ARKG, AI Investing, And Stablecoins | August mARKet Update

0:00 Hello and welcome to the August market update webinar. My name is Dan White, one of ARK's associate portfolio managers, and I will be moderating today's call. This call was pre-recorded on August 15, 2025. With that, I will now turn it over to ARC CEO and CIO, Kathy Wood. Kathy, >> thank you, Dan. Uh, and thank you everyone for joining us. Uh, just a quick market update. As usual, um you can go to our in then uh podcast and and YouTube video to get more detail. Uh but just to update that a bit, um it does seem like the uh outlook is uh evolving the way we expected. Uh GDP growth is

0:46 slowing. I think we're hearing a lot more stories about a weak consumer. uh uh the labor market uh numbers were terrible. Uh especially the massive I think it was a 257,000 downward revision uh that has caused a lot of drama uh in the administration. Uh and uh and yet that's very typical when we are entering or when we're in a recession I should say. Now we've been saying rolling recession for some time. We think we're coming actually to the end of it. Uh employment is a lagging indicator. Uh so this does not surprise us. Inflation very interestingly the CPI

1:32 and PPI which uh did come out after in the no. The CPI is behaving very well and the PPI is not uh it is more hit uh it has been hit more or boosted more I should say by the tariffs. Uh but what I think this is saying is that the consumer-based companies cannot push along these cost increases and therefore their margins are suffering if they try to raise prices. uh they're losing units and uh that is very damaging to to margins for these manufacturing companies. Uh and the other thing we know from the CPI is

2:18 we're going to have at least a one-year tailwind uh from rents uh that the rent increases decelerating and actually then declining. Uh that should be get getting in for the next year. Uh so we've got growth slowing, inflation slowing. If you look at trueflation, uh which is the real time indicator, uh focused on thousands of items in real time that is at just about 2%. Um it had been as low as 1.78 recently. So the tariff increases have taken it to 2% on a year-over-year basis. Uh so again a surprise on the low side of expectations

3:06 um uh given the impact of tariffs. Uh then we come to interest rates. Uh interest rates have been very well behaved even after the PPI uh came out much stronger than expected. Uh interest rates uh yes they went up a bit but they're still quite low uh by even recent historical standards. Uh especially on the long end uh if you look at the yield curve on the short end so the 2-year Treasury yield versus the three-month Treasury yield um it still suggests that the Fed is restrictive uh and maybe increasingly so. So, we do think uh that there will be interest rate cuts uh in the month and maybe

3:54 months ahead. Uh if we're right on inflation, we're going to see quite a few interest rate cuts. And then finally on profits, it's interesting uh while the consumer companies are having uh trouble, uh profit growth has held up pretty well on balance. I think the S&P earnings uh are running around 10% on a year-over-year basis. Now, how can this be when nominal GDP, which includes both real GDP and inflation, uh is slowing down? The difference is productivity. And so, we're seeing uh innovation, especially AI, help companies uh become more efficient and more profitable. Uh

4:43 so very interestingly when we think about productivity we we we can see companies turning it around in many ways. It can drop directly to margins and it seems like that's what's happening right now. But over time we do think it will be dissipated in other ways or turned around in other ways. better said uh it could be that uh it'll turn around in lower prices uh or increased R&D because of the massive opportunities evolving now uh in the world of disruptive innovation or in perhaps workers comp uh especially in the more competitive industries. So, um,

5:30 we're looking forward to a very interesting, uh, next year. Uh, we actually think we're going into a a a b a boom, a productivityled boom. Uh, and it will, the boom we think will be even stronger if prices start falling, which we think they could, certainly in the areas of innovation around which we've centered our research and investing. Uh so with that, Dan, take it away. >> Yeah, thank you Kathy. Uh we'll be getting to that investment question very soon, but but before we get into the Q&A, uh these questions are submitted through our clients, our distribution partners, and sent to our funds X account and Instagram account at arc-f

6:18 funs. So if any listeners have any questions, please uh reach out through those channels. And this first one, Kathy, as promised, coming back to you. So, uh, how do you structure your investment process to adapt quickly to the rapid pace of AI while maintaining a long-term portfolio focus? >> Uh, well, this is a very interesting question. I I founded ARC to focus exclusively on disruptive innovation technologically enabled. Uh so we are set up for this environment. We expected rapid growth. In fact, one of our uh taglines is invest at the pace of innovation. Interestingly,

