Tony Nash - Forecasting Global Markets with AI

JohnPaul: [00:00:00] Hey everyone, welcome to the podcast. I’m your host JohnPaul and this is Digital Gold. Known to many as the Bitcoin Kid, I started my own cryptocurrency out of my parents’ basement back in 2013. The goal of this show is to simplify the crypto world and explore how it changes the way the world thinks about money through conversations with thought leaders [00:00:16] in this space.

[00:00:17] JohnPaul is the founder and CEO of Orm Capital Ventures. All opinions expressed by JP and podcast guests are solely their own and do not reflect the opinions of Orm Capital Ventures. This podcast is intended for informational purposes only and should not be relied upon for investment decisions.

JohnPaul: [00:00:42] Tony Nash is the CEO and founder of Complete Intelligence. Using advanced AI, Complete Intelligence provides highly accurate market, costs, and revenue forecasts fueled by billions of enterprise and public data points. Previously, Tony built and led the global research business for the economist in the Asia consulting business for IHS. He’s also an associate entrepreneur, [00:00:59] media entrepreneur, writer, and consultant. I’d like to welcome Tony Nash to the show.

Tony: [00:01:04] Great. Thanks JP.

JohnPaul: [00:01:05] Tony, as I mentioned, you’re the founder of Complete Intelligence. Can you tell me a

Tony: [00:01:09] little bit more about what Complete Intelligence does and how you work with your clients? Sure, yeah. As you mentioned in the intro, I led global research for British firm called The Economist. I led Asia Consulting for an American firm called IHS Market. In that time, over about a decade, I had a bunch of clients come to be saying, we have two problems. [00:01:33] First, forecasts are terrible. That was a comment both on the work of the firms that I worked with as well as just the market generally. They said forecasts are terrible. There’s no accountability of the forecasters and nobody tracks their historical data. We have to try to dig it out ourselves.

Tony: [00:01:53] Forecast accuracy is a huge issue. The second issue is the appropriateness of a forecast. If you make a chemical or a mobile phone or cake mix, there are specific items within that product that you need to know the cost of, but you may not be able to do that internally. Most companies have, major companies have hundreds of Excel workbooks floating around with their [00:02:20] forecast for sales or for costs or whatever. It’s just really confusing. What ends up happening is people manually estimate costs and revenues. What we wanted to do was automate that entire process, company-wide. We wanted to take out the human bias that comes with forecasting industry and internal forecasts and all that stuff. We really wanted to build

Tony: [00:02:46] products that allowed the machines to learn how markets move, so that’s currencies, commodities, equities, and so on, as well as how company revenue and spend changes over time. When doing some of my initial research on complete intelligence, basically just to paraphrase, you guys are taking the spot of what an analyst would do. Is that correct? [00:03:10] Yeah, but here’s what we don’t do. We don’t put together a report on what’s going to happen in industry X or with commodity Y, because what we find is when that stuff is put together, so when an analyst puts a report together on some aspect of an industry, it’s really loaded with a house view on something or a personal bias. We do have a weekly newsletter and we do

Tony: [00:03:40] video podcasts, that sort of thing, but we don’t have industry notes because we don’t want our clients to feel like we have bias towards the oil and gas sector or toward industrial metals or that were for or against gold or for or against crypto or something. There’s so much of that loaded into forecasting today, and it has been that way for decades, that we just want to let the data [00:04:07] and the sophistication of the data. We’re doing billions and billions of calculations every time we run our process. Humans do this, but they’re not aware of it. Humans also aren’t aware of the amount of bias that they put into their calculation. What we do is we track this and we track it based on error rates and we allow the machines to correct based upon how they’ve made error over

Tony: [00:04:36] time. It’s just like an infant learns. You touch a hot stove and you learn not to do that again. It’s very similar the way we reinforce the behaviors that we want within our platform. I guess my question to you is when it comes to these machines, they’re learning in the background. You don’t have a team of a thousand analysts and said you have a team of a thousand [00:04:56] neural networks or machines basically working for you, running these calculations 24 seven on all these different commodities. Are they just making assumptions and then confirming if those assumptions

JohnPaul: [00:05:06] are right and then the models that do better end up getting weighted more? How does that work?

Tony: [00:05:12] I guess. How do those questions and answers work in those data testing points? Is A.B. testing that you mentioned? It’s a good question. We’re running tens of thousands of scenarios for everything we forecast, every time we forecast. Then we’re looking at which ones best reflect the market as it stands right now. Then we add in the different approaches on a weighted basis to make [00:05:36] sure that they reflect where the market is. It’s a multi-layer analysis. It’s not just a basic regression correlations driver, that sort of thing. We’re also looking at the methodologies themselves. Some of these are very fundamental traditional statistical methodologies. Some of them are more technically driven state decision trees, those sorts of things, types of machine learning

Tony: [00:06:00] models. We’re looking at how on a proportional basis those different methodologies best understand the market at this point in time. That’s a long way of saying yes to your question. I think that was a great answer. You guys are looking at currencies, equities, and in July,

JohnPaul: [00:06:20] you discussed gold and silver being nature’s Bitcoin. Can you explain to our listeners what

