Welcome to Lokotech Group's midsummer update and dockside chat. My name is Yngve Johansen. I am the Chairman of Lokotech Group, and today I have with me Ola Stene-Johansen. He is the CEO of Lokotech Group. We're going to get some updates from him. We're going to ask some questions that we have received from investors. First, let me extend my thanks to Pareto Securities and Investorweb for setting up this event. We are having this event in location of launching new investors relations pages, also simply because we have had much that have happened lately for the company. Here we are at Aker Brygge when most people are on vacation, perhaps even in Miami, getting ready for our football match. It's a really nice day here in Oslo. There's a lot of sun.
We are next to Aker Brygge, well, it's warm and well compared to Trondheim, where I'm normally sitting. It's raining up there. With that, one last thing. I'd like also to thank all the investors that have contributed and sent in questions. Some of the questions I have reformulated a little bit. I have merged them for the sake of efficiency, so if you don't recognize your exact question, that is the reason. I believe that we will get across the information that is needed. Okay, Ola. With that, tell me more about what's going on. Let's start perhaps with Lokotech Group's largest investment and largest project.
Sure. You're referring to our ASIC-
Yes
our application-specific integrated circuit.
Yes.
In layman's terms, a chip. It's a dual-purpose ASIC. The main purpose of this chip was to hash Scrypt. Scrypt is an algorithm that secures multiple blockchains. Second-largest proof-of-work algorithm currently, it's a memory-intensive algorithm by design. We have tapped out the design. The mask set is progressing nicely, as expected. The wafers have started processing, so right now for us, it's time to prepare and scale and get ready for receiving the first physical silicon. It's also a dual-purpose ASIC, so it has design features that fit very well to AI edge inferencing. Right now, I think the next milestone is the power simulations. It's a simulation of the actual submitted design as opposed to the EDA tools that are used in earlier stages of chip design.
It's all about power efficiency for Scrypt. It's by far the most important metric, although we've included several other quality features in our design, at least.
When you compare it to our primarily Chinese competitors.
Okay. Let me then just go because there was many questions on this, on the efficiency bit. When you say power simulations, can you elaborate a bit more on that? How is that done now and compared to earlier? Foremost, when can we expect those results?
The results will be completed imminently, and when I say imminently, I mean we're counting days, not weeks.
Oh, okay.
That's an important milestone that will be met shortly.
These power simulations, they differ from any previous estimations that we've done because it takes a snapshot of the physical design, includes all the improvements and the physical parameters that have been implemented over the last years.
Versus a EDA tool estimation. EDA tools typically are conservative. It's impossible to say exactly how conservative.
It improves the accuracy of what the physical silicon will be able to achieve, and it's a bit cumbersome process in many ways, but it's important for us to verify so that we can complete certain system design questions that remain-
unverified as of now.
Well, also of course, the business case rely on it also.
Obviously, it's the biggest driver for value in.
Yes
vertical, so.
Okay. No, but.
Fingers crossed.
Yeah. Very interesting
The tools are really conservative in our case.
Yeah, yeah.
The developers of these tools, they continue to iterate and update, and they would also like to be accurate.
It's difficult to say for certain before you've actually done it.
No. That's good. You also mentioned that one of the metrics is the power efficiency, naturally. Also, you mentioned some other functions that will make a difference for the user.
Sure. We've primarily marketed the graceful degradation, the fact that this chip has 255 cores that compute in parallel. With a bit of design smartness, you're able to create a product that is more typically Western design.
If a single core fails, the rest of the chip will continue to operate.
Okay.
There's also system design choices that allows us to circumvent a single chip if it's a full chip failure.
What it translates to is reliability and quality-of-life improvements for the clients.
There's also security features in there that might be more important for AI. It's a clear differentiator in that vertical.
We will get back to that.
we'll get back to that.
Yeah.
It's exciting stuff, I think.
This sounds great. What is the maturity then of the Scrypt functionality? How well tested is it? How ready is it? Is it ready to go in the product, or what is the maturity?
I brought a prototype board with us.
Okay.
Everything surrounding the chip itself is completed, right? The firmware, thermals, system design. What I need now is chips to populate this board with, and we should be ready to plug it in and start hashing.
Oh, really?
Okay. Maybe.
