← Blockspace
Loading summary
Transcript92 lines
- [00:00]
A
Foreign.
- [00:05]
B
What's going on, y'?
- [00:06]
A
All?
- [00:06]
B
Welcome back to Blockspace Live, brought to you by CleanSpark. Charlie we are bringing the heat today with OpenAI announcing Jalapeno, its first ASIC chip. You heard that right. ASIC chip for inference compute in partnership with Broadcom and what implications this might bring for Nvidia's chokehold on the AI computer value flow. That's our opening story for our interview today. We have the head of Bitcoin strategy at Smarter Web company, Jesse Myers, on to talk about what else but strategy and specifically Stretch continuing to trade under its $100 peg, why investors might be spooked right now, what's leading to this unwind and how strategy in the market could correct it. For our third or, sorry, second story, Iron's tax abatement for an Oklahoma data center is up in the air. We've got resident Oklahoman Charlie Spears here with all the details on that. And for our final two stories, SpaceX is raising 25 billion in an unsecured senior note sale that was oversubscribed by to 89 billion. The numbers just continue to inflate and we are not surprised, but a little bit amazed at how much interest we're seeing there. And finally, CoreWeave is deploying its GPUs in stock or, sorry, excuse me, Sweden for its second Swedish deployment.
- [01:36]
C
Block Space goes live weekdays at 1pm Eastern, featuring quick hits on the latest in AI data centers, Bitcoin mining, emerging tech and markets. Make sure to like and subscribe. Hit the notification bell. If you're on YouTube, get the push notification comment in the chat. You can type your words of wisdom in the Twitter and X chat if you want. If you like this stream, it turns into a podcast shortly after we wrap anywhere podcasts are found. Search Block Space in the search bar of your podcast player of choice. We also have a newsletter newsletter.blockspace media.com like and subscribe. Check that out. Quick Hits delivered to your inbox every day. This show is brought to you by CleanSpark. Nasdaq listed ticker CLSK. More on CleanSpark later on in the show. Colin It's a fun one. Nvidia finally gets a shot across its bow with OpenAI announcing they're putting out they're working on an ASIC chip with Broadcom Asics. Finally, we're back in our comfort zone.
- [02:50]
B
Kind of, we are back in our comfort zone, except it's a totally different form of compute, much more difficult than running hashes, and it's not a jalopy It's a jalapeno, Charlie. So we've got the tweet up here from OpenAI. We've designed and built our first ASIC chip or our first AI chip, Jalapeno. Designed from the ground up by OpenAI and brought to production with Broadcom as the silicon implementation partner. It's purpose built for LLM workloads powering ChatGPT codecs, the API and future agentic products. So what the straight skinny here is that OpenAI has taped out this ASIC and actually Charlie, they're already using it, which is pretty incredible.
- [03:36]
C
This isn't an announcement, this is a, hey, we're already using it. We're going to roll it out significantly, I believe. They target what one point something they're
- [03:46]
B
targeting a gigawatt, a gigawatt of by the end of 2026, which is kind of crazy. I mean I will say I would flag that as very aspirational. Obviously I don't know how far along they are in the, in the tape out. And they did not release any information for how much they have produced currently. But these chips are currently running ML machine learning workloads in a lab at a production Target frequency for GPT 5.3 and Codex.
- [04:18]
C
Spark.
- [04:19]
B
A few other things here. OpenAI is just a few housekeeping items. OpenAI is the chip architect and they used ChatGPT to help design this ASIC chip. And according to their own disclosures here, they said that they went from design to tape out in nine months. If that's true, that is absolutely breakneck speed. And it kind of reminds me of to point towards one of their chief rivals, Xai, the building of Colossus in 122 days. The time to market for some of these technologies are really breaking people's models for how fast you can get some of these things done. Broadcom, like we said, is the Silicon implementation and networking partner. And I believe I'm pronouncing this correctly. Celestica is building the board rack and system level integration. So they're building the shoebox for the chip to live in. And one last point on the 1 gigawatt rollout. Microsoft will be their premier partner for this. They plan to run these chips at Microsoft owned and co located data centers across the US. I still do just want to highlight that 1 gigawatt, even though they taped this out in 9 months according to their own disclosures is very, very fast.
- [05:40]
C
Yeah, and so why does this matter? I think if you aren't really familiar with the physical level of compute Again, we are armchair Claude driven compute researchers here. Why is this a big deal? And Colin, I'm going to invite you to give a little bit of history but context for why, you know, the past 20 years of compute regime have been GPU based. But these ASICs, these are chips which are designed specifically to be optimal at one thing. They're, these are going to be optimized for inference. You have the training of the models, then you have the inference of the models which answer and infer answers to the questions that you act. They're not, these are not really designed for training models, which again is a little bit less clear how you physically design the chips for those. But for inference, the real challenge is, is that you need all the software architecture and the types of things you're telling the chips to do to be a lot more standardized and vertically integrated. And so that is one of the reasons why Nvidia has dominated because they have this much agile GPU stack. I'm going to toss it to you, Colin, maybe to give a little more context on this.
- [06:57]
B
Yeah, so this is something I've been curious on for a while, I've been hearing about and we covered on the show ramp ups in ASIC design for AI. And so my first question is the obvious one. Why have we not seen more ASIC chips for AI deployments until recently? Part of the reason for that is because Nvidia has carved out a moat not specifically for their chips.