Tony: [00:06:25] you mean by that and provide your thoughts on Bitcoin as a store of value and where you see that blockchain space going? I think one of the key aspects of cryptocurrencies is that there should be a fixed amount of it. If it really is immutable, then there’s only so much of it. If there really is demand for something that’s limited, then the value should rise or fall based [00:06:51] upon the availability of that fixed good. Gold is similar in that I can’t necessarily go out and buy a car with gold. I’m sure I could. I can’t buy a loaf of bread with gold. I think cryptocurrencies is becoming a bit more spendable than precious metals, a bit more useful depending on which cryptocurrency you’re looking at. But yeah, it is similar in that cryptocurrencies to date

Tony: [00:07:18] have been more of an asset than a currency. They’ve behaved more like an asset than a currency, meaning the value goes up and down pretty dramatically based upon the perception of scarcity. Currencies don’t necessarily act that way. Currencies act as units of value so that you can buy other stuff. Gold is on some level kind of nature’s Bitcoin or nature’s cryptocurrency, but I think [00:07:46] we’re coming to a point where there’s a division between those two, where cryptocurrencies are starting to be used as, and when I say starting, of course they have already been, but more broadly be used as vehicles to buy other stuff, not just stores of value. So the former is a currency, the latter is an asset. Yeah, I definitely agree with you on that point as we move down this line of utilization,

Tony: [00:08:11] we saw with PayPal news that we recently came out, Square News. Hopefully people will start using Bitcoin more as a day-to-day currency. It’s one of the biggest questions I get is it’s too hard to use Bitcoin. When am I going to use that to store less of actually Bitcoin as a store of value, especially from some of the retail clients coming into this space? So regarding Bitcoin [00:08:29] and complete intelligence, are you guys forecasting anything in the digital currency space? Are you forecasting the currencies themselves? Maybe the mining profitability or any of the mining machines, and can you speak a little bit farther on that? We do. We started forecasting limited cryptos about six months ago. And as I’m sure you can imagine, there’s been a lot of volatility in

Tony: [00:08:51] cryptocurrencies over the last couple of years. And because we’re a machine learning platform, it takes a while for the machines to understand how cryptocurrencies trade and move. And so just because we started forecasting cryptocurrencies doesn’t necessarily mean that we would recommend people making trades or taking positions based upon what we forecast. It’s different for things [00:09:19] like copper or whatever that we’ve been doing for a long time. And those are also relatively stable markets, say, industrial metals, that sort of thing. But cryptocurrencies, very volatile, very new. And the market is still learning how to value them. This is one of the key things about cryptocurrencies that I think is misunderstood is the market is still learning how to value them.

Tony: [00:09:44] That’s not a comment on whether I think they’re undervalued or overvalued right now. I just think the market isn’t really sure how to value them. And so in our platform, we expected to take really another couple months before we’re confident in where our platform is saying cryptocurrencies will go. Again, because it’s such a complicated asset in the way it moves, [00:10:11] and because there’s so little institutional and historical knowledge about it, we have to iterate a couple billion more times for us to really understand where it’s going. Are you seeing a lack of data or trading data, network data, in making these decisions, that making it harder than traditional markets? Or have you seen that the data in the Bitcoin

Tony: [00:10:31] space is relatively open and well established? I don’t really see an issue with data. I think part of the problem with cryptocurrencies is that it doesn’t really trade on fundamentals. So what we’re utilizing is a configuration of methodologies that balance out fundamentals and technicals. Some months, certain assets lean more toward technicals. Some months, they lean [00:10:54] more toward fundamentals. Cryptocurrencies don’t really have fundamentals to lean on. And so then you’re looking at a lot of relatively short-term and ultra short-term approaches to understand the value of something. So the memory of the price, it’s either sticky or it’s not. I know that sounds a little bit silly, but cryptocurrencies move in bursts,

Tony: [00:11:20] or they languish. There’s really not a lot of in-between. And so understanding which technical approach is to take and within what configurations to take them is what’s really kind of confounding our platform right now. And I would say our error rates for cryptocurrency is probably, I think, three times what our average error rate is. So our average error rates for across our assets [00:11:52] on an absolute percent error basis is between five and 7%, something like that, across currencies, commodities, equities. For cryptos, we’re looking at probably a 15-ish to 20-ish percent error. And so it might be a little bit lower than that now, but it’s settling within the range that we’re comfortable with. We’re really comfortable when things are, say,

Tony: [00:12:14] less than 10% error. And we expect to be there very soon. But part of what’s different about what we’re doing is that we’re not afraid to talk about our error rates. We’ll be very transparent with people about what our current and historical error rates are and have been, because our clients are making decisions based upon the data that we bring to them [00:12:40] and the forecast that we bring to them. So when I say to you, look, our error rates for cryptocurrencies is between 15 and 20%. I’m not really sure you can find many other people who would admit that publicly. But if traders are making decisions based upon the forecast that we bring to market, then they need to know that. They need to know how to hedge against that error range.

