Despite 2 Scary Words – Bitcoin And Diversification, I’m A Riot Platforms Fan

Summary Riot Platforms is a challenging stock. It involves two words should scare off the fainthearted… (1) bitcoin, and (2) diversification. For many, the word “bitcoin” is an automatic red flag… gambling, etc. But actually, it and crypto are gaining respectability and even functionality (as real money). RIOT is a big-time miner. The number of possible new coins is capped. We won’t reach the end till about 2140, but growth will decelerate as we progress toward then. So, it makes sense for RIOT to diversify. Like other miners, it’s eyeing the data center business, a power-hungry cousin to mining. Success here won’t be automatic for anybody. But…. I think RIOT will succeed. It has a lot of good land near Dallas, an important Tier 1 data center region. This contributes to my story oriented “Buy” rating. Remember when… Remember when Amazon ( AMZN ) sold books online. The company still does that. But it’s better known for other things today (selling countless different things, cloud). Remember when Microsoft ( MSFT ) was about operating systems. The company still does that. But it’s better known for other things today (cloud, application software). Remember when MicroStrategy sold enterprise analytics software. The company still does that. But with a new name, Strategy ( MSTR ), it’s now better known for other things (bitcoin treasury holdings). Remember when Riot Blockchain was was about bitcoin mining. It still is and has lately been getting getter at it. But the company rebranded in 2023 as Riot Platforms (NASDAQ: RIOT). That reflects its wish to diversify. RIOT can be uncomfortable for investors. It involves two words many find scary… I suggested one of them above… diversification. This sounds respectable, even essential, on its face. Investors constantly hear this when it comes to asset allocation. It can also serve companies well… so long as it’s done wisely. But many companies do it badly. As investment legend Peter Lynch says, one should guard against “ diworsification .” Rather than generating shareholder wealth, they instead make things worse… and generate accounting write-offs. Diversification might sound less challenging if we could at least understand and relax about RIOT’s core competency. But its established business confronts us with the other scary word. One, of course, is bitcoin. Many swear by it as the wave of the future in finance, as a store of value and as a better alternative to conventional money. Many others scorn it as nonsense or gambling. An investment case for or against RIOT stock should be based how one evaluates it in terms of the two challenging words, diversification and bitcoin. Let’s start looking more closely at the latter… My Stance Toward Bitcoin Has Evolved and is Now Much More Positive Than it was a Year or So Ago It took me a long time before I came around on bitcoin, and crypto in general. Like many, I thought it was nonsense at worst and at best, gambling without cool casinos, fun shows and countless restaurants. Too, it can be easy (and self-gratifying) to see market participants as an ignorant or dysfunctional herd. Folks like Ben Graham and Warren Buffet helped to legitimize that view with the Mr. Market allegory . But as I accumulate years of experience analyzing investments (since late 1979), I’ve learned more and more about how important it is to respect the market. Information has become increasingly plentiful. And different views are more likely to reflect different investment goals and risk preferences. True stupidity is quite rare. That helped me see bitcoin less as folly and more like a collectible… a digital collectible. How much different, once we adapt to the now-ubiquitous digital world, is a bitcoin than a gold bar or a baseball card picturing a great 1950s-era player. (See generally, here .) Of course, some things are more collectible than other things. The bursting of the 1600s Dutch tulipmania proved that. As it turned out, the more I learned about the potentially functional properties of crypto and blockchain , the less necessary it became to delve deeper into the philosophy of collectibles. RIOT holds some bitcoin in treasury (19,223 coins, amounting to about $5 per diluted RIOT share at bitcoin’s recent price of about $108,000.) But unlike MSTR, it doesn’t necessarily hold permanently. RIOT has recently shown itself to be willing to use its bitcoin to help it raise capital for other things. (See below.) And operationally… RIOT is One of the Major Domestic Bitcoin Miners. Bitcoin mining is a something of a brain scrambler. That’s so even for those who’ve learned to accept crypto as a legitimate asset. It’s a complex process. But it’s also quite interesting. (I explain more, in