7:05 uh, and as many of you know, if you follow our research, which we do give away, um, we we start, uh, our analysis at the top, trying to size new markets based on the learning curves associated with new technologies and how quickly uh, costs are falling because of those learning curves. And we we estimate how long it will take for these technologies to scale across sectors or industries uh during the next five years. Uh so we're set up for exactly this kind of environment. And then our the bottomup steps of our process uh are are focused

7:51 importantly on uh and and scored for the first four. There are four six metrics. The first four are scored um uh company, people, management and culture. So that's the first one. Do we have a visionary leader who will stand up to shortterm oriented shareholders or venture capitalists uh and invest aggressively uh to really come out ahead and capitalize on these super exponential growth opportunities? Uh execution. Uh are these companies spending enough on R&D as a percent of revenues and are as important maybe more important is are they spending uh their

8:37 R&D dollars in the right places and again we have a point of view informed by our top-down analysis on on the right uh areas for R&D focus uh another one is uh moat uh this is very important the moat many people think is um very importantly about patents and in some cases especially in healthcare that's true. Uh but today uh it seems the moat or the barriers to entry uh are becoming or are focused more on speed of execution and uh and moving down the learning curves uh meaning costs coming

9:22 down uh as rapidly as possible to scale across sectors. Uh then there's product and service leadership. Uh, of course there are a lot of polls and social media is full of uh responses to different products and services. So we can learn a lot there as well. And then of course valuation is critical to us. Um uh we've uh gone through valuation a number of times but we have a five-year investment time horizon. And we assume that our companies um or the stocks of our companies their multiples uh enterprise value divided by adjusted EBITDA uh that they

10:09 decay or compress to uh nearly a market multiple in most cases or to the average multiple over uh the re last 10 uh for for a given company. Uh so we assume multiple compression if uh our portfolios are selling at a premium to the market. Uh and they are now so that uh compression is an automatic part of our valuation. And then uh finally thesis risk and very often that's uh government regulation. And I think one of the most wonderful things happening today is the amazing deregulation that's

10:55 taking place uh in this uh in this administration. They've made it a primary focus and they're running hard with it and uh even in healthcare. So we're pretty excited about what's going on. >> Thank you Kathy. Uh our next question here is for Frank. So, Frank, uh, is there the potential for large large language models to become commoditized and, uh, kind of a followup to here. Is this a winner takemost market? >> Yeah, it's a it's a great question and we get it, um, pretty often, so happy to happy to walk through it. And I think when you when you think of the the model builder companies um what we've seen is is almost you know simultaneously or two

11:40 two opposing trends at the same time and we're observing more competition uh while also in other ways uh a narrowing of the field uh and so to explain the former uh for the past a large majority of the last two years it was really open AI setting new frontiers and then the other competitive model building companies like Google, Anthropic, XAI catching up to the frontier at some lag. And what we've seen is that the the pace at which companies catch up and the the the size of the leaprogging as new models come out has become much more competitive. Uh and all four of those companies that are really playing at the frontier kind of trade off who has the best model and who's you know having the next increment that's better across

12:27 different domains. And so in that lens whereas open II have been on the model side uh really pushing the step function changes it's gotten more competitive. Uh we've also seen on the you know open- source model side uh China become increasingly competitive first with Deepseek but now models from Alibaba and other companies uh are increasingly competitive with models in the US both uh models from US companies I should say. a lot of these open source models coming out of China are ultimately run on US clouds and used by US companies um which is an important nuance but um there's there's kind of competition from that front as well now when I say the field is narrowing uh there are a lot of companies over the last year or two years that raised a lot of money uh that uh did not train a good enough model did

13:12 not get enough momentum to compete at the frontier so you have companies like inflection AI and stability AI that had uh turnover in the teams a pivot or a total restructuring uh and the list goes on through a longer tale that can no longer compete with the capex required uh with that to train a frontier capable model. Uh and so you know at the same time as you have growing competition of the companies that are at the frontier the frontier is separating from the rest of the pack. Um, and as you're thinking of these kind of offsetting dynamics, uh, in answer to this question, I think it's important to remember that the foundation models themselves, the large language models are just one layer of the stack of an AI solution. And you really when you look at the the products and services that AI