Tony: [00:13:04] So referring to that the cryptocurrencies are much harder to predict. Is that keeping any of your current clients from moving over to the digital currency space? Are they looking at this space for growth opportunities or for potential revenue generating opportunities or even a way to hedge from the current macro environment? I think everyone is either involved and trading, [00:13:28] let’s say, even at a small level, or they’re very committed. I think the approach that we’ve tried to take, the number of firms that get very hypey about cryptocurrencies and almost feel like they’re trying to push it on to their clients. We’re not that way. We don’t care if someone invests in iron ore or invests in cryptocurrencies. It’s really what is their profile and how well

Tony: [00:13:51] can we forecast it? But I think the interest in cryptocurrencies obviously is still very high because nobody really knows what’s happening there. Nobody really knows what the future is there. And nobody really wants to miss out. Actually, I know maybe two or three people who want to miss out on that. But very few people want to miss out on it. And so they’re keeping an eye on it or [00:14:18] dipping a toe in if they’re not already in in a big way. And I think you have to be fair on these sorts of things. It’s not as if, say, the main cryptocurrencies have kind of fizzled out, they’re still around. They didn’t fizzle out after, say, two years. They’re still around. People still trade them. We’re still trying to figure out how to get them into some sort of monetary system

Tony: [00:14:43] or some sort of transmission mechanism. And until that’s figured out, I think that unless they fizzle out, the main ones, I think it’s still necessary to stay involved. So we’re not seeing a massive demand for what we’re doing in terms of forecasting. And when I say forecasting, I’m not talking about the next say five to seven days, I’m talking about the next 12 [00:15:05] months, okay, monthly intervals over the next 12 months. So for something like cryptocurrencies that have a relatively short term horizon, because it has been pretty speculative from an investment perspective, it’s been pretty hard to look at this stuff over a longer term. But we’re getting better at it. And I think as these things become more predictive, there will be a lot more interest.

Tony: [00:15:31] And that’s largely the market coming to agreement on what the various cryptocurrencies are actually worth. And following up on that, how do you value them as being a common trend? It seems like in the analysis that you guys are doing, as a large Bitcoin miner in this space, we believe the stock to flow ratio is a huge component of giving value to underlying cryptocurrency. And so that is when [00:15:55] the halving occurs, did your models take that into account? Or did they, how do they kind of work with that event? Because I think the halving is an event where you don’t really have that in any other industry where you’re losing half of your new coins coming in or half a new supply coming in on a daily basis. Well, I think you, you know what, you do see this a bit with say, central bank

Tony: [00:16:15] money supply, that sort of thing. So and you do see, let’s say with a dollar or the euro, the Japanese yen or something like that, you do see central bank money supply coming in. And the pickup of that money supply is not fundamentally dissimilar from cryptocurrencies. Although I think with cryptocurrencies, it’s a it’s a fair bit more technical. But I think it’s, you know, understanding [00:16:40] both the stock and the flow is critical to understanding where that value is. If there’s too much stock, then, you know, it’s obviously not not valuable unless there’s the demand, the flow going into demand. So so yeah, I think it’s it’s in but until people can have a normalized discussion around it, whereas it where it’s similar to say central banks, then I think it’s really hard for

Tony: [00:17:09] people to contextualize within their kind of trading and valuation framework. So look, you know, if you look, for example, you know, the Chinese government introduced this this coin into Shenzhen a few weeks ago, right? They effectively gave people the equivalent of $30 in this Chinese crypto currency to spend. And then it was gone. So they’re calling that a study on how widespread adoption [00:17:37] of cryptocurrencies will work. And I’m sure it was gone within a day, right? I mean, if I’m given 30 bucks to spend for free, then I’m going to spend it probably today. So, you know, I think until we have a better baseline for widespread adoption. And I think the government endorsement on some level kind of matters, because let’s let’s look at that $30. It’s effectively like a voucher or a

Tony: [00:18:04] gift card, right? That they’ve given people a gave people a $30 gift card for free. It doesn’t matter what currency it’s in. Okay, it’s going to get spent, right? I don’t necessarily think that that’s a valid test of the adoption of a cryptocurrency. I think you have to have something more widespread and more enduring, because there you have a fixed amount of stock that spent over a very [00:18:29] abbreviated period. It doesn’t really mean anything, right? But I think until we have a wider spread adoption for spend, we’re not necessarily going to get a fundamental based value. Okay, we’ll get that technically based value, meaning looking at the stocks and the flows and trying to understand based on stocks and flows, but not necessarily based on the inherent value that you get with a

Tony: [00:18:59] legit currency, not that cryptocurrency is illegitimate. That was probably a bad word choice. But let’s say a central bank endorsed currency. We’ll say that much. And on the central bank endorsed currency chain of thought, when you see the United States and in Europe and also China adopting these different types of cryptocurrencies, or I guess you could [00:19:22] say ways to distribute capital to individuals for stimulus, how are you seeing China and the US and any other major players deploying these central bank currencies over the next two or three years? As you did mention, China is already doing it. In the US, I’m not aware of us doing any type of central bank currencies or deploying central bank currencies to citizens. But are you seeing,

JohnPaul: [00:19:50] I guess, how do you see that playing out over the next two or three years, if not, and maybe longer?