- [07:20]
A
They're good.
- [07:22]
B
That's not the real competitive advantage though. They have a software stack and human capital advantage. So Nvidia released what was known as CUDA, a software platform in 2006 which allowed GPUs to be used in parallel for multi purpose compute. So they weren't just being used for graphics interfaces, they were also being used for other forms of compute. And then the real inflection point came in 2012 when Alexnet, which was like a breakthrough neural net model back in the day, was trained on two Nvidia GPUs. From here, CUDA became an industry and academic standard for neural networking, giving Nvidia kind of a compounding human capital and software effect. But now this is why we're seeing OpenAI come out with this. Now the economics are catching up with this software note. Nvidia's advantage was again never really about the chips. It was about being the first to make GPUs programmable for general AI workloads, then locking that in with software before anyone realized how valuable it would become. But now OpenAI is making a bunch of money from their inference capabilities, obviously somewhat from running models and generating new models, but really the cash cow is coming from inference. And now they have the money to deploy into making these asics for inference specifically. And my understanding is that really now that the economics make sense, building your own Asics, it's a multi year, hundred, hundred million dollar undertaking. It makes sense now because you can kind of predictably amortize the cost for that with your revenue from inference. And also these ASIC chips are really, they can do inference much more easily than they can do training. They need to, as Charlie said, with the asics being optimized for one thing or a few things with inference, you can really just tell it to run that inference. But when you're training a model, you have to keep tweaking the parameters. And so GPUs are still king for that. But we'll start probably seeing asics ramp up in terms of inference workload going forward.
- [09:26]
C
And if we consider there's multiple dimensions here, if we consider that ASICs are roughly 5 to 20 times more efficient at inference when we get everything dialed in, these are the numbers I'm seeing online. They're one to two orders of magnitude max more efficient than a GPU at inference. And we consider that we have these multiple dimensions of bottlenecks to roll out more compute. The manual, with all the ASICs, can do comparably more inference per unit of power than his competitor. And then also we consider the fact that Nvidia has had this, the chokehold on this like GPU and CUDA regime. And this could signal a change in the weather. I'm going to show this chart which I found really interesting, which kind of shows visually what we were projecting back in 23 and 24 for the rollout of GPUs versus ASICs in AI compute, specifically inference. And at the time it was projecting that sometime in 2026 there would be a crossover where more ASICs would be deployed than GPUs. And now the whole story has been GPUs, B3 hundreds, Nvidia, Blackwells, Google, TPUs. But here we have OpenAI saying we've actually been using our own proprietary ASICS with Broadcom and we're going to expand to a gigawatt by the end of the year. So maybe this inflection point has already been happening and could be crossed soon.
- [11:16]
B
And one thing that could accelerate this last note on this. Broadcom has helped produce custom ASICs for Google and for Meta in the past. This was pre AI you have to wonder are we going to see them get involved with Anthropic? Are we going to see them get involved with Google for Gemini? Is Broadcom kind of carving out its own moat here to be the premier silicon implementation partner for AIA6.
- [11:41]
C
Something to watch so I'll hit you with some skeptical or cautious takes one while Broadcom CEO Hawk Tan fun name while he says that the jalapeno reportedly matches the like Nvidia Blackwell performance levels in certain areas. We don't have public benchmarks. Always good to be skeptical whenever someone says they have a chip that does X and then we get to see how it actually performs. And then Asics again you are trading a flexibility and dynamism for efficiency and cost. This industry changes rapidly almost one thing is for sure which is that it will be different in 10 years. How much do you bet a lot of this becomes execution cost and a little bit dependent upon like compute agility and then you have like ramp challenges. How fully vertically integrated is your software stack? How do you integrate it physically? And these are big questions. It also increases your capex requirements per gigawatt so you're actually spending more money to roll out a gigawatt. It does become a capital problem although the money side seems to be largely solved right now as infinite money is pouring in to AI. I guess maybe I'll share one meme from our boy Stu here. This is great shows Ricky from everyone's favorite Canadian show Jalapeno vs. Jalapeno Chips.
- [13:18]
A
So
- [13:21]
C
Colin, you got any other last
- [13:22]
B
thoughts on no, I think that does it. We'll we'll move on to yeah ad roll and then bring on Jesse Myers to give everyone a chill outtake on the stretch drama.
- [13:34]
C
We're going to talk about stretch before we bring Jesse on. We are going to hear from our sponsor CleanSpark.
- [13:40]
B
We are CleanSpark, America's Bitcoin miner. A publicly traded company with the largest operating hash rate powered entirely by self operated infrastructure across four states.
- [13:56]
A
This is our proof of work and
- [13:59]
B
we are setting the standard for what's next. Learn more about the intersection of energy and Bitcoin at CleanSpark. If Bitcoin's actually the best money and it's the thing that people should accumulate is the best risk adjusted asset. I lose zero sleep about whether or not that's going to happen. I just ask the question of when is liberty matrix math that you're running
- [14:27]
C
on large pieces of data.
- [14:28]
A
The bitcoin miners can absorb that energy
- [14:31]
C
and in many ways this feels like
- [14:32]
A
a second bite at the apple to build a new Internet.