The core functionality has also been proven in silicon and has been running smoothly since, what is it? Q3 2024 or so.
Okay.
We know that the logic works, and we're eager to populate these boards.
Okay. That's good.
Components are being stacked up.
Okay. All right then. Let's move then to something else related to the other functionality. Let me just first, I saw there was our recent notice, and some of the investors, they have submitted questions related to that, and they want to know more about that letter of intent we signed with Aprochip our JV to explore the AI functionality of the chip. What can you share on that?
Sure. Obviously, it wouldn't be smart for our shareholders' best interest to reveal any of the negotiation parts.
Okay.
I think the intent is very clear. We understand that small team in Oslo might not have the reach or the competencies to hit the ground running when we receive this first silicon. This JV negotiation is also a preparatory thing where we bring in competence, network distribution, and so forth. The AI functionalities of lower maturity, it's TRL, technology readiness level 3, 4, depending on which ladder you want to market against.
Yeah.
In order to reach higher maturity levels quickly, we think that these counterparts could be a very good fit for us.
Yeah.
Makes the JV more investable, more marketable, and quicker to a commercialized product.
Okay. Also, could you elaborate the product in terms of software for that functionality, and testing of models.
Sure
Use cases.
The AI or even just the AI inference being a sub-vertical of AI as a whole.
Yeah
is populated with serious competition.
To me, it's about finding and identifying the right niche. Where does this chip fit in in the broader market?
We know that it has features that none of the competitors have. You mentioned software, and I think the fact that this chip could be a true agentic AI chip, meaning that it can perceive, decide, and act on that information, and even feed back that data to a broader system that could self-learn or self-calibrate, is the biggest differentiator. It fits somewhere in between the very specialized inferencing engines and the more power-hungry generalists.
There's work to be done, and that's why we started ramping up both hiring and also searching for the right partners.
Okay. Yeah, talking of hiring
There were questions with regards to that too, and you can see that on your own pages now and also on FINN.no and other places that you are looking for new talents. Could you tell us a bit more about that?
Sure. I feel like it relates a little bit back to my letter in the annual report. Lokotech and Lokotech Group is going from being a startup with a very limited spend to a scale-up where you need the manpower and competency to scale properly. We started to search for engineers and site managers, sales and service. As you know, our previous CTO is no longer working for us, so that's also an important role to fill.
How is that going?
We've identified what I think is a correct path. It's a proper recruitment firm that will assist us in identifying the right candidate.
That work is barely started.
But-
I think we've received many interesting CVs and talents.
Yeah, I think the latest count is like 30.
Wow
engineers.
Okay.
Some, a majority, located in Norway. You have both Oslo, Trondheim represented. We also have less interesting CVs, if I'm going to be honest. Junior personnel that I think it's more right to hire a bit later.
You need to start somewhere. Yeah.
Yeah. We have started.
Unsure if we will start interviewing next week.
That's soon. Yeah
It starts soon.
Yeah.
The next phase of that hiring.
Just out of interest, with regards to the AI JV, will you also start looking for people there soon?
Yeah, as I mentioned, the goal is to make Arctic AI a standalone entity that can carry its own weight, for that reason, we will need competent people.
AI competencies are even more specialized than even more general-purpose electrical engineering. That is a search that has started.
I don't have any names for you.
No
Just yet. It needs, at least in my opinion, an experienced engineer or experienced CEO that has had the opportunity to land large projects.
like this before.
Mm. Yeah. No, exciting. That part is really exciting. Okay, there are a lot of moving parts in this, clearly, and it is really interesting to hear the status on that. I am sure also the investors and shareholders appreciate that. Can you share a little bit more on when will we start to see wafers coming out of this process?
It is quite interesting, actually, because we do not have a final out-of-fab date yet. Still sticking to my previous communication that both the wafers and tested chips in packages, they are all due second half of this year. There is no delays on that sense.
the pace is picking up, and I feel like that is also connected with the fact that we are starting to really recruit now.
Mm. Yeah. In one meeting I joined not long time ago, I see more and more people coming into the process, clearly, things are going well.
Sure. You have specialized subcontractors that now are a lot-
Yeah
more tied into the whole-
project, if you like.
How is that for you? Are you excited? I've been doing this for a while.
Do you ski, Yngve?
Yes. I ski and snowboard.
Snowboard. Sure.