regular people language, in my 5/27/25 MARA report .) Here are the key aspects of mining you need to know to assess the investment merits of RIOT: Miners get newly issued coins by successfully answering and an absurdly complex mathematical puzzle. Success requires a monstrous amount of computing trial and error. This is every electricity intensive. So, mining profitability isn’t only about bitcoin prices. It’s also about controlling power costs and mining equipment energy efficiency. Each trial is called a hash. Mining power is expressed as “hashrate” (number of trial answers per second). More trials, higher hashrates, increase the likelihood of success. The number of possible bitcoins is capped 21 million. To postpone the end date as far as possible, the number of coins awarded per successful hash is periodically cut in half. As miners go, RIOT is quite good. Scale alone helps RIOT manage power costs. But its primary Texas location may be a much bigger plus. The deregulated power market there opens up opportunities for creativity. A big thing for RIOT is its power curtailment strategy . Based on this, RIOT shuts mining operations during local periods of peak electricity demand. It especially times this in relation to the June through September Electric Reliability Council of Texas (ERCOT) Four Coincident Peak (4CP) periods. But it doesn’t just stop trying to get bitcoin. RIOT also sells power back to the grid during high-demand intervals. With electricity prices spiking at such times, RIOT does well, perhaps even better than it could by mining, on these occasions. According to CFO Colin Yee on the 5/1/25 earnings call : Direct cost of mine, excluding depreciation, in the first quarter of 2025 was $43,808 per Bitcoin, of which power costs amounted to $35,313 per Bitcoin or 81% of total direct cost per Bitcoin. Quarter-over-quarter, our net power costs decreased from $0.038 a kilowatt hour to $0.034 a kilowatt hour as Riot’s power strategy continued to yield strong results. Another plus is RIOT’s dual cooling approach . It optimizes as appropriate between immersion and air-cooled systems. That’s another way to control power costs. Also, RIOT has a small Energy/Infrastructure operation. It contributes little in the way of revenues and profits. But it helps make RIOT vertically integrated . That means gives the company more than usual control over its mining hardware supply chain. The unit makes equipment specifically geared for RIOT’s needs. Meanwhile, RIOT has dramatically increased its hashrate over the past year. RIOT IR The company has, lately, spent heavily to buy cutting-edge mining equipment. An acquisition of sorts (a deal involving Corsicana – see below) added to RIOT’s already-large scale. So bitcoin mining, the company’s dominant current business, is on a roll. Analyst Compilation and Calculations based on data from pages F47 – F48 of latest 10-K and pages 31-32 of latest 10-Q Analyst Compilation and Calculations based on data from pages F47 – F48 of latest 10-K and pages 31-32 of latest 10-Q Analyst Compilation and Calculations based on data from pages F47 – F48 of latest 10-K and pages 31-32 of latest 10-Q But let’s remember… As Great as Things Now Look for RIOT Mining, This Business has a Ceiling RIOT’s forward hashrate expectations do not look especially buoyant. RIOT IR It could probably work harder to mine better… if it wished to. And as the slide notes, a recent acquisition (Rhodium) allows for some guidance upside (more below). Ultimately, though, the company strategically decided to refrain from adding much to mining capital spending. In fact, its latest 38.4 EH/s hashrate forecast is the result of management’s downward revision. It previously forecast 46.7 EH/s . (NOTE: EH/s means exahashes per second, or one quintillion hashes per second. A quintillion is one followed by 18 zeroes!) This relates to the last of the prior section’s bullet points. It bears repeating. The number of possible bitcoins is capped 21 million. To postpone the end date as far as possible, the number of coins awarded per successful hash is periodically cut in half. Don’t panic… Mining isn’t likely to max out until about 2140 . But halving events (periodic reductions in the number of coins awarded for successful hashes) will continue to make this a progressively harder and probably less lucrative business. That seems opposed to what a growth investors want. (Value investors don’t even get there… They’re likely stuck on the word “bitcoin.”) Rest assured RIOT management does NOT have its head in the sand about this. As noted above, it’s diversifying. To support this, RIOT is shifting resources and attention away from trying to max out on mining. That said… Diversification is not inevitable. The company could simply live with the