- [14:45]
B
All right, Charlie, let's get Jesse up in here.
- [14:47]
C
Let's do it. We're going to bring on Jesse Myers. We haven't done a mic check, but let's see. Jesse, can you hear us?
- [14:52]
A
Hey, gentlemen, how are you?
- [14:54]
B
Welcome to Great, man. Thank you for joining.
- [14:57]
A
Thanks for having me. Excited to be here. It's, it's quite a moment in time to be here as, as bitcoin's price is tumbling today.
- [15:07]
B
I'll tell you man, that's it.
- [15:08]
C
Just cross 60k. I'm glued to the screen.
- [15:11]
B
We keep coming back to it, just talking about how it has been a pretty demoralizing year with the AI trade sucking up all of the liquidity. Maybe I'm just being a blind bull, but I've often found that bitcoin surprises to the upside and the downside. It usually defies expectations. Makes me hope that we're setting ourselves up for some sort of countercyclical trade. But that aside, the real reason we wanted to grab you is because you've had some temperature reading tweets about the stretch drama with stretch trading below $100 peg and what it means for strategy. And I would love for you to open this by just saying, giving us a rundown. Why do you think this sell off is happening and what has strategy done to address this?
- [16:00]
A
Yeah, it's kind of hard to say definitively. I think it's, and it's a combination of factors. Ultimately I think it's, I was just looking at the bitcoin chart and comparing that to stretch and SATA and seeing when were these sell off events happening. And it's been over the last two weeks during market trading hours. Bitcoin has been selling off at certain points and those periods of sell off for bitcoin have been the periods of sell off or stretch Also that suggests that there might be some larger players who are squeezing the bitcoin price lower and punishing leverage longs and just trying to cause panic. That has been a profitable trade in the past for deep pocketed. Realistically, Wall street hedge funds that play in these waters because they can cause panic and irrationality and they can trigger leverage liquidations in bitcoin and now in stretch. So I think that what's been happening over the last six months with stretch, for a long time there, the narrative was increasingly this thing is now trading at $100, maybe down to $99 and it's never going to be $90 again. That confidence built for a while. And whenever that exists, you create the conditions to invite leverage. Because if you believe Stretch will never go to $90 or even $95, you can take on a leverage position. You can 10x leverage, and that works great until it doesn't. Right? So if there's ever any sort of destabilizing event or a narrative that goes against holding Stretch or having confidence in Stretch, then things can accelerate and you can trigger liquidation cascades as the small percentage of people who engage in this trade had a lot of leverage. And you can also trigger panic where people are like, well, I thought this was a low volatility asset. It's no longer a low volatility asset. I'm going to trim some of it right now because this is sort of against what I thought it was. So all those things have added up over the last couple of weeks and created the conditions for Stretch to trade. At the moment, we're at $80, and that's quite a surprise to me. But I think it's the free market nature of Stretch playing out and biting the people who got overconfident that this thing was always going to trade at 99 to $100. And stretch is less than a year old. And this is, I think, the first big learning event for these instruments that it'll trade close to par until there's a market panic with bitcoin or some narrative about strategy is in danger because bitcoin is going to zero, and then obviously their credit instruments wouldn't be serviceable from that point. And fear spirals at bitcoin bottoms or times of bitcoin panic. And we're seeing how that translates now into a stretch panic. But I'm sure we'll dig into this now. But the fundamentals for Stretch and for strategy are still completely unchanged.
- [19:59]
B
Charlie, did you.
- [20:00]
C
Yeah, well, I was going to point out, like, I mean, I'm looking at the strategy dashboard here. Stretch teetering right on the edge. $80. I think I saw it flicker to 79 WA while we were talking, but the yield's gone up. That's kind of the plan. That's the design. Jesse, I'm kind of curious about, like, I mean, I don't know how to price these things. I don't know what the right yield is for this. How many commentary or thoughts on, like, on how the yields price does that? Does that. Do you think that'll sufficiently incentivize people to, you know, buy more Stretch, bring the price back up to 100? What are your thoughts on this system and balance here.