Yeah.
What's your preference?
Sorry. At least if it's powder, it's snowboard.
Snowboard?
Yes.
To me now, this timing, this summer is like the season's first ski trip, right? Only that you've been on a six-year lift ride.
Ooh.
The powder is ready. It's there. Just exiting the lift and about to have the ride of your life. Easy to say that I'm excited.
Yeah. No, that sounds great. It's been a long lift ride, but finally ready. No, but that's good. I can then perhaps move to a different kind of questions that we got. One was quite good. It was related to now. Given that everything performs as expected, how are we doing on the rest of the supply chain? Are we scheduled to start pumping out large volumes, ramp up quickly? Please consider also can GlobalFoundries deliver in that respect?
A lot of the value chain has been trimmed and designed so that we're enabling a rapid scalable production. Has to do with wafer probers, probe cards can be duplicated, and you can run multiple in parallel. We're planning the cycles so that it fits engineers' shifts to not make bottlenecks. We also have a good relationship with Arrow. Unsure about their size, but they have to be top 3 components distributors in the world. They are stocking up components that we expect to use.
Assembly is not a issue. The pick-and-place suppliers are used to delivering higher volumes than what we are estimating. It comes down to GlobalFoundries and wafers. Thankfully, they are U.S.-based, so it's a money talks type of a question in the end. Which again, relates to the market. The market is difficult for us to foresee. Cryptocurrency is volatile by nature. If the market picks back up, then the demand is definitely there. It's a question of how quickly can we turn them around and recycle that capital.
Remains to be seen. I'm optimistic. We have witnessed, as you know, competitors that brought just a slightly more efficient product to market, and they take a market share of 20%-25% overnight.
I'm confident as long as we are able to charge the prices we think we will be able to. Yeah.
We're rigging for mass production.
Okay. There is a question on cost estimates. I don't know how much you want to go into that.
I'm not going to reveal any specific cost estimates.
guidance of the gross margins.
General trend is that wafer prices are increasing. We saw a, I don't know, 7% increase this year.
I don't think that's the make it or break it for us at this moment. It's also very volume dependent, so again, relates back to the market.
Mm-hmm. Yeah. Okay. Let's move over to a different question. It's slightly similar to what we talked about five, 10 minutes ago, but it's on the claims on our power efficiency.
I reckon from you that you still think that we are more than 100% efficient than competitors.
Mm-hmm. At least
at least, yeah.
It's gone from most likely to at least, I think now. That.
in a couple of days or so, we'll have a definitive answer, right?
Okay. All right. it won't be a boring summer, I guess.
I hope not, no.
All right. Okay. we had another one. We can move to something slightly different then. End of 2024, we have a notice to the market where we had to share that we have received such interest from a party that we have to notify the market about it.
Basically, they expressed interest in the company even for acquisition purposes.
There was a question on, is this still on the table? Are we still in touch with this party?
We are in touch with them from time to time, not necessarily on discussing this topic-
There are other topics related either via PowerPool or for Lokotech ourselves. They're still very much in the business. They're still executing on their M&A strategies. They had a firm demand to see a physical chip, right? So-
They are waiting on us basically to prove the physical chip.
That's the status.
It seems like there is more than enough that is happening very near term. There was a question also with regards to future development. There was mention of, they were looking on other algorithms or according to migrating to lower nodes.
I think that's a very natural thing to do. If we prove the first ASIC, we should start exploring the second generation. From a capital side, if you're taping out on different lower nanometer, it's expensive. Timewise, it's not nearly as long as what we've seen for this first design.
Finger in the air, 12-18 months perhaps, it could go full speed ahead.
Yeah, other algorithms, I think SHA-256 is the natural next one.
Exactly how we're going to attack it, if we're going for low cost, power efficient or a high-density system, that's two different strategies. We have some promising leads and ideas, but nothing hard right now, no.
Mm. Okay. All right. Well, I guess that is a sufficient answer to that. We can move slightly also a bit over to some people, they want to talk about sales and numbers. One guy, he wonders about when will presale start again?
presales will start when we have the physical silicon. When it's tested and running, then marketing efforts will pick up drastically and while we operate the first boards and potentially sell the hash rate while we certify all the form factors-
we'll take in the next batch, and that will size up the wafer order, which then relates back to-
the next production batch.
important part of that is also finding the yield, and we have estimates, figures that-
How does that yield curve look and how does that impact the economy of the product?