14:00 is embedded into, you have a kind of a much more nuanced uh, uh, look depending on the use case and the user. Uh so on the consumer side uh you have kind of the vision of of a personal agent uh that can do all of these tasks for you and that's the market that chatbt is really despite maybe not having daybyday the top model or the step function changing model that's the market that they're really leading in with over 700 million weekly active users. Um the fastest growing interestingly is Google. Uh which is why if you're paying close attention we've uh acknowledged that as well as the strength in in Google cloud uh and added Alphabet into strategies like Arkw. Um and uh then shifting to the enterprise side I think you're going to have just like in uh cloud uh software

14:47 today a mix of horizontal platforms and vertical specific platforms. We're seeing that develop where the enterprise versions of the AI agents are one example of a horizontal productivity tool. Uh companies like Palunteer that are building the critical data and application infrastructure to make make AI useful within enterprises is another example. um developing in this horizontal category and then vertical c uh vertical categories and winners like the coding agent space which is one of the fastest growing areas of software where you have companies like cursor and replet that have grown their revenue base by over 10 times uh in less than a year. Uh so growing tremendously quickly and focused on a specific niche which in this case is autonomous software development. Uh, and so I think you'll

15:33 see specialization and different companies winning at each of these tiers and they're picking from the best model available. And so OpenAI may be positioned to use their models the best because they're creating them and that creates an advantage. But a company like Replet is switching from the best coding model that's available. um opening it they just released uh GPT5 uh actually announced that cursor the most popular coding agent uh is switching from using anthropic to uh GPT5 uh to support their product and if Google or Anthropic comes out with a better coding model than uh than GPT5 they will probably switch back. Um so depending on the uh use case you will have more commoditized uh use cases and others that are are more differentiated.

16:18 Another important nuance is that outside of the maximum intelligence or raw performance of a model, there's more nuance characteristics like the personality of the model and the level of personalization that you're building into the product uh that leads to a sticky product over time. And I think that's what uh OpenAI has done well in adding memory and more personalization to chat GPT and they can use their you know base of hundreds of millions of users to improve that product over time. uh but also where they got blowback with launching GPT5 because it had a new personality than uh the prior generation model that turned some users off. Um so we're really learning new dynamics of of the competitive vectors here. Um and I guess just the last thing I'll say in summary on the um the the the enterprise

17:04 side of things, there'll be horizontal plays, there'll be vertical plays. Uh but we think a majority of the value will be in these horizontal plays, which is why Palunteer is a is a a high conviction name for us. Thank you, Frank. I uh Alex Karp mentioned something on the earnings call with uh you know, kind of on that, not just the LLMs. What was that again? >> Uh well, I think he said something along the lines of uh LLMs don't work without Palunteer and don't work without the ontology. Um which in in a business context is really providing not just the model that is intelligent, but providing it with the right data at the right time and the right tools to actually affect change within an organization. that is really what the Palunteer platform is built for. So it makes sense that they're, you know, seeing the adoption

17:49 they're seeing and being able to deliver these amazing examples of value and productivity for their end clients because the clients all want to use AI and Palunteer really gives the tools to to make that possible and integrate AI into your existing data assets um uh inside of organization. >> Yeah, I'll just add something there. It was a very bold statement toward the beginning of the quarterly call and he said large language models do not have common understanding. Uh they need Palunteer if they're going to work within an organization, right? Something like that. >> Yeah. >> Great. Uh thank you Frank and thank you Kathy for this next one. This uh next question is for the multiomics team. Um

18:36 so how how are broader government macro and and FDA factors influencing the health care sector and how might uh those dynamics shape and shift ARKG's uh path forward to closing the performance gap with the uh rest of the ARC suite. >> Yeah, thank you. It's a great question. So from a big picture standpoint, we see a few major forces here shifting that could support support ARKG's performance. So thinking on the macro backdrop, Kathy has highlighted several elements here earlier. But to hone in on a point here, high interest rates and tight capital have been pressure points for healthcare innovation. So when you consider many of the companies in the space are still years from profitability