- [20:39]
A
Yeah. So from the beginning, these things are designed to rely on free market price discovery. This is designed basically like a bond, where it's a variable rate bond, where they're currently offering 11.5% dividends on the notional value. So that's on the $100 value. But as you pointed out there on that dashboard right now, because it's trading off of the $100 value so much, the effective yield has gone up because they're still doing 11.5% per year on $100. But you can buy it for $80 right now. That's an effective yield of 14 and change. 14% and change. And so basically, the instrument has gotten more attractive if you believe that they will continue to pay the dividend payments. And that's where, you know, your. That's where understanding how this thing is constructed matters a lot. Stretch itself is not in danger at all. Stretch depends on strategy continuing to pay dividend payments. So the thing that matters is strategy's health. And what is strategy? Strategy is a bitcoin balance sheet. So what matters is Bitcoin. Ultimately, they have a ton of bitcoin and they have fundraising capabilities, whether that's through issuing more equity or issuing more prefs or doing traditional debt. And they can use those proceeds to pay dividends, or they can sell some of their bitcoin if they need to. That's kind of the last resort. So for strategy to be in trouble, bitcoin needs to go to under $10,000 per coin and stay there forever. And so then it becomes a question of, do you have conviction that bitcoin is simply in a bear market and doing what it does in a bear market and feeling really gloomy here at what's probably at or near the bottom, and that it would be kind of up from here. We won't visit $10,000, and we certainly won't stay there for multiple years. That's the question that the market has to answer. But a lot of people are forced sellers of stretch. Whether that's because of leverage or now this thing has volatility, it didn't before. I don't want to hold it. And that just creates an opportunity for people who think, wait a second, this thing is now mispriced because they still have all this bitcoin on their balance sheet. They're still in an incredible position. Bitcoin's not going to go to $10,000 per coin. I think that's up to the individual and therefore their capacity to pay the dividends is unimpacted and they will keep paying dividends on stretch. And therefore I'm interested in 14% and change yield effective yield right now by stepping in as a buyer at $80 per stretch share. So I think that's what's going on. And if that is what's going on, then the current turmoil resolves itself as people start to realize strategy is still paying the dividend payments and their capacity to pay those dividend payments is really unaffected. And therefore this thing just got more attractive. And I might be interested in being a buyer here at 14% effective yield. And if buyers outweigh sellers, the price will march back up to the price equilibrium that the market feels is appropriate. Now meanwhile, strategy has the incentive to keep raising the dividend rate because it's a variable rate. I expect they'll probably on June 30th, which is the next date where they would announce what the upcoming dividend rate will be. They'll probably raise it to 12%. Yeah, so yeah, so the next payout date was sort of coincides with the date where they would announce the rates for the next period, which is July. And yeah, they'll probably raise rates maybe 50bps. So that would be up to 12% and that makes it that much more attractive as an asset. Yeah, so we'll see what happens with this. But I think the market is just going through its first learning experience with digital credit of these things are going to sell off when there's panic. But that just makes them more attractive. If you have dry powder and understand that the capacity to service these debts is essentially unimpacted, you're confident that the
- [25:34]
B
peg will be restored. Once enough investors realize that there's a trade to be had here. And there's a. At this point you can make a pretty penny if you were to long from here. And it was the bottom. Going back to your point about ultimately stretch is only as good as strategy is. And like you said, they only really would face stress if Bitcoin were to drop below 10,000 and stay there for a long time. Is there anything else that you think would spook the market enough to lose confidence in these instruments or is it really just spoil? Does it really just boil down to Bitcoin's price dropping that low?
- [26:10]
A
Yeah, I struggle to see what else would cause an existential risk to strategy. They have minimized their traditional debt to a point where the vast majority of what they have outstanding is preferred equity obligations. And there is no, there's no holder put, there's no liquidation risk with preferred equity obligations as they've structured them. So the only question is whether or not they can continue paying the dividend payments. And if they can't, then they will suspend the dividend payments. That's the exchange between strategy as, as an issuer of preferred equity obligations and the investors get a higher yield in exchange for giving up a lot of the traditional rights that are attached to traditional debt. And that means that that's what protects strategy in the event of 80% bitcoin drawdown. That's why you get rid of the liquidation risk that would happen if you, if you had traditional debt. So yeah, I mean, the health of strategy is entirely based on the health of their Bitcoin treasury and that is directly related to Bitcoin's price and pretty much nothing else. So that's why I think this is all overblown. There's swirling fears about is strategy being attacked or are they risk of liquidation. But they're, they're literally not like the, the structure of the preferred equities make it impossible to liquidate them. So, so I think it's a, a nothing burger.
- [28:07]
C
I feel like you've already addressed this, but I want to like, ask the question directly. And you've had a few tweets kind of explaining why analogies to, I'm going to say it, Terra Luna are misguided and they're not like strong analogies. How do you, how do you explain this? What's your argument here?
- [28:26]
B
Also, for listeners who might not understand what Terra Luna was, if you could just give a very brief history lesson.
- [28:33]
A
Yeah, so. So Terra Luna was a structure that came about, it imploded in May 2022 to, to cut to the end there, the idea was let's have two separate assets. And you know, I don't know, I don't remember the details of it super well, but the basics of it was let's have two assets, one's Luna and one's Terra. And Terra is going to be our stable coin. It's going to be pegged at $1 and we're going to use the Bitcoin treasury of Luna to. We're going to sell Bitcoin and use those proceeds to buy Terra whenever it dips below a dollar. And then we'll, and then we'll sell it when it's back at a dollar and replenish the reserve. So in theory, that system is, you know, it was an algorithmic stablecoin where it was going to self balance, right? It was going to maintain the peg at a dollar, which works great until there's some sort of pressure on that system. And, you know, there started to be pressure on, downward pressure on Terra's peg, and the system automatically started using the bitcoin reserves to bid up Terra.
- [30:01]
C
But I'll also interject, they had a bunch of other assets which were not Bitcoin.
- [30:06]
A
Fair point.
- [30:07]
C
They made Bitcoin look stable with how volatile.