Tony: [00:19:55] Sure. China, the China central bank did a first test of a cryptocurrency, I think in January of 2017. So they’ve been trying to figure this out for some time. And I think China sees it as a potential way to rival the US dollar. The problem is there is no trust in the people’s bank, China, nobody else had to try to really trust it. So the immutable aspect of a cryptocurrency [00:20:27] doesn’t have validity outside of probably the walls of the central of the people’s bank of China building. And without that limited supply, without the immutability of it, then again, it’s just a gift card. It’s just a voucher. Now, I think the PBOC, the Chinese central bank has had, but with each day, it’s kind of passing. I think they’ve had an opportunity to utilize

Tony: [00:20:55] cryptocurrencies for things like trade finance, which is a really opaque aspect of international finance related to trade. And if they had, let’s say, gone to some of their trade partners and said, look, in Europe or the Middle East or somewhere, we can get around using the US dollar by utilizing this digital Chinese yuan or something. I think there was a time when people would have been [00:21:22] open to it, especially if it made payments faster and less costly. But I think that window has passed, at least for now. I think it’s really hard for China to insert itself. I think if they had done this, say, in 2015-16, I think they would have had a real opportunity and they could have done a lot to displace some US dollar-denominated trade finance and probably displace a lot of

Tony: [00:21:52] euro-denominated trade finance, but they didn’t do it. They’ll keep trying. I’m not sure how successful they’ll be outside of those places that have to trade with them, meaning North Korea, Iran, and those sorts of economies, Venezuela, and so on. With Europe and the US, I don’t think the central bankers fully understand what a cryptocurrency is. And I don’t think that they really have, say, [00:22:23] the patience to understand how to, say, deploy it in a credible way, if that makes sense. And so I think you’ll almost have these parallel currency regimes with cryptocurrencies. The problem, though, is I don’t necessarily, at least for the next few years, see them displacing a currency like the dollar. They may displace, say, secondary or tertiary currencies within, say,

Tony: [00:22:52] international trade, trade finance, cross-border payments, these sorts of things, and even domestic payments where, say, a central bank doesn’t really have credibility. That makes a lot of sense. But I’m not necessarily sure that I see it displacing, say, US dollar or euro transactions, let’s say, in kind of main, say, kind of day-to-day activities. If you look at a government like [00:23:18] Venezuela or Turkey or something like that where you see a real currency crisis, I think it’s possible. I’m not necessarily saying it’s probable at a place like Turkey, but I think it’s possible that you could see adoption of something like cryptocurrency, especially if the government puts a restriction on US dollar use. Tony, do you see, I mean, it seems like you’re saying that the

Tony: [00:23:43] Western China will have its own central bank digital currency and maybe the United States will try to deploy theirs as well. Do you think this is going to move the global economy into being a more closed system? Or do you think this will actually open up finance and trade and make it better for everyone? Or do you think, well, I end up having this almost finance war of, you know, [00:24:02] we already do have that, but on the digital currency level now, where it’s traceable and trackable by a single entity and the capital or the cost to deploy these systems is much lower? It’s a great question. I think the people who accept the digital Chinese yuan are going to have to decide if they want a centralized authority in China tracking all of their activities

Tony: [00:24:28] in that digital CNY. You know, I think that’s a real decision and a real trade off that those people who trade in that currency are going to have to figure out. Although dollars are traceable, you know, you can kind of transmit them and other currencies, you can kind of transmit them. I wouldn’t really say it in an anonymous way, but you can kind of get around tracking of [00:24:53] every single transaction. But with cryptocurrencies, you know, the ledger tracks everything. And so if you have, say, the PBOC in China tracking every single transaction for every single digital CNY that’s out there, that’s kind of next level of information out there, right? It’s not just Google understanding what’s in your email. And it’s not just Alexa tracking what you’re saying.

Tony: [00:25:18] It’s every single penny you put out there being tracked by a central ledger. And I think you said that perfectly. You know, China will be tracking every transaction. And that will help these central bank digital currencies, if it’s China, if it’s the US, if it’s, you know, somewhere in Europe and as these different currencies are deployed, they’ll really be able to build [00:25:37] almost a very well put together social graph of who you’re paying. I mean, it’s very similar to Venmo when Venmo had the kind of privacy error, where you could see every transaction, if you add your transaction on public, that you sent all your friends. This is almost like that, but the central bank can see that for every single person. Now we know who interacts with who,

Tony: [00:25:56] where you go, you know, if you’re going to get coffee at Starbucks every morning, where you’re going to be, you know, it’s very interesting to see the amount of power that, you know, these central banks in my opinion are going to start, are going to gain over deploying a currency where it’s traceable, trackable, and it’s on a single ledger. [00:26:13] Right. Well, so imagine, you know, right now we have, we have macroeconomic data releases like gross domestic product or industrial production or retail sales, those sorts of things. Imagine, you know, right now, the way that happens is a statistics ministry doesn’t estimate of what that economic activity is. And they release it like a month after it actually happens,

Tony: [00:26:37] and then they revise it four times before they finally give up and say that this macroeconomic variable is finished. If you do have a centralized kind of ledger for this stuff, you can actually look at national and global economic activity on a real-time basis, right? So you can actually see through, you could actually see through COVID, you could see the US economy [00:27:02] declining on a real-time basis or the Europe economy declining on a real-time basis, which would be pretty scary actually, but that’s the reality of it. If you have this centralized ledger, you can see, let’s say, the velocity of that currency grinding to a halt as people don’t spend money, which from a central bank perspective can help you understand how to