19:21 and rely on outside funding to advance their programs, having the cost of capital improve is really important. This is meaningful for early stage innovators, making it less expensive to fund their multi-year clinical programs and helping extend their financial runways. Thinking on the level of strategic activity, well, we have to remember so large pharma here is facing a massive patent cliff by the close of this decade. So over 200 billion in annual revenue at risk by 2030. Now historically, large pharma will often look to innovative biotech companies to replenish its pipelines. So thinking in early stage biotech, these transa excuse me, transactions are a key form of strategic price discovery. So they set real world benchmarks for what innovative platforms are worth and help public markets calibrate valuations accordingly. And a relevance here to

20:08 just highlight. So back in June, we saw Eli Liy acquire BER therapeutics for 1.3 billion. Verb is a biotech company leveraging beam space editing technology to develop invivo single course therapies aimed at lowering risk for cardiovascular disease. So for the first time we're seeing a major pharmaceutical company making a substantial bet on invivo gene editing targeting a common disease and this is really important validation for gene editing space and I would not be surprised to see this kind of strategic activity continue and the third one is really important here that I would highlight and that's the new FDA that we're entering. So D Dr. Marty Mari uh was appointed this spring as FDA commissioner and he's been very vocal about ushering in a more modern and responsive agency. So the state priorities include gold standard science and common sense and regulatory

20:54 decision-making. So this is looking like partnering earlier with innovators, integrating more modern tools like generative AI and big data into the review process, for example. And the ultimate goal here is to make the FDA more efficient, streamline regulatory pathway, and reduce regulatory uncertainty. These are really key steps that would be expected to shorten development timelines and lower capital burn for companies. So we can't promise outcomes, but we believe these shifts are uh important and create conditions within ARKG and in our multiomics exposures for our ARK are certainly right for playing catch-up. Uh just to put an ex exclamation point on um on the FDA uh this I think this has been the biggest surprise in in terms of deregulation and how quickly it

21:41 is happening. Uh and another example is uh the FTC which in the last administration really did not allow uh much in the way of M&A especially not by large companies like large pharma. Uh that is changing as well in this new administration. So those are two examples >> and m maybe just to add u one more aspect here is that we know that this administration has been very kind of focused on the u China uh as the potent potential competitor and across the different u sectors and you know I think in the

22:27 biotech and the healthcare you know they had been really ramping up uh on China side. So, uh we have definitely noticed a trend that many of the pharma companies actually now go and acquire assets in the China versus here because they are cheaper. uh and so they they're definitely more efficient in actually bringing to phase one or phase two the assets um and so the far will go there and acquire the asset and so we have to kind of ensure and I think this administration is in line with that that the center of innovation is actually staying here and is not moving to China which at the moment uh it seems it could be a a possible option. Thank you multiomics team. Some exciting

23:13 tailwinds uh for the entire genomics space and uh Nemo I actually have one more for you too. So Tempest AI is uh one of the larger holdings in the ARKG fund. Uh they recently had a earnings report. Could you uh please just provide the listeners an update? >> Yeah, absolutely. Uh happy to. Uh so Tempest had a really good uh second quarter. So the revenue was about 340 million and where the genomics was about 240 and the data and services were 72 million. Uh you know just for reference data and services growth is 35% year-over-year. Uh interesting thing here that embryomics their acquisition had a really good quarter as well. Um

23:58 which was not expected. Uh and so just a reminder, embryomics basically going to focuses on whole exom sequencing and whole genome sequencing in this in the mostly for the rare disorders. Uh uh so they had been doing doing really well and I think uh you know what is this signaling to me is that u tempos acquired amigenomics because they actually wanted to be exposed more in the rare disorders space and maybe even the newborn uh sequencing eventually. uh and again kind thinking this all big picture what Tempest is trying to build is the kind of ultimate like true powerhouse when it comes to diagnostics you know they they started first cancer they have a huge database that's their kind of biggest franchise but then they have been developing on the cardio side