- [30:12]
A
So they're using their resources to bid up their algorithmic stablecoin, but depleting their resources in the process. And so that actually just caused a vicious cycle of people became more and more concerned. Wait a second, you're running out of assets? I need to get out of Terra. So everybody ran for the exit and they ran out of Bitcoin and the price peg disappeared. The Terra went to zero, and boom, it all went bust. So that's what happened four years ago. And crypto people are afraid of anything that looks like it's pegged to a certain price level as a result of that learning. But Stretch is the opposite design here. They're not saying we're going to maintain a peg. They're not saying anything. We're not going to maintain anything. They're saying we're going to allow the free market price discovery to find the appropriate price level that informs the effective yield that investors should be getting on this instrument based on all the known information at that point in time in the market. They're not using resources, they're not depleting their Bitcoin treasury to defend the peg. They're using the Bitcoin treasury to pay the notional dividend payments. Well, that's their last resort. They're going to raise capital by issuing more equity or more prefs as their first and second line of defense to pay the dividend payments. But at no point did they get in a forced position where they're having to deplete resources above and beyond what they, what they planned on. They simply have to meet the dividend payments. So, you know, it's. I understand how crypto people, people who were lived through the Terra Luna thing, have the scar tissue from four years ago, but it's the opposite design. You know, this is not an architected system where we're going to, you know, maintain a peg by force. Instead, it's designed to allow the market to tell, to decide what the appropriate price level is and structured in a way where the effective yield goes up as the price goes down, such that it should invite buyers as the price goes down. That's the Manner in which itself heals is that the lower the price goes, the more attractive this asset gets. So long as you believe that they're going to continue to pay dividend payments, which. Which they will, because they're not depleting their resources to defend the peg. So they still have their incredible bitcoin balance sheet and all their fundraising tools in order to pay dividends.
- [33:01]
B
They also have a decent cash reserve. There was a lot of chatter about they depleted it by 1.5 billion, but they retired the worst note in their capital stack.
- [33:12]
A
Right. They had an opportunity, I guess like a month ago to retire a billion and a half dollars of convertible notes, an 8% discount, which is very attractive to shareholders, to the company and shareholders. And so they did that. They used some resources from their USD cash reserve to do that. And then suddenly their USD cash reserve for dividends, what they've set aside for dividend payments that's separate from their bitcoin treasury, was now six months. And that was, I think, part of the narrative that's popped up in the last two weeks of like, will Stretch continue to be dividend payments continue to be paid? And that just kind of allowed a swirling effect of fears that maybe strategies running out of money. But no, people are just forgetting that they just retired a billion and a half dollars of debt at an 8% discount. And so how did strategy respond to the sell off of Stretch last week? Well, they sold $335 million more of MSTR shares. So they issued those into the market at a 1.14 MNAV, so technically accretive and used about 90% of that to set aside to add to top up their USD cash reserve. And the rest of it they bought more bitcoin with. So even last week during this spiraling fear about Stretch and strategy, they were able to raise enough, enough new capital to increase their dividend reserves from six months to 10 months. That was in one week of panic. I expect they'll continue to top up their USD reserves probably up to two years now, because I think they're seeing the market is spooked. And that's a good way to shore up confidence about the capacity to continue paying Stretch dividends in a bear market. So that's probably what they'll do. And then a month from now, everybody will look back at this. It's like, why were we all afraid? Of course they were good for it because they had all this 10 months of cash set aside and basically infinite capacity to keep fundraising.
- [35:43]
B
Well, Jesse, we'll let you go. We've had enough time or taken enough of your time. Thank you for the rundown. And there's a little tongue in cheek sign off. We'll be keeping an eye on the stretch price until then. Steady lads.
- [35:54]
A
Deploying more capital.
- [35:56]
C
Thanks Jesse.
- [35:57]
A
Thanks for having me guys.
- [35:58]
B
Thanks man.
- [35:59]
C
Cheers.
- [36:01]
B
That is Jesse Myers goes by Croceus Bitcoin BTC on Twitter. Also, most people don't know wrote one of the, I believe, seminal essays on the disconnect between bitcoin culture and wider elite culture called why the Yuppie Elite Dismiss Bitcoin. Highly recommend looking it up. If you're wondering why that one friend who is all in on everything else doesn't quite understand bitcoin. Although maybe the arguments don't land so well when bitcoin's dipping below 60,000. But that's it for that.
- [36:33]
C
It's definitely an I told you so piece. There's a lot of these. So if you yeah, that was like a great time to write some good substacks about why you should buy bitcoin. Think of it as like a long term investment. We are going to keep rolling. We are going to talk about Iron hitting some snags here in my home state of Oklahoma, SpaceX and Corwe. Before we go to that, a word from our sponsor, Luxor.
- [37:04]
B
This episode is brought to you by Luxor's Commander Bitcoin miner management software for enterprise operations. Commander gives you real time fleet monitoring, bulk remote commands across your fleet and intelligent miner that is an automated profitability engine that runs every five minutes and tests your fleet's power settings against live energy and hash rate markets. ERCOT back tests show 10% improved profitability with intelligent mining versus binary mining. Commander Pro is $100 per megawatt or a 25 basis point pool fee adder. Roughly half the price of competition and you can try it for free for 60 days. So if you'd like to learn more, go to Luxor Tech forward slash Commander Charlie yes on with this tax abatement and why is Iron potentially not going to get the tax abatement that they want?
