from the Bitcoin to AI has stopped if anything may be getting pulled out. And on the other side, the dollar entry is happening, so high risk long-term high risk investments have slowed down there. So they say two strikes have been hit against the AI world, right? Which it’s you know, it’s not these aren’t by far fatal, these are just you know, they create some noise, they create some adjustments. If it just stays at this and nothing m major happens after this point, okay, so the so I think we’re gonna be I think the uh open AI um IPO is gonna do great. Right. I think everything will be will be just fine and at the long run we’ll look back and look at these corrections as hey, this was healthy. It corrected the markets to where it should be. But if one more blow hits this market you know from some other direction, and who knows what, right? It could be anything. It could be another war, it could be something crazy happening with oil prices, yeah. Nobody knows what could happen, right? So if one more blow hits the market and disrupts this capital flow, okay, then we could see um a systemic collapse of these AI companies and then the contagion effect, because those data centers are heavily levered. Right. So these come in the banks. And either banks or REITs, right, because they’re data centers, they’re real estate in effect, so they could go into REITs, right? So you got these heavily levered banks and other major systemic institutions, right, that that could con you could have kind of a 2008 contagion effect and there’s and could cause this cascading bank failures and major fund failures. And if that happens, you know this it would be basically the Fed will have to bail out again. And OpenAI, I think the CTO of OpenAI, the CFO of OpenAI, actually suggested that, that we’re too big to fail. And she didn’t she didn’t say exactly. She didn’t say that exactly, but she implied that, hey, you know what, the Fed is going to need to step in.
[David] It’s crazy stuff.
[Pavan] Right. So, you know, I guess when you know the old saying is, if you owe the bank five million dollars, the bank owns you, right? But if you owe the bank five billion dollars, you own the bank, right? And Open AI and few of these companies that put themselves in that position where they own the bank.
[David] Amazing. Fascinating. Well I want to get more into that in the next episode. Thank you so much, Pavan, for coming in and sharing some of the latest activities, what’s going on. Any if what’s some angel AI updates? I keep seeing all these wonderful testimonies real quickly that they just keep coming out and coming out and coming out. How the people are discovering the power of angel AI and what how it is working and transforming so many businesses.
[Pavan] Well, close off with something really fun. our newest angel AI athlete, Kelly. she is this is like if you’ve ever seen the movie Cool Runnings, right? Which is the Bob the Jamaican bobsled team going to the Olympics, right? Yeah, so she is the only Winter Olympian from Puerto Rico.
[David] Love it.
[Pavan] It’s not a bobsled, it’s that single sled. It’s a single sled, yeah. It’s a single sled where she goes goes down really fast. She hit over 100 miles an hour at the bottom. So she just won the gold in the North American Cup. I think just happened two days ago. it was so cool. She went up, took her gold medal wearing an angel AI jacket. That was amazing.
[David] I love the fact that you sponsor so many athletes, young athletes that are up and comers that are are absolutely winning at the game and their craft. And it’s so exactly what you’re doing. Pavan, thanks so much for being here again. It’s congratulations, kudos to you, and looking forward to spending some time with you at the Christmas party.
[Pavan] Thank you, David. See you there.
[David] Appreciate you.
Important Links
- Michael Burry Bets $1.1 Billion against the AI Bubble
- SunWest Mortgage website
- Angel AI website
- LinkedIn – Pavan Agarwal
Pavan Agarwal is a renowned leader in the mortgage lending industry and a pioneer in bringing artificial intelligence to the financial markets. Agarwal serves as the President and CEO of Sun West Mortgage Company and Celligence International, LLC.