24:44 algorithmic tests they have narogenomics kind of test uh with army genomics they they want to enter rare disorders to so I think they want to be like a one-stop shop all for the for the for the for diagnostics and it's not only kind of diagnostics how we are used to that you know you do a test and then there's like a binary decision but it's really thinking about this uh intelligent diagnostics how they call it or like how I think it's more like a continuous diagnostic where they want to be able to kind of order life lifetime they can kind of diagnose different things for a patient so I would say that you know in our mind they're the kind of only company that's really uh taking this AI first approach when it comes to diagnostic. A lot of our portfolio

25:30 companies is is using AI in the diagnostics but Tempest is really taking this to another layer and actually coming first from that angle and you know in when we were talking I think last week I think we talk you know in my mind it came Tempest reminds me a bit of the Tesla you know Tesla is primarily AI company which is using the vehicles basically to collect the data to make the AI model. So this is how tempest approach it talking think about diagnostic all these tests which are helping the patients and doing the things they for them the vehicle the way to actually collect the AI collect the data to train the AI on that uh and and kind of build a foundational model. Yeah. And uh yes, I guess uh again another exclamation point here. Uh I think Tempest

26:15 probably is the company we've been looking for uh that will be really the the information backbone of the US economy. Uh right Nemo thinking about it in those terms and AI is win or take most. We talk about Tesla that way and in this way we think this company probably is the closest uh or has the highest odds for winter take most. >> Yeah. Yeah. It could be uh absolutely you know um so again we here we are thinking about where diagnostics is truly being preventive predictive um and you know where actually the big the biggest value that's is being created is actually in collecting the data from the

27:01 patients and so campus is really positioning themselves there to be uh at the center of that. Um so we think that this really you know and you know if you think that how healthcare is a big spend right um and so and having a temp as a center of that because it's kind of collecting the data and then enabling this kind of preventive and predictive medicine it really puts them as I said at the center of the of all this sector. >> Thank you Nemo and Kathy certainly a really exciting company in the healthcare space. All right, these next questions are Sam turning to you on the robotics and autonomous team. So according to your modeling and your team's modeling, um what does timing look like for humanoid robots uh at mass scale and adoption?

27:48 >> Sure. I think people are very excited about mass scale and adoption. Uh we think this is probably more in the five to 10 year time frame. And to contextualize this, it's not about making them. Um, it's really about what can they do. And there we're seeing tremendous improvement on the AI front. Um, there's a number of companies looking to produce in the thousands in the next year. Elon and Tesla have come out saying, you know, they want to get to roughly a million in five years. But if you think about what is mass adoption, you know, at scale, there's the potential for billions of these, right? So it's like a million is impressive in the scale of industrial

28:33 robots where they stand today where you know just 500 roughly 500,000 are being sold on an annual basis. So this is going to completely transform the robotics industry. Uh but the potential here is so large that really getting to mass mass scale uh is going to take time. Uh at the same time there's exciting uh developments that are being ticked off in the near ter in the near term. Uh useful applications just in factories uh you know other demonstrations of folding laundry or or some task at home. Uh and so it's going to be a very exciting future on the way to true mass scale which is going to we believe transform what the economy even

29:20 means. Um but that will take time. >> Yeah. So uh moving from the cool and exciting dance videos to real functionality across different industries. >> That's right. Don't sleep on the power of dance, Dan. >> Yeah. I I I need some help there. We all know that, too. Um, so the next one, Sam, I'm going to stick with you here in the Eevee Tall space. Um, could you please compare your five-year outlook uh for Joby and Archer, two names found in uh the ARC funds? >> Sure. And I'll take I think this is a good question just to tee it up because on one side the question gets it right, I think, and on the other side uh confuses it where it's like where is this question on the nose? From our

30:06 research, these are the two companies uh that are pushing the frontier of EV toll. uh where do I think this question kind of frames it incorrectly is you know p putting them one against the other right this is a industry that's at a nent stage uh that should have multiple winners and so I don't think it's it's one versus the other and when we look at both of them you know they're tackling things in different ways right Job's testing in different ways they're hoping to own and operate at the start uh even this morning you know they just did a test flight between two public airports. Archer uh is moving very quickly. You know, they started later but have caught up in a tremendous way. They're looking to sell aircraft, not