- [37:56]
C
Yeah, so here's, here's the deal. So Yesterday iron reported Q3 2026 earnings and again everybody's kind of waiting with bated breath what deals they're going to announce. Blah blah blah. They tease this one a few months ago that they said we've got a bunch of power capacity in Oklahoma. I as an Oklahoman and somewhat dialed into these deals was like, I was like oh, I haven't heard about this. Where is this coming from? Well, we have a little more information because as we can see here, they updated their Oklahoma Data center update, noting that this particular power capacity in this plan had hit a local community pushback snag as Iron was applying or working towards a tax abatement. Basically waive some of the tax, lower taxes, and we'll invest a lot into the region. And this is in Pittsburgh county, which is center of Oklahoma on the eastern side around the Lake Eufaula area. That's a very, very cheap place to find power. Not surprised whenever we put the dots together there. So it was in Pittsburgh County. And what happened is there in the process of applying for this tax abatement, the community pushed back. It's kind of a story that we've been hearing forever. This is not a data center moratorium, but rather the Pittsburgh county delayed decision on proposed tax data center tax increment district. So a tax increment district, or TID as it's abbreviated here, basically changes the tax regime of a area either for or against to increase or reduce taxes based upon some agreement. This is different. This is different. This is decided differently than at the city council level or the county level. This is a separate body which. And this is typically the case in most like counties and areas. It's a separate body or governing body that decides what the regional tax policy will be. So the commissioners decided to table the decision, which Iron had been pushing for and advocating for, because just a few days ago, at the community engagement meeting, over 200 people showed up with various concerns from the. Let's see, from this story. I thought I had a different story pulled up, but basically it's the standard kind of community concerns that you're hearing in across the country. Water, power, noise. Citizens pushed back and said, we don't see the specifics of the deal. We don't see, like, decibel. You know, we don't see decibel limits. We don't see, like, actual plans for it. And in the face of, like, broad, like, rapidly increasing, like, data center pushback and skepticism, or as I've seen in Oklahoma, almost like fiery, white hot rage, this is not a surprise, and this is funny. This is a 1.6 gigawatt project which has now been delayed, not canceled, but delayed because of what is really a. A $24 million tax abatement request. Iron said Iron claims that the. The additional property tax alone generated from this development would generate about 60 million a year. And they were asking for, I think, about 24 million in, like, tax abatement for this proposal. And this doesn't include like what Iron says are additional jobs, infrastructure, commerce just to the region. So it's, it's wild. This is a major, major project which has now been delayed. We don't have a timeline for how long it's been delayed, but because of a TID in the tens of millions. Colin, I'll toss it to you for some more commentary.
- [42:12]
B
Well, investors and traders are not liking the news. Iron down 8.4% today. Now Bitcoin's also selling off right now. So it's hard to really say whether or not this is mostly noise. Is down 15% over the last five days and 16% over the last month. So could just be another story that's negatively weighing on the stock price as well as bitcoin because you know they still have a decent amount of hash rate.
- [42:42]
A
Right.
- [42:43]
B
So they're going to trade somewhat with bitcoin as well. I don't have too much to say about this other than oh, go ahead.
- [42:50]
C
Oh I, I do. I have some other additional thoughts if you've.
- [42:54]
B
Well, I guess my one thing is I, I don't want to dismiss these communities concerns completely in the sense that as we've talked about some of them, like I think energy use and energy prices probably the one that matters the most in my view in terms of the negative impact these data centers could potentially have. As long as the power authority and the utility though are not just over allocating towards these things, that shouldn't be a huge problem. There's obviously a lot of research that's been done on either side. Some research suggests that power prices actually fall in places where data centers have been rolled out. I'm a little skeptical of those claims just because I don't know what they're actually looking at and who's funding that research. But the other thing that I would say, I understand that there's a tax abatement and that there's a lot, a lot of data centers have taken these property tax breaks for the first few years of them being online partly because the localities want to court them to get those tax, those taxes and to have that tax revenue in the future. But I would say for these residents, I would just ask, do you not see any value or benefit in having that money coming into your community? I don't know anything about this part of Oklahoma. I would imagine though that while it may not be a life changing sum of money, you know, $11 million to the local school district, to Kiowa Public Schools sounds Like something that a lot of localities would actually want to have. The other thing I will say I do wonder just in the game theory
- [44:37]
C
of
- [44:40]
B
finding the right locality for these data centers and the leverage that these localities will continue to have over these companies since power is a bottleneck and there are not as many suitable sites for AI as there as there as there could be, I do wonder how much sway the data centers will have in these conversations going forward to actually get these tax breaks or if they're just going to have to just not or if they're going to exit the equation entirely. Because if you think of this, if this locality has a few different data center providers bidding on this space, then they could probably tell all of them to go pound sand. We're not going to give you a tax break in the future. Since there's so much competition for these sites.