Yeah.
Important details that we'll learn second half of this year.
Yeah. There is another one on why are we not just pre-selling all the time? Why do we have this hold up? Is it related to pricing or?
Yeah, it's a combination of pricing and market dynamics. The vast majority of the agreements that we have entered into has a fixed markup, right?
We don't want to risk bleeding margin or pricing the product out.
Repeating and increasing those type of orders have been not prioritized for us.
We get a lot more precise data, more precise pricing, and I personally believe that the market will pick back up by then.
Okay.
It's a timing thing as well.
Okay. There is also someone here that want to ask a bit about the pre-order. We have $5 million-$13 million.
that we have. What is the status of those orders? Are they still there?
Sure. They paid a deposit and they are eagerly waiting for a chance to see the first plates.
Okay. You think they will come over here and check it out when we are?
I think so, yeah.
Yeah?
Definitely.
Okay.
We can send somebody there if they don't have the ability to come here.
There is a second one, a follow-up question to that, on what is the cancellation rights on these orders? You say that they are still effective, but is there any risk that somebody might change their mind, or can you say something about that?
Change their mind, no, but if we, for some reason, have overpromised on efficiency, they have at least a chance to cancel it.
I don't think they will, as long as we beat the competitors in any percentage ways because there's a large need in the market for a non-Chinese player.
Okay.
Just based on that, if we're on parity for any reasons, then pretty sure they will still commit.
There's also the geopolitical situation with Trump and tariffs and export control.
Okay. Did I just increase my volume? Is that?
Yeah.
Okay.
It's not. It's better.
Okay. I can talk louder. With the NOK 5 million-NOK 13 million, these numbers, are they visible in the published annual report or in the balance sheets in any way?
It's visible in the report, sure. Prepayment is, I believe, visible in the balance sheet.
Obviously also comes from liability, that's the same size as the deposit.
Yeah. The deposit, yeah. Okay. Well, I guess we can skip this one. It's sort of something you have answered quite well before. When you have evidence on the chip is functioning as we want it to, and the Scrypt functionality is there with the exact marketable efficiency, what is the marketing strategy? How will we go about to move forward then?
You've got to remember that first of all, the cryptocurrency community is a relatively close-knit community.
Yeah.
There are plenty of very specific events that we can attend. We have all the largest YouTubers who are dying to test our product and market it on our behalf.
I think gaining that brand recognition now is a relatively easy thing to do because it's so narrow in a sense.
Yeah, I don't know if that answers your question, but we have, I don't know, 100, 110 large clients that we keep warm, if you like.
I think Ben spends on average two to three hours a day to
discuss various topics with these
different potential future customers.
It's still like a business-to-business sort of focus then?
No. For political reasons, we have decided to keep the MOQ as low as possible.
Down to a single blade.
It's also designed for sitting in a home computer.
In that sense, we also want to take care of the little guy, and that's been part of the.
The whole range basically.
Yeah.
Yeah. That's cool.
We also have PowerPool and PowerPool hosting. There is 14,000 customers right there that we can market to immediately.
Mm. Yeah, that's starting to become some nice synergy effects to that. It leads over to the next group of questions related to PowerPool. How is it going for PowerPool, the development? There are some comments on.
It-
Yeah
We just launched a new firmware.
Yeah.
It is custom firmware that improves the efficiency or even tunability of our competitor's hardware.
Really nice to have feature. You can't afford to leave 10%-15% of your top line on the table if you're using the wrong firmware, right?
Yeah.
There's challenging sales cycles. Very often people would like to partner up with a pool service provider when they take new sites online.
You're very dependent on their development schedules.
Deployment schedules.
So-
This is to attract new customers and then perhaps the larger kind?
Sure. I think it depends on all of the specifics, but my guess is that PowerPool's footprint, all the.
All the clients' hardware measured in megawatts.
Today should be perhaps 50 MW.
There are several clients out there that have shown interest that have tenfold that in their portfolio.
Really?
It's a little bit of a nightmare to reprogram farms depending on if they even have a fleet management software.
There's also a bit of political things happening. Russian miners often want to use a Russian pool. Chinese miners want to use a Chinese pool.