halvings. Investors in natural resource royalty trusts and target date ETFs or mutual funds do something like that. (They face liquidation targets.) So, too, could bitcoin mining shareholders adjust (assuming they appropriately study up and really learn what’s going on). Combined with the RIOT’s owning bitcoin and serving as an asset play, this could work. So, RIOT can’t just diversify and insist all will be well. It also needs to diversify sensibly and well . Diversification must be better than alternative solutions. Diworsification won’t do! That’s why I referred above to diversification as one of the scary RIOT words. But there’s good news… RIOT’s Planned Move into Data Centers Stands an Excellent Chance of Success Demand for Data Centers in General Anybody with any interest in AI, HPC (High Performance Computing), etc. had to wonder about this earlier this year. Who among us could forget DeepSeek Apocalypse ( so named by me ). That happened when Hangzhou DeepSeek Artificial Intelligence ( DEEPSEEK ) announced it created a great AI model that cost much less than popular models then in use. AI related stocks tanked. So, too, did those that relate to data centers. Investors assumed we were overspending. Many observers and investors feared cutbacks, and lower data center demand. But eventually, analytic heads with boots-on-the-ground vantage points prevailed. And once-beaten-down stocks rebounded. On the 6/12/25 Oracle ( ORCL ) earnings call , Larry Ellison, that company’s Co-Founder, Chairman and CTO, was quite definitive: Let me add, we recently got an order that said we’ll take all the capacity you have wherever it is. it could be in Europe, it could be in Asia. We’ll just take everything. I mean we never got an order like that before. We had to move things around. We did the best we could to give them the capacity they needed. The demand is astronomical… [D]emand right now seems almost insatiable. I mean I don’t know how to describe it. I’ve never seen anything remotely like this. I mean people are calling up and asking us, please, can you find us more capacity. We’ll take it wherever. It’s in Malaysia, we’ll take it fine. We’ll take it there. We got some wherever. EMCOR Group ( EME ) management shared similar views. See my 6/20/25 EME report for more. RIOT agrees with this assessment. RIOT IR CEO Jason Les verbalized it thusly on the 5/1/25 earnings call : [W]e continue to be very encouraged by the feedback that we’ve received not just from hyperscalers but potential financing partners, which will be critical in this development as well. And everyone continues to recognize the value of power that can be delivered in the near term. In this case, we’re talking about 2026. The fact remains. There’s approximately 5 gigawatts of data center capacity being used for generative AI today, and the forecast is we’re going to need about 30 gigawatts by 2030. So that’s a pretty big delta to fill in a pretty short period of time. And as a result, we continue to see strong demand. Bottom line: Demand for data centers isn’t shrinking. It’s soaring. Nobody should quibble with this as a chosen field for RIOT’s diversification efforts. Of course, that alone won’t support an investment case for RIOT. We need to go further. We can’t naively assume every data center venture will succeed. Introducing the Data Center Business… plus Location Location and Location In general, if you hear of a bitcoin miner aspiring to data center work, start with a presumption of innocence. Bitcoin mining has a voracious appetite for power. Ditto modern AI- and HPC-oriented data centers. So, they’re compatible bedfellows. Focusing on AI, today’s main headline grabber, data needs vary. If you’re thinking about model training, think big… or rather HUGE! This is where we expose models to monstrous amounts of information. Using that data, the models learn how to recognize patterns, spit back old information, and answer new questions. Think of training as a university where AI models study. (And as in human universities, there’s continuing education. New information/data is always being generated. So models need to keep updating their training. In human terms, we refer to this as lifelong learning .) Then, there’s inference. This is where graduates go out into the world, get jobs, and do regular work. This is about inference. It’s what you’re doing if you use ChatGPT, Gemini, Perplexity, etc. It’s what Seeking Alpha does when it creates AI stock reports or Earnings Call Insights. Training tasks are monumental, but not necessarily frequent, Inference tasks tend to be smaller (by AI standards, which is still huge compared to non-AI tasks). But they’re frequent. Some data centers are huge. Others are smaller. Either way, we still need heretofore unimaginable amounts