- [45:28]
C
It's pretty clear to me that the the power is now in the hands of the local officials. According to non doc, this site is a $50 billion data center site. And if we kind of like gauge and look at what the average cost per gigawatt that people are talking about is, that's not too far off from what the new Jensen Huang like 100 billion per gigawatt cost is going to be. But if you look at this, I mean are you willing to now you got to do some new calculus. What is the value of delaying a $50 billion data center indefinitely or incurring long term community resentment or push even just a few months or a year when the total cost of the project is 50 billion? Do you want, you know, is 10, 20, 30 million going to move the needle here? So I think like maybe the meta going forward is how do you reduce friction like local community friction and surface area for local community pushback. And I'll also point about, I'll also make this point. People think Oklahoma is this super red state and they would be right. That doesn't mean that Oklahoma's pro data center. In fact I was doing some research on this and it looks like Oklahoma may be second to Virginia. The not ground zero, but ground one for the second most organic grassroots pushback of any part of the country. The there have been moratoriums in Luther county for Beltline Energy Broken arrow. Just last week a suburb of Tulsa put a six month freeze. Edmond, Oklahoma City, Tulsa county or rather Tulsa Municipal. And that is just the short list. I see there are like I'm on Facebook and I'm seeing people organizing groups to go to city council meetings and commissioner Hearings to push back against data centers. It is the most, it's funny like the, the best boots on the ground data center intel is in your local Facebook group, which is anti data center where people are driving by, taking pictures, they're locking every single like city and municipal, you know, regulatory minutiae and logging it in their community Facebook group to oppose these data centers. Yeah, it's kind of amazing just to take a look. I mean here's, here's Oklahoma down near Lake Eufaula. Right here you have, just south of Kiowa county. You have the specific area that this data center would be. I think they say it's a couple miles south of Kiowa here. And just if you're wondering, if you zoom in on like all these little white beige tan dots here, those are all gas wells. So those are all oil wells, small vertically drilled oil wells, very shallow. And this part of the state is flush with very well developed aged mature gas field. And that's kind of why you have all the, all this gas infrastructure. Not saying oh, deal is related to that, but like this is, this is like a typical Oklahoma vignette here.
- [49:01]
B
I'm sure that if they can actually get gas on site, that's going to be a huge value add for this data center. I'd assume that's part of the reason why they're in the area. Last thing I'll say it's interesting, obviously Oklahoma, very red state, so you probably wouldn't get this anyway. But I would imagine that the oil and gas industry didn't get much. Too much. I mean, some pushback, but not in the sense of like, you know, local councils and local.
- [49:25]
C
No, you don't. And the thing is like oil and gas in general, broadly in the United States is very, very well, like regulated. It's very clear what the policies and laws you are, you have to follow, how it's taxed, how it's permitted, it's really, really mature.
- [49:40]
B
Right. But the other big thing here I think is that the wider community really can engage in that industry and actually clear, tangible, tangible benefits from it. They might work in oil and gas. Maybe they own property that was prospected and they sold it. Right.
- [49:57]
C
I keep mentioning this, like, look, each of these little well pads here, each one of those is revenue to whoever owned the land prior. Ongoing revenue.
- [50:06]
B
Yeah. And so ultimately part of the pro, the PR problem for these data centers is that outside of saying, hey, we're going to put X amount of money into your schools or here's how much the property tax revenue will be. There's no clear benefit for the wider community, especially when you consider, yeah, these data centers do there. There are jobs associated with them. It's really not as much as the companies would like you to think. And the bulk of the jobs are contractors to build the sites and a lot of those guys are probably coming in from out of town.
- [50:36]
A
Yeah, right.
- [50:37]
C
I'll point out that once a well's drilled and it's just out there, you have a similar. You got, you know, your pumpkin, your pumper who goes out there like it doesn't employ that much sustainably. So the analogies are actually pretty similar, but the actual wealth effect is very different also. It is as know a lot of the liberal coastal elites absolutely hate this and this is their nightmare. But the rest of us, middle America, have a sense of romanticism. Listening to country music in our old Ford as we drive past pump jacks. So as we go eat, you know, chicken fried steak at the local diner
- [51:12]
B
in Kiowa, I mean, I will say I've driven across the country multiple times and I do get a little fired up when I'm going into, you know, western Kansas and eastern Colorado. And you see these, your super old pump jacks just slowly rotating.
- [51:28]
C
Yeah.
- [51:28]
B
And then I see the windmills and I get enraged. No, it's all cool. All right, I say we leave that one there, Charlie, and let's, let's go
- [51:39]
C
on to the next one. We got it. We're gonna stop looking down at the ground. We're gonna go look up at space here. So SpaceX after a word from our sponsor, Lygos,
- [51:53]
B