Tony: [00:27:24] incentivize people to spend money if they have it. So from a kind of centralized monitoring of the economy perspective, I could see that being beneficial. From a consumer and an individual save or spender perspective, I can see that being a little bit scary. It is a little bit scary, but I agree with you also with the COVID situation. The stimulus, [00:27:47] really in my opinion, didn’t get to the people as well as it should have. And central bank digital currencies will allow these central banks to give stimulus to those who are most affected, at least in theory, and to be able to provide potentially different access to credit for different types of individuals. We’re taking different types of risks being business owners

Tony: [00:28:05] or just employees. On the COVID analysis, and as you guys with CI were doing the analysis on the equity markets and oil and different types of currencies, did you guys see any indicators as COVID was picking up in the analysis of the market and how did it affect your predictions in these broadly over the different markets that you guys predict and watch? [00:28:30] I think what we saw in the wake of COVID was, and this is no surprise to anybody, I don’t think, is a move to very short-term thinking. What data points are coming out? What’s moving? What are people doing? Let’s track day-to-day what’s actually happening. Also, an eye on what is the government doing? What stimulus is coming out? When is it coming out? How much is it? Where’s

Tony: [00:28:53] it going? That sort of thing. I think for the probably three to four months, I would say until July or August, a lot of trading and forecasting was really done on that basis. The news moved the market. It was fear and news that really moved markets. We had to come to a place where the size of the dump truck of stimulus was bigger than the fear that people had of COVID. [00:29:22] When we got to a number big enough, you started to see markets break higher, which was, I guess, a positive thing for people who weren’t working, but getting stimulus from government so they could day trade and make some money in markets to shore up some of their bills. Now that the stimulus has gone out, and now that we see at least some markets coming back to, I wouldn’t say

Tony: [00:29:51] normal, but at least to a significant level, we’re starting to see, or we’ve started to see over the past, say, six to 10 weeks, more fundamental basis put into markets and put into some of those value decisions, whether it’s an equity or whether it’s a commodity or something. It’s still playing out. In a number of ways, a lot of the techs still very sentiment and [00:30:18] stimulus based. We see things like some of the commodities that are still very much based on that, or I would say more than 50% based on that, but we’re starting to see markets move back into a direction that’s a bit more traditionally based. I use that term very loosely, traditionally based, but with at least a bit of fundamental analysis. Look at something like Tesla, for example,

Tony: [00:30:45] the price to earnings ratio is around 1100, I think, something like that. You may love Tesla, but that’s a pretty healthy multiple. At some point, and I’m not necessarily predicting Tesla will fall to Earth, but at some point, something will catch up with the valuations of these things, whether they’re commodities or whether they’re equities, and we’ll start to value things [00:31:13] on a more traditional, again, that’s a loose application there, but on a more traditional basis. One of the things that I’ve been noticing in just conversations is it seems like the stock market is almost, I would say, really turning into a casino where you have people just buying stocks, they heard on the news, they’re getting the motley fool every week, and they have so many

Tony: [00:31:33] decisions to make, so many different options. I’ve noticed that it seems to be just too complex for, I would say, normal retail, robin hurt traders. They get overwhelmed with so many decisions. I think one of the nice things about value as we talked about value in crypto is at least with Bitcoin, you know what you’re getting. You know that this is an asset with a stable monetary supply

JohnPaul: [00:31:53] with a stable issuance rate over the next 100 years. What are your thoughts on how Bitcoin mining,

Tony: [00:32:02] I’m actually going to change it up and move to a separate topic, a different topic, but what are your thoughts on Bitcoin mining and how it relies on the global supply chain starts in semiconductor factories in China, and you mentioned the supply chain optimization a lot on your website as a function of complete intelligence. Can you walk through a little bit how you guys optimize the [00:32:24] supply chain? And then I’d love to talk with you through potentially how the Bitcoin mining supply chain works on our end and see where optimizations are and how COVID or any of these other things impact supply chains and what you guys are seeing on a worldwide basis. Sure, that’s great. I think with any supply chain, you have really three factors. You have cost,

Tony: [00:32:46] you have distance, and you have time. And so, I mean, there’s quality as well, but if you assume that you can get equal quality in multiple locations, you have cost, distance, and time. And so, we help people initially with cost. We’re helping them to arbitrage the best cost locations. We have a client who manufactures confectionary, they make candies and sweets, and they buy sugar, [00:33:13] I think, at eight different places around the world. And so, we help them understand where the sugar prices, because it’s not a single global sugar price, right? They’re local factors. So, we help them understand where sugar prices will change and at what magnitude they change, so that their factories can be prepared and that they can have the right margin they need,

Tony: [00:33:36] so that they can take in the right inventory, so that they can make the right transactions at the right time. So, I think from a pure cost basis with commodities, for example, like sugar, it’s possible to do that. When you look at something like semiconductors with a very sophisticated manufacturing process, cost is probably not the only, [00:34:01] well, I can assure you, it’s not the only factor associated with the decision. So, then you start looking at things like time and you look at things like distance. And so, when we go back to, say, March, April, May, a lot of semiconductors travel by air and we had air freight rates from Asia to the US that were normally, say, $1.50 a kilogram that had in many cases been jacked up to, say,