of power. Expect RIOT to work with larger centers. It’ll likely serve hyperscalers. (That term refers to the giants like Azure or Google Cloud.) Geography plays a big role in that choice. Some parts of the country are better than others for big data centers. Electricity regulatory climates play a big role here. So, too, do access to good power sources and well-educated work forces. (See generally, here and here .) Number one is Data Center Alley. That’s a nickname for a large northern Virginia concentration of data centers. RIOT is not there. But RIOT is right near Dallas Texas. That’s one of the Tier 1 U.S. data center markets. RIOT IR Dallas ranks number two , right behind Virginia. (See also, here .) Pivoting to Take Advantage of Its Texas-Sized Data Center Opportunity In 2022, RIOT bought a large land parcel, known as Corsicana, in Navarro County, Texas. It planned to create “ the world’s largest Bitcoin mining facility .” In May 2025, RIOT bought adjoining parcels. It thus more than doubled its Corsicana footprint. The latter … is part of Riot’s plan to build large-scale data centers tailored for enterprise and hyperscale clients. CEO Jason Les emphasized that these data centers require significantly larger footprints than traditional mining operations, underscoring the company’s ambition to diversify its revenue streams. To lead this initiative, Riot hired industry veteran Jonathan Gibbs as Chief Data Center Officer, signaling its commitment to expanding into the fast-growing market for AI-ready infrastructure. Here’s a visual presentation of the outcome of the land transactions. RIOT IR Les expects to serve big-league data center tenants at Corssicana: There are a number of hyperscalers and even large companies beyond hyperscalers that are in high need of power data center capacity to meet their growth targets. So I would say, especially because of the location of our site that allows it to be used for generative AI training, inference or cloud applications being 60 miles away from Dallas, a Tier 1 data center, there’s a lot of different things you can do at that site. So our view is the whole hyperscaler landscape and beyond that are potential tenants in this — for the Corsicana facility. As far as feedback, feedback from our advisers has matched really what we’ve gotten from tenants. And that’s why we’ve been working on improving the land, adding water and additional fiber connectivity. These weren’t things that were holding us back, but we’ve gotten feedback on what we can do to make the offering more compelling and put ourselves in the best position possible to be successful. As great as all this sounds, let’s maintain perspective… This is a development situation. It’s not a here-and-now money-geyser for RIOT. Even in Texas, as business friendly as it is, it takes time to get something like this off the ground. Here are the results of the study RIOT conducted to assess what it can do with Corsicana. RIOT IR And here’s a sense of RIOT’s path forward in data canters. RIOT IR RIOT IR So as intriguing as RIOT seems, let’s now switch gears. We need to discuss topics that’ll help keep us away from the dreaded irrational exuberance… Numbers (To the Extent We Have Any Worth Examining) If you love numbers, you’ll hate RIOT. This is not a pre-revenue development entity. But neither is it an old timer. It didn’t really get going, commerce-wise, until after the pandemic. Besides the small commercial sample, two more things make RIOT hard to quantitatively analyze. 1. Accounting for Bitcoin Operations Traditionally, companies had to write the value of bitcoin holding to account for any paper losses. OKI. Prices fall. Holdings lose value. I get it. But they were not allowed to write those values back up when bitcoin prices recovered! Seriously! Finally, in late 2023, the accounting powers that be recognized how silly that was. So, they changed the rule to allow companies to write assets back up when appropriate. (It took effect in late 2024.) Nice! Who could argue with more economic reality. But it still leaves workaday stock analysts in a quandary. Now we have two-way quarter-to-quarter profit fluctuations as bitcoin zigs and zags. Accounting standards setters pose no solutions for this volatility. I can’t think of anything either. So, it looks like we’ll just have to grit our collective teeth and cope. 2. RIOT’s Strategic Spending and Strategic Pivot Recall that Corsicana first became a gleam in RIOT’s corporate eye in 2022. It started out planning to seriously expand mining capabilities. Translation: big spending. And of course there’s the pivot to data centers. Translation: More spending. Numbers We Can Look At Given the above discussion, here are the only numbers I feel able to show you with a straight face. Analyst