30:52 necessarily operate them. Uh but both of them, you know, in the next five years, I think are on a similar goal trajectory, which is to commercialize in the next year in the Middle East. Uh then get that approval in the United States, right? Archer, you know, hopefully within the next five years as the official uh EV toll provider in the LA Olympics will make it uh truly magnificent games where you're not stuck on the, you know, eight lane highway, but you can actually see uh the the sports you want to. Uh and at the same time, you know, we would hope that Joby is approved in the US as well, and this will relieve all of the increased traffic from the autonomous vehicles on the road. Um just one one more uh

31:40 difference I think is the exclusive relationship between Andreil and Archer. So it is a a defense play as well. >> Thank you both. Uh Sam and Frank, I think this next one's for both of you here. So uh you know big headline the other week with Tesla shutting down Dojo. um what does this mean for the company and for its shareholders? >> Yeah, I can I can start on that. And just to give a little bit of background, Tesla has been building their own custom chips. Uh and they've had them in production since 2019. They were talking about it and working on it years before that. Uh and really started with in the

32:25 car uh where they've been very successful. They're on their second generation chip that's in cars now. Um the the naming is a bit confusing, but it's it's the AI4 chip. It's their fourth generation of hardware. They used uh Nvidia prior to swapping over to their own custom chip uh in 2019 with uh what they called hardware 3. Uh and now the current generation chip is AI4. Uh and that's uh the reason to build your own chips is because it allows you to if you have a very specific workload uh uh improve the performance and have greater uh a fine fine level of control over um how the chip is built purpose for uh specifically for the workload. Um and it also allows you to move at your own timeline versus the timeline of a vendor

33:10 like Nvidia for example. So that's been very successful in the car and Tesla uh seeing the success of that uh starting in 2019 but really you know talking about it more publicly in 2020 and 2021 um launched efforts to build dojo which is their custom chip for training uh in the data center. So this would be a competition with Nvidia for example. Uh and I think the announcement of shutting down dojo they they did dojo one and have been working on dojo 2 uh is a is a recognition of a few things. One is a consolidation of resources and a focusing on uh their custom chip efforts on the incar chip and a realization they had real competitive differentiation there. uh whereas in the data center side keeping up with the pace of change

33:56 of Nvidia uh is hard and if you're thinking about consolidating resources consolidating down into the chip that is really differentiated in the car where the end workload is working makes sense. Uh and it's also because they've been using that chip in the training process itself. Uh so Tesla has I think they've said over 20,000 of uh hardware 4 or AI4 chips in the data center that they use for part of the workload um that they use to train their models when they're taking raw video data into the data center and doing the perception processing before it gets fed into the GPUs for for the uh downstream parts of the training process. And what Tesla announced um shortly before the the termination of the dojo project is that the AI6 chip uh so not the generation

34:42 that's going into cars next year which will be AI5 but the generation after that uh their intention is to make it as good as possible at both inference in the cars and training in the data center. Uh so we'll see as we get closer to that timeline. It's going to be probably three plus years out uh what they're able to do uh from a design perspective to make it efficient in both use cases. Um we've uh the other piece of news related to this is that uh Tesla announced that they are going to be working on the AI6 chip with Samsung on their USbased fab in uh Texas and they have a special agreement where Elon will be walking the floor of the fab to maximize efficiency. So this is another uh angle of of vertical integration and specialization going down to the uh to

35:28 the chip manufacturing level itself, not just the design level. um that should yield uh if successful uh even more differentiation over time. >> Great. Thank you, Frank. Uh this next one, moving into the the crypto sector, lot of uh kind of news and excitement around this area recently. So, Lorenzo, a new name that's found its way into the ARC funds, BitMine Immersion Technologies. Could you walk us through the uh investment thesis for this name? >> Yeah, of course. So uh to get a to give a bit of background so Bitmine immersion is led by uh Tom Lee and um right now it's the biggest Ethereum DAT so digital

36:15 asset treasuries and these are companies uh that are trying to accumulate um as much ether as as possible. And so um really when we looked at the ways to get exposure to uh Ethereum um right now the main uh problem with the ETFs are is staking right the big difference for Ethereum and Salana compared to Bitcoin for example and Micro Strategy um these networks have a native yield so for Ethereum it's about 3% right now for Salana it's closer to four to five. And so the ETF right now are not allowing uh um staking. And um