Tony: [00:34:29] $15 a kilogram, so 10 times or more of the normal price. So, that’s where distance becomes, or let’s say cost becomes a function of distance, right? And so, that chip set that semiconductor may cost the same X factory, but getting it to the destination is increasingly critical and increasingly costly. So, that’s where we help people also to understand what the cost of that distance is and what the [00:35:00] cost of that time is. Because you could put it on a vessel and you could ship it and it could take three weeks to get where it needs to go. But in many cases, the cost of those, the finished goods are high enough that you can absorb some of that transport cost. So, there are a number of ways that we help people understand those transactions, but at the end of the day, it all has to do

Tony: [00:35:27] with the cost of that bill of material, meaning the cost of the good that go into that finished item that’s ultimately sold to a customer. So, when we look at semiconductors, for example, and you look at what has happened over the last, particularly last year, and if you look at, say, TSMC, Taiwan semiconductor, moving one of their locations to, I think it’s Arizona in the US, [00:35:54] we’re starting to get more of that high value supply chain in the US, more as a function to de-risk supply chains in the wake of COVID, meaning factories in China closed during COVID, people still had to make stuff and they had to still have their business open, but they couldn’t because the factories in China were closed. Once the factories in China opened, there was constrained

Tony: [00:36:23] transport capacity, so it would cost them a lot more. So, they had goods that were late and they had goods that were a lot more expensive than normal. And so, I think what a lot of manufacturers have done, especially in the wake of COVID, has said, look, we need to diversify our supply chains and have multiple sources for some of these high value goods. And we, complete intelligence, [00:36:48] have been talking about regionalization of trade since 2017. We wrote about it more formally in, say, starting Feb of 18, when the steel and aluminum tariffs were put on by the current administration. But we’ve believed for years that we would start to see a re-regionalization of trade, and that cuts out some of the risk associated with supply chains. And some of those costs,

Tony: [00:37:20] maybe transport costs that maybe lower are offset by maybe marginally higher, say, labor or taxes or something like that, either in the US or Mexico or something. So, one of the things that many people don’t necessarily understand is when China came into the WTO in 2000, the US was in the first decade of the NAFTA agreement, the North American Free Trade Agreement. At the time, there were a lot of [00:37:45] manufactured, there was a lot of manufacturing for the US done in Mexico. Part of the reason a lot of factories moved to China was because electricity in Mexico was really, really expensive at the time. Okay. And the electricity in China was really cheap. So, a lot of these manufacturing, especially energy intensive manufacturing firms moved to China to save on their electricity, which was a

Tony: [00:38:13] large factor within their total cost. So, what’s happened in Mexico over the last, I think, four years is laws were passed to deregulate the electricity market in Mexico. So, now you have power in Mexico that’s a lot cheaper than it was 15, 20 years ago. So, the attractiveness of Mexico as a location, at least from a cost basis, is quite a bit higher than it was in the past, and especially quite a [00:38:43] bit higher than it was when firms were leaving Mexico to go to China. So, Tony, you mentioned the impact of COVID on the supply chains. And I want to talk a little bit about something that we have in Bitcoin mining called the supply gap. And it basically, what that is, is when the price of Bitcoin is skyrocketing and is hitting an all-time high like it did back in 2017,

Tony: [00:39:05] the underlying value of these Bitcoin miners really relies on the profitability of those machines. And that is heavily, heavily relies on the price of Bitcoin. So, what we see is that the supply chains, they shrivel up almost. They’re being able to order machines over a three-month period and ends up going out to six months. You won’t be able to get your machines until six months [00:39:28] later. Do you see this central, not centralization, but going from globalization back to Mexico, back to these localized economies? Do you see that helping these massive supply fluctuations or kind of, I guess, events that occur specifically with Bitcoin price and Bitcoin miners, but I guess also globally with events like COVID that really do shock the system we know of today?

Tony: [00:39:51] Yeah, I do. I think that, of course, we’re going to have some difficulties in the early days of it. We’re going to have some awkward moments where things don’t work as people plan that sort of thing. Whenever you have large systemic change, you always have some moments that are a little bit embarrassing and caused you to second guess the decision. We’re going to have those. That’s normal. [00:40:10] But I think over time, what we’re building is a more robust global supply chain. Something like 40% of all manufactured goods are made in Northeast Asia, China, Korea, Japan. And as we have re-regionalization of manufacturing, and that’s to North America, that’s to Europe, and so on, we have a diversity of manufacturing locations. And so if there is, let’s say, COVID in China or in

Tony: [00:40:38] Asia, but it hasn’t hit the US yet, then it’s possible to use additional capacity and say, US or European factories to help meet the needs of Bitcoin miners, right? Depending on what we’re doing, depending on the sophistication of those factories and the capacity of those factories. But I believe that as we have regionalization of supply chains, you have much more robustness [00:41:04] in those supply chains. I also think that in the wake of COVID, so I lived in Asia for 15 years. I just moved back to the US in 2017. I lived through probably five or six pandemics in that time. And so we got a little bit used to it. In the US, it’s relatively new. And I think people here are trying to figure out how to contend with it. And the calibration of risk in the US