Compilation based on Data from Seeking Alpha Financial Presentations Notice the footnote below the table! You won’t be able to reconcile the operating margin I show with RIOT’s SEC documents. Nor will they match the Income Statement numbers Seeking Alpha presents based on what it gets from its data vendor. I did my own things there. I eliminated non-recurring items, especially the bitcoin valuation write-downs and write-ups. My solution is not perfect. I can’t eliminate bouncy numbers from ongoing bitcoin transactions. (Even “Cash from Operations” is impacted.) Let’s just say I’m doing the best I can. I think the best information you can extract from that table lies in the Cash etc. line and, in sections relating to Cash From Investing and Cash from Financing. You see there numerical equivalents to matters discussed above, mainly expansion of mining and the pivot toward data centers. You also see that RIOT needs outside capital to keep going. That’s also so for many of today’s emerging companies. Risk Needless to say, risk here is high. Conservative investors, look but don’t touch. Even risk tolerant folks should size positions appropriately (i.e., small) relative to your total portfolio. Any in-development venture is necessarily risky. Will it succeed? I explained why I think RIOT will succeed. But until that actually happens, it’s a source of risk. Further… Dependence on infusions of outside capital are always risky. But the markets will likely oblige and give RIOT what it needs so long as its strategy and progress remain credible. (That’s how it usually is in situations like this. See, e.g., the financial strength discussion in my 5/21/24 CVNA writeup .) Helping, too, is RIOT’s willingness to get creative — using its bitcoin — in terms of capital strategy. As Jason Chung, Head of Corporate Development and Strategy explained on the 5/1/25 earnings call (Note: ATM means at-the-market equity issuance): I think our financing activities this quarter could give a good sense of how we’re thinking about all the different financing options that we have available to us. So we saw extremely limited use of the ATM, and therefore, very minimal dilution to our shareholders. When we did issue into the ATM, we achieved an average share price of $13.05 per share. So we feel pretty good about our timing when we did utilize the ATM on a limited basis. But in addition to that, we took two other steps, right? In April, we started selling our monthly Bitcoin production as well, raising additional funds that way. And now with the Coinbase facility, we’ve opened up another avenue of financing for us that, again, side steps the ATM. What to do About RIOT Stock Don’t even think in terms of valuation ratios. There are none here that are meaningful. Here’s the price chart. StockCharts.com There’s most recently been a lot of very positive sentiment toward RIOT. We see that in the way the 10-day exponential moving average ( EMA ) has been above and running away from the also strong 50-day EMA. We also see it in Chaikin Money Flow ( CMF ) and the Chaikin Oscillator (CO). Both measure which party to trades is more motivated. CMF does it for institutional investors. CO does it for the market in general. Both indicators have been very strong. Buyers being more motivated than sellers typically exerts upward pressure on stock prices. But trees don’t grow unabated to the sky. We have to expect a near-term breather. (We’ve seen big ups and downs before. And we already see an indication of price fatigue in the very latest price bar next to the right-side edge.) All in all, nothing I see in the chart runs counter to my judgment regarding this stock. And in terms of that judgment… This is one of many emerging growth stocks today for which no meaningful valuation numbers can be calculated. It’s about the story. But it can’t be any story. It has to be a good story. And it has to be a credible story. My favorite example is value-guru William Miller’s very early investment into Amazon ( AMZN ) almost on day one. I’ve described other examples on 1/12/25 with RGTI, and on 6/10/25 with QS. I see RIOT as one among that group. When it comes to rating stocks, I think in terms of probable future stock performance relative to the market, rather than literally buying, holding or selling. And since I’m not rating based on a quant model, I’ll eschew “Strong …” extremes. Absent an exceptional degree of conviction, I think “Buy,” Hold,” or “Sell” are enough. I see RIOT as likely outperforming the broader 3- to 5-year market. I base this on the favorable stance I adopted lately towards crypto and bitcoin, and on my belief RIOT will succeed in its diversification plan. I translate that to the Seeking Alpha rating taxonomy by rating RIOT as a “Buy.”

Source: Seeking Alpha