It's not even legal to mine in China, but it's 60% of the market.
You've got those type of.
Yeah
Frictions as well.
We recently met a Nordic miner that runs a couple of megawatts worth stock firmware, mines to a Chinese pool. I don't see a reason why he should not switch.
Closer to home, easier to gain immediate trust. Ruben also launched Equihash, the Zcash algorithm recently.
Okay. Yeah.
It's picked up pretty decently. Right now, I think it's growing, but 6%-8% of our net revenue from that algorithm.
Oh, really?
I know he's in talks with large Zcash miners.
Those type of developments. If they happen, we most likely have to announce it to the market, right?
Because it's significant changes in the.
Yeah
The business.
Yeah.
Right now, mining as an activity is in a low cycle.
I think that's a good opportunity to innovate and find cost savings. Right time for us to grow the fundamentals, even though in terms of NOK it doesn't immediately show.
When the prices pick back up again, or for some reason hash rate in other pools disappear,
we get more of the cake. It will be very much reasonable in our opinion.
Yeah. That's fine then. To the next question then, which is mentioning that we earlier have said that we are in contact with the players in this market that can increase the revenue and profit of PowerPool 5, 10, 15 times. We are still in touch with these kinds of players .
Sure
It might happen.
We're pursuing.
Yeah
that a lot. I feel also, it's a bit weird actually because some of these clients, I believe, is a lot easier to convert if we do it on the custom firmware side.
Meaning that they see a real benefit. We can charge them on a custom firmware rather than on a traditional PowerPool.
Yeah.
Also, we have these SPDUs, right? That our samples are now currently in production. Shipped them to some of the largest hosters in the world.
That has the most to gain on accurate measuring the client's power consumption.
I think that's a smart catalyst to provide something that is unique and way cheaper than using industrial PDUs.
For instance.
Better functionality, more functionality, developed with the use case in mind.
That's where Ruben really excels.
Yeah
Translating customer needs into technical viable product and services.
Yeah. I saw you moved him to our new position now recently.
Yeah.
Chief Product Officer.
Yeah.
It also includes services.
title is still.
Okay
CPO.
That's probably then related to what you mentioned, that they have a special capability to understand what the client needs.
Also then to execute on those designs. Yeah. You mentioned the SPDU. There was one last thing on There is also a bit on our data center initiatives, but I think we fill in for on the other products, the SPDU, you mentioned that. Tell when is that possible to buy for When is it CE certified and ready to go?
About a month, I'd say.
Okay.
We can also ship uncertified units while they are being certified as samples, right?
Lead time, believe it's three weeks until we have the first 200 samples.
Some will go to certification; some will go into our own operations.
Some will go as samples to clients, so.
yeah.
Yeah.
Actually quite excited about that product as well.
Okay. With that, I think we have reached over many of the questions we actually got in. Also we will start rounding up a bit. I want to go back to the chip again. Do you have anything you'd like to add on the power SIMs? When the mask set and wafer is done, is all this then progressing as planned?
Yeah. I don't see any.
bumps in the road on that topic. Test specifications are developing nicely. The spec sheet is developing nicely. After the test spec is done, the physical work on most of the test software and test hardware is progressing as planned.
Might be some days in between, depending on particulars, but there's nothing, no major roadblocks that I see.
Okay. Are you ready for the football match tomorrow?
Yeah. I'm contemplating whether I should watch it in an intimate setting out in Archipelago or downtown Oslo. I'm very much stoked. Bets have been placed.
Oh, really? Yeah.
I hope it doesn't create friction with our British participants in the team.
In the team. Yeah, that's true. Yeah. Someone was rooting for England, yeah. A very important one also.
Yeah.
Yeah. All right. Well, in that case, I think we're starting to get ready to close this chat. Thank you. You have been talking nonstop for 45 minutes.
The sun is scorching my back now as well.
Yeah.
I moved a little bit.
That's how it is in Oslo. You can come to Trondheim. There you're gonna get rain.
There's no problem there.
Yeah.
No, it's nice to spend a little bit time outdoors.
Yeah.
Thank you, Pareto.
Thank you Investorweb also for setting up this stream and the investors relations pages. I'm sure we all want to get some feedback on how the investors like that. We are stoked and are eagerly waiting for next bit of news. Thank you.
Thank you.
Thanks so much