Tony: [00:41:27] to pandemics is it’s new, so people aren’t really sure what it means or doesn’t mean. So the global transmission of viruses is not something that’s really going away. So will we have more COVID-like viruses coming out of Asia or coming out of Europe or the US? It’s likely. And so we’re at a point where we have to have regionalization of supply chains. So first, we have robust supply chains [00:41:57] where we can source from the US, Europe, Asia, wherever we want as capacity, as demand, and as costs require. But also, we have the flexibility. If there is one of those events, whether it’s a disease event or whether it’s, let’s say, a war or something like that, we have the flexibility to make stuff in other parts of the world too. So if there was a devastating conflict to

Tony: [00:42:24] Northeast Asia today, global supply chains would be paralyzed. That’s just a fact. And so the sooner we can get regionalized supply chains, the better we’re all off because the risk of a, let’s say a conflict to Northeast Asia, if it ever happens, it won’t impact everyone on the planet as much as it would today. [00:42:50] We definitely agree are seeing that de-risking and big huge news with the semiconductor and TSMC moving to potentially the United States to build a facility, hopefully reducing on that distance for Bitcoin miners specifically. I found it very interesting that you mentioned about Mexico and the electricity prices there to understand that those manufacturers actually had to leave

Tony: [00:43:12] Mexico and went to China because it was too expensive to extract or to complete that manufacturing process. I view Bitcoin mining as a way to almost extracting Bitcoin from the network through a manufacturing process where we’re using these Bitcoin miners and large amounts of energy to do just that. And so I wanted to talk farther about how you’ve worked with clients in either [00:43:36] the natural gas or the energy sectors in the United States specifically and pricing out those

JohnPaul: [00:43:42] markets. And where do you see the future of this industry going? The electricity market

Tony: [00:43:47] specifically and the cost of power in the United States. Sure. So I’m in Texas. The cost of natural gas is very low and the abundance of natural gas is very high. So electricity prices, to be honest, is not really something we worry about here. I know in other parts of the country and other parts of the world, it is a worry. Electricity is something that has always been very regional [00:44:13] and it has always been very feed stock specific if you’re burning oil to make electricity or coal or nuclear or whatever. And you really have to look at that blended cost. But in Texas, we’re looking at a lot of natural gas to fuel our electricity. So not that much of a worry for us. And in this region, it’s not that much of a worry. I think in places like Europe where

Tony: [00:44:38] they’re net gas importers, I think it’s more of a worry. And there’s always a lot of discussion around importing gas from say Russia or from the Middle East or from the US. I think they have an abundance of choice there, but it’s relatively more expensive there than it is say here in the US. I think in Asia, you have a lot of imports from the Middle East, particularly places like Qatar, [00:45:03] these sorts of things for natural gas. China uses a lot of coal, something like 70 plus percent of their power generation is from coal. And it’s really hard to to wing themselves off of that. Japan is a very large LNG and natural gas importer because they shut off their nuclear power after the incidents in 2010 or 2012, sorry, with the reactors, the Fukushima reactors. So it really

Tony: [00:45:33] all depends on the local power generation capacity and feedstocks. But I think generally, we’re not necessarily seeing a world where hydrocarbons become all that expensive for quite some time. When we look at what COVID did to demand, the demand destruction that COVID brought about is pretty shocking. That applies to industries and then applies to consumers. So we don’t see, say, [00:46:00] oil prices or natural gas prices hitting, let’s say, the highs of 2008 for quite some time. And since they are relatively global commodities, although there are differences in certain aspects of them, it also pushes down the prices, let’s say, in other parts of the world, say, the Middle East and so on and so forth. So we don’t see electricity prices outside of, say,

Tony: [00:46:23] regulatory impacts or things like fixed investment requirements. So let’s say there’s a regulatory requirement that a power station can only be, say, 20 years old. That’s a significant cost that would add to electricity prices. But other than that, it seems to us that the feedstocks, although we don’t necessarily expect to see kind of negative $37 oil like we saw in April, [00:46:48] we don’t necessarily see energy price inflation coming anytime in the next, say, 24 months. And if we look at things like gasoline, I know this is electricity, but things like gasoline, prices are down, say, 30% from where they were a year or so ago. And they’re expected to remain that low, at least for the next six to 12 months. So it’s not just electricity, it’s also gasoline

Tony: [00:47:17] or petrol as well, where because of muted demand, prices will remain relatively low. I think that’s great news for miners in the United States. And I really cross the world as more and more energy generation comes online. We’re seeing that that cost to produce coins is continuing to get cheaper, which allows miners here in the US to compete, if not beat miners in [00:47:43] China on the cost per kilowatt hour. Tony, was there any other trends that you guys are focusing on right now in regards to your investment portfolio analysis that you wanted to highlight on the show today? I think there are hundreds of trends for following. But I think we’ve cut both of the main ones. I think really it’s understanding risk of any asset that we follow or our clients

Tony: [00:48:10] follow is really, really important, whether it’s cryptocurrencies or whether it’s oil and gas or whether it’s, I don’t know, the SMB 500. Understanding the risk there is really critical. We’re always trying to figure out how to balance the risk and opportunity associated with the assets that we forecast. And that’s, I would say for any of your listeners, that’s the really critical part to [00:48:35] understand. So we could pursue this down any avenue. And I’m sure we can talk for another hour on just about any asset. So I really appreciated the time today. It’s been a fantastic discussion. Thank you very much. Yes. Thank you, Tony. It was great to have you on. I want to offer you the opportunity to have any questions that you want to ask me about Bitcoin,