Bitcoin is below $60,000 and Michael Saylor just lost his shirt. Kidding. That being said, there's a lot of stress in the market right now and if you're not holding your own keys when you're using a Bitcoin lender, you are opening yourself up to liquidation risk. Because who knows what some of those other lenders are doing with your Bitcoin. But you know when you use Lygos that you are always in control of the keys to your kingdom. Lygos is our preferred Bitcoin backed lender. Here at Block Space they use Bitcoin native smart contracts to make sure that you are always in control of your keys and nobody else can touch them but you. With Lygos you always know where your Bitcoin is. No bridging, no rehypothecation, no wrapping. Just competitive rates as low as 8.5 to 10% APR. Go to Lygos Finance to Learn more. All right, Charlie. SpaceX, SpaceX, SpaceX, it seems X We used to cover bitcoin, Bitcoin, bitcoin. Now we're just covering SpaceX, SpaceX, SpaceX. But with good reason because the headlines that keep coming out post IPO and that came out pre IPO continue to dazzle to the upside. This is coming to us from CNBC. SpaceX raises 25 billion in debt sale less than two weeks after IPO and this, this senior secured note sale was over subscribed to 89 billion, a figure originally reported by Bloomberg here. Cnbc says nearly 90 billion. So they can feel like they have a different take on it, you know. But we covered this senior secured or senior unsecured note raise yesterday. It had not been priced yet. It was announced on Monday. And now we actually have specifics from the raise itself. I'm going to go over those really quickly here. The notes have not been registered under the securities act and may only be offered or sold under an exempt registration. So this is a private placement, not a public offering. Basically what that means. There are fewer reporting requirements for this and only qualified institutional investors are allowed. This allows them to close the raise faster, but it also means that it can only be open to institutions, asset managers, insurance companies and pension funds rather than retail investors. And again, faster legal pathway to get the sale done. That's part of the reason why they were able to turn it around so quickly. There are five tranches of notes. Notes due 2031. 7 billion worth with a 5.35% coupon. Notes due 2033. 6 billion worth at a 5.65% coupon. Notes due 2036. 6 billion at a 5.875% coupon. Notes due to 2046, 2.5 billion at 6.6% and notes due 2056. 3.5 billion worth at 6.65%. The blended average is 5.855% for a total of 25 billion billion. Now, SpaceX will use the majority of this raise after closing costs to pay down a bridge note that it took on before it IPO'd believe that note was for roughly 20 billion. They might have a little bit left over, might just pad that cash position. Or they might end up, I don't know, maybe after closing costs and after paying down the note, there won't be that much leftover. But it did seem like they upsized it because original reporting from CNBC said that they were only going to raise 20 billion. Maybe that the source was just wrong about that. But it seemed like they were actually citing SpaceX investor materials. So maybe with the over subscription they ended up bumping it up by 5. Might as well take the money when you can get it. Charlie, what's your take here?
- [55:39]
C
I don't have too many takes. Except maybe SpaceX should have raised more than 20 billion. There's certainly appetite for it. In a rare turn of events, SpaceX might be a little conservative here.
- [55:51]
B
Right. And you know, they really don't need the cash though.
- [55:55]
A
Yeah, they don't.
- [55:55]
C
They already have a hundred billion dollars of cash. Well, I mean 120 billion that can
- [56:00]
B
build like what, one data center? That's one data center, Michael.
- [56:03]
A
How much?
- [56:05]
C
Half a gigawatt.
- [56:09]
B
So yeah, not much more to say there. Just that the institutional appetite for this stock continues to surprise to the upside. I'm good with leaving it there, Charlie. I think we can end with our final story on Core Weave.
- [56:21]
C
Yep.
- [56:23]
B
So little quick Diddy here, but wanted to get Core Weave in the news partly because it's playing into a trend that we've covered here over a few segments and that's Core Weave signs co location agreement with Canopto in Sweden. This will cover two campuses in Stockholm and it is with data center operator Canopto. And the initial capacity for this build out is already live at Kanapto Stockholm 4South site. Now Stockholm 4South is Connapto's newest and largest data center. It's 32 megawatts of IT capacity and we don't know how much of that is going to coreweave. This was not announced in the press release. Also couldn't find it in any SEC filings. But another notable part of this is Core. We will deploy both Nvidia Blackwells and the Vera Rubin GPUs at this site. It is 100% hydropowered. Because that's Sweden. They have abundant hydro with the fjords there. I believe if I recall correctly, Norway and Sweden is like roughly 99% of all their power comes from hydroelectric and also very common in this part of the world. Some of the heat from this deployment will be recycled for municipal heating. We had a really cool segment with Yaron Mellorud of Hash Labs, I believe a year ago at this point where they are deploying not in Sweden, but Finland, bitcoin mining capacity and then that heat is recycled into these municipal heating plants. So for some of these Nordic countries, they don't actually have atomized individual heating units within residences. They have these massive municipal plants that basically generate heat for an entire municipality. And then pipe it to the homes. And there was a reason for this. I can't quite remember what it was because it seems inefficient to me, but there's a specific reason that they do that. And now some of these data centers and bitcoin miners are actually capturing the waste heat from their sites and sending it to those municipal heating plants for distribution throughout the rest of the locality. Last point on this. This is Core Weave's second deployment in Sweden. It also has deployments in the uk, Spain and Norway. And again, we're starting to see some of these companies move out into Europe. We saw it most recently with Iron, also with their Australian site as well. But Iron signed a deal in Spain for I believe, what was like 400 odd megawatts. And so you're starting to see them grab power wherever they can get it. Also Europe, not to be left behind. They're still going to need inference computer if people want to use AI reliably in some of these locations. So there's opportunities there for these Neo clouds.
- [59:07]
C
Not much more to add there. The EU finally getting a piece of the pie. Not a very big piece and it's a very clean piece because it's very nuclear and water. But someone's got to do it. The US I guess, will bite the bullet. We will be the ones to extract the natural resources and put those in the, you know, power our beautiful, beautiful turbines with clean, clean natural gas. On that note, thank you so much for listening to Blockspace live. We do this Every single weekday, 1pm Eastern. Free train quick hits on AI, data centers, Bitcoin, emerging markets and tech, in that order. If you like what you hear. This turns into a newsletter and podcast. Podcasts anywhere. Podcasts are streamed and newsletter@newsletter.blockspace media.com this show is brought to you by CleanSpark. NASDAQ listed ticker CLSK I'm Charlie. I'm Colin and we'll see you tomorrow.