37:02 even if if that were the case really uh because of liquidity and how these networks uh work um right now because of the high usage of the of the network, you have about two weeks to stake or unstake uh um Ethereum. And so this really um makes it very difficult for uh ETFs to uh stake the majority of the assets. So we think they're probably going to be able to stake between 30 and 50%. Which obviously lowers the the net yield. And so we think uh Bitmine and other DATs with the permanent capital structure are actually a very efficient wrapper and and uh uh are actually

37:47 better. And so uh you know regarding BitMine in particular for uh the the investment thesis and and and the outlook, I think really the the focus for the next I would say a foreseeable for the foreseeable future is going to be to accumulate um Ethereum. And so when you look at just to give an example when you look at the window from July 13th to August 13th um Ethereum uh went up 60% in that 31 days window and Bitmine Immersion was able to increase their ETH per share by

38:32 130%. And so actually the way Bitine is issuing shares at an MNAF premium to buy back Ethereum has actually been very accreative to to shareholders uh in that in that time frame. Um and so if you look at long-term uh with the yield right now Bitine has about 1.2 2 million ether. That's about $5.5 billion more or less depending on price. And so the yield is about $150 million yearly, which is very significant for a company that is very lean. If they were to, you know, use that in a future for other purposes, they would obviously have to to sell Ether, but that is pretty

39:17 sizable. And you know, we think the the use of ETH as collateral within DeFi, um there are many use cases and things they can explore. And so, um we we do think it's a winner takes most, uh uh right now in in the market. And so, we're excited to, you know, to to back and invest in in Bit. >> Great. Thank you, Lorenzo. Uh one, uh next one's for you too as well. This our final question in the stable coin world. So uh and maybe you can explain sort of the acronyms here as well as as who owns these. So how are USDT and USDC different? >> Yeah, so USDC is the stable coin issued

40:04 by uh Circle and USDT is the stable coin issued by Tether. Uh in terms of the technology really underlying it's extremely similar. These are smart contracts on uh on Ethereum rails and on Ethereum and Tron mostly for uh for tether. Uh so technologically it's very similar. The biggest difference is I think the user focus the geography and really the backing of the of the collateral uh right now. Uh so between both they have I think more than 90% of market share. So, they're really the the two biggest stable coins issuer. Um, USDC uh is backed by a short-term US

40:50 Treasury bills and USDT um is also backed by short-term uh securities uh but has a portion of the collateral uh that is invested in um precious metals, corporate bonds, Bitcoin and and secured loans. And so right now in terms of the the collateral that is the the biggest difference and in terms of of uh use cases and really geographies um USDT is really dominating emerging markets. So there are extremely present uh in Asia, Africa and Latin America and USDC uh especially with now the the Genius Act is more focused on uh um on developed

41:38 markets and so uh Europe and and the the United States. So those I would say are the are the the main differences. >> And I'll just add uh I'll add one thing to that. Um the There has been controversy over uh the collateral backing USDT. Uh as Lorenzo said, it's a combination of treasury securities as well as Bitcoin, gold uh secured loans. Uh as as Lorenzo said, uh I think and please correct me, Lorenzo, 80% 75 to 80% is US Treasury securities. Uh, and one of the interesting nuances here that we've been exploring is what does stable coin mean?

42:27 Right now with USDC, it's effectively pegged. It's well, it's pegged to the dollar and backed by US treasuries. uh USDT with uh the addition of Bitcoin and gold has a little bit of a different flavor of quote unquote stable in that uh uh it's geared almost for a stable purchasing power um way of thinking. Now uh they're they're they're both going to evolve in in different ways. So we'll see what happens. um and we don't know what they're going to do, but that's just a a little hint that there is that difference between them evolving.

43:16 >> And for for anybody that's interested in diving into that difference in detail, we did a podcast with uh Paulo from Tether and the Art Laugher, which I think is really interesting to uh tune into if you want to explore the topic more. >> Thank you, Frank. >> Great. Thank you all. Uh, and with that, that will conclude today's webinar. We will briefly show performance statistics that can also be found on arc-funds.com. Thank you to everyone who listened to today's call and once again thank you to the ARC