Tony: [00:48:57] specifically, that you want the audience to make sure that they hear anything that’s on your mind. No, I guess what I am curious about Bitcoin is, we saw a bump in 2017, I think largely driven by broad awareness or a more broad awareness of the opportunities in Bitcoin. What will drive the next bump in Bitcoin or crypto value? What do you see driving that next rise, [00:49:27] let’s say 30 to 40 to 50% rise in the value of cryptocurrencies? So the way I view the cryptocurrency marketing really Bitcoin specifically is, I’m all about the stock to flow ratio and how that Bitcoin is created. So when that having event occurs, I got into cryptocurrency back in 2013. So I’ve been through two of these

Tony: [00:49:47] having events now. And when that having even occurred in 2016, we see that it kicks off like a real, almost momentum moving into this space where the cost of creating these new coins is exponentially higher, makes it so that all these older machines have to come offline. And it really does a disservice or really degrades the value of these mining machines. It makes the profitability cut in half. [00:50:09] And so when that happens, I think that there are these, the lack of coins, new coins coming into the system creates the momentum, which is needed to push the price up to those 2017 highs you were talking about, or potentially 2021, 2022 highs, simply saying it doesn’t happen instantly, because it does take a while to get there. But I expect that to happen in the next coming years,

Tony: [00:50:31] not necessarily because of one event, but simply because of the schedule of new coins coming out into the market. So sorry, if I understood you, Craig, are you also saying that the age of the infrastructure that the miners are working on has an impact on the vet? So the replacement cost of that infrastructure also puts upward pressure on the price of Bitcoin? [00:50:52] I would say that exactly. So the fact that we have to replace machines that have less efficiency, so the jewels per terra hash, or how well they can turn one watt of energy into one terra hash of mining power, it needs to be upgraded by 50%. So if you have a machine that was running a hundred jewels per terra hash like the S9, that machine is no longer, and it was just barely

Tony: [00:51:12] making money, that machine is no longer going to be even anywhere close to profitable because of this having event. You now now need to go upgrade all of your machines so that they run at the 50 jewels per terra hash level, or you need to find half the cost of electricity. And that is very hard to do, especially because these facilities are massive with hundreds of megawatts of power. [00:51:31] So that’s what I drive as the underlying driver to this Bitcoin price push that we see every four years. And if you look back on the chart, it happens every four years, simply because the miners place such the one of the biggest components of the ecosystem, there’s about $5 billion in mining awards today, every year. And that’s a huge driver in a relatively small market where Bitcoin is

Tony: [00:51:52] currently sitting. Interesting. So that replacement cycle, like you said, and this is a question, it’s not a statement, that’s about every four years, give or take. Every four years, give or take, you either have to replace your equipment with newer machines, which now you’re waiting in line, because everyone else in the whole Bitcoin dollar has to do that, or you’re moving to power where [00:52:11] it’s half as expensive, but all miners are always searching for the cheapest power. So that’s something that’s always occurring. Okay. So with the kind of the supply chain hiccups that we saw with COVID, does that push that replacement cycle back? Like, is that replacement cycle being pushed back by six to nine months? Or is that, do we have a pent up kind of inflation, meaning,

Tony: [00:52:42] do you believe that the value of Bitcoin being driven up will last for longer because of the supply chain issues we saw in COVID? So with this, definitely with the supply chains issues in COVID are shipping rates. As you mentioned, those increased dramatically. It affected how fast machines could get out. It actually caused Bitmain and some of the other major manufacturers to delay their [00:53:06] shipping by two or three months. So if you were to buy a batch to be delivered in November, it still hasn’t been delivered. So there is that pushback and we’ve seen that greatly affect the market regarding the deployment of these machines and kind of scaling with the recent Bitcoin price while it rise. New machines are very hard to get. I would say about maybe 10,000 to 15,000 new

Tony: [00:53:28] machines per month are coming to the US and that might be even on the higher range. So that’s about 50 megawatts of power per month coming to the US and coming out of these factories, which is only $50 million worth of capital. So we have huge constraints on the semiconductor themselves and making those mining machines. And when the price of Bitcoin even jumps up like it has over [00:53:50] the past couple of days, up to the 13,000 mark, that’s going to create even more external pressure, even more interest in mining, which makes it even harder to get those machines and will push out the timeline even farther. So yes, it’s a huge issue when it comes to supply chain management because of COVID and the Bitcoin price increasing investors appetite to get exposure to the space.

JohnPaul: [00:54:08] Fantastic. That’s really interesting. Thanks for that. Of course, Tony. Well, thank you for coming on. I appreciate it. And I’m glad we’re able to have you on. Thanks again, Tony.

Tony: [00:54:17] Thank you. Hope to speak soon. Have a good day. Thanks, JP. Bye. Bye.

[00:54:21] I hope you enjoyed today’s episode of Digital Gold. Be sure to subscribe so you’re notified when the new episode drops. Don’t forget to leave us a five star review to support our journey to

Tony: [00:54:31] become the number one crypto podcast. Thanks so much for listening. And until next time, mine off.