management to expand infrastructure, highlighted by a 10-year, $311M+ hosting contract with AMD. Installed mining capacity is set to reach 43.5 EH/s by Q1 2026, supporting liquidity and margins amid bitcoin price volatility. Despite dilution risk from future share issuance, RIOT trades at attractive valuation multiples relative to the sector and its long-term AI-driven upside. Riot Platforms, Inc.

Riot Platforms: AMD Today, Many AI Players Tomorrow
Summary Riot Platforms is rated Buy, driven by its strategic pivot to AI data center hosting while maintaining robust bitcoin mining operations. RIOT leverages its bitcoin holdings and flexible energy management to expand infrastructure, highlighted by a 10-year, $311M+ hosting contract with AMD. Installed mining capacity is set to reach 43.5 EH/s by Q1 2026, supporting liquidity and margins amid bitcoin price volatility. Despite dilution risk from future share issuance, RIOT trades at attractive valuation multiples relative to the sector and its long-term AI-driven upside. Riot Platforms, Inc. ( RIOT ) is a bitcoin (BTC-USD) mining company that chose to ride the AI wave. I don't see the same focus that Bitfarms Ltd. ( BITF ) has been using, a company I analyzed in a recent article . BITF is choosing a total transition from cryptocurrency mining to infrastructure for AI and HPC infrastructure. RIOT, on the other hand, is reconverting its three large mining farms into facilities to host AI services, but for the time being, it will not abandon Bitcoin mining. Even more, what I like the most about RIOT is that it's been using its bitcoin for that conversion. On January16 th it announced the purchase of 200 acres at the Rockdale facilities, Texas, for almost $96 million derived from the sale of 1,080 bitcoin. The announcement meant not only an extension of its operational capabilities, but also an agreement with Advanced Micro Devices, Inc. ( AMD ) to supply 25 MW during the first quarter 2026 with the possibility to expand the buying to 200 MW. The contract extends for 10 years and the company expects long-term revenues of $311 million, with the possibility of expanding revenues to $1 billion if all extensions are exercised. RIOT I think this was a great announcement for the company, and I hope it will be the first of many semiconductor companies or companies linked to AI development that will require RIOT's services. I hope that because I am rating the company as a Buy due to its future potential thanks to its electrical load management (which differentiates it from data center REITs) and because its installed capacity is not so far from reaching 1.7 or 1.8 GW based on its projections. RIOT The Advantage Of Buying Electricity Flexible management of electricity purchases for bitcoin miners is a key point in my investment thesis. In my previous articles I talked about how the possible increase in energy prices led me to buy companies with the adequate infrastructure to take advantage of that future potential, as could be the case of TransAlta Corporation (TAC). In fact, part of my thesis in that article was based on the deregulation of the electric market in Alberta. In the case of RIOT, this deregulation scheme is already working because bitcoin miners are using cheap energy in moments of lower demand, and to re-sell the energy they have bought in moments of higher demand. This is explained by the company on its own website , in the window about network support. Bitcoin miners, due to the network’s decentralized nature, can easily adjust power usage, unlike traditional data centers and manufacturers. Miners are the ideal consumer of new power generation because they provide a consistent, baseload customer during periods of low demand for power and also provide immediate, low-cost, demand-side flexibility during peak hours. For many experts on this subject this could sound repetitive. But for an investor like myself, who is not an energy engineer, for example, it is very important to me. This is one of the advantages to keep the bitcoin mining business, which of course will have the inherent volatility risks associated with a digital currency, but I believe the benefits will outweigh the risks. The Company Will Remain A Leading Producer Of Bitcoin On the one hand, I believe that not abandoning bitcoin mining will continue to be a fantastic source of liquidity. I reached that conclusion after the effective use of bitcoin for the extension of Rockdale facilities. Later, I'll talk about the company's financing for its capital investments, and I'll explain why I believe it is a risk that could postpone the upside of the stock. On the other hand, the advantage of remaining operating in the cryptocurrency market for RIOT is that it boasts a large installed mining capacity of 36.5 EH/s, as reported in Q3 2025, specifically on slide 44 of the 10-Q report. In Q3 2024 the company had 17.9 EH/s, an increase of 103%. RIOT For those who are not familiar with this terminology, the higher the EH/s, the more efficient the company will be at mining bitcoin while using less electricity. The expectation is to reach 38.5 EH/s in Q4 2024 and 43.5 EH/s in Q1 2026. Therefore, I feel that RIOT can maintain attractive margins in cryptocurrency while transforming into a premium AI hosting provider. In the EH/s analysis we must consider another key element: the global metric because if it goes up a lot, RIOT will be left with less bitcoin. It's an element of constant volatility, but I think the company knows how it works. I'll rephrase it: I am a bitcoin investor, ergo I also like mining. Then, if your savings are in gold or silver, I believe you'll be interested too in the mining companies of each metal. AMD Is Just The First Tenant: RIOT's Potential To Become A Premium AI Hosting Provider I believe RIOT will position as a great player in AI data centers because it already has the infrastructure needed to do it, and the connection with AMD is an example of that. If AMD was quick to sign a contract with the company, why wouldn't more similar companies follow suit and hire its services? The installed capacity of RIOT could reach 1.8 GW, and that's why it is a very attractive option for semiconductors, which need a powerful digital infrastructure in data centers to supply the demand of hyperscalers. The company's portfolio is based on three big facilities: Kentucky (162 MW), Rockdale (700 MW) and Corsicana (1000 MW). The largest power generation facilities are in Corsicana and Rockdale, although the three facilities need a strong capital investment to extend its capabilities. In fact, Kentucky , due to its operational capacity of 162 MW, actually has a potential capacity of 300 MW. Capital investment for 2025 was estimated at approximately $450 million across all improvements. Corsicana's facilities accounted for the majority of investment spending in 2025, excluding cryptocurrency mining costs. RIOT Corsicana represents 925 acres in Texas and, with the building development, Core & Shell, the company expects to accelerate the speed of adaptation and complete a custom-built data center by 2027 (slide 12 of the Q3 Presentation ). The other key establishment is Rockdale, the one that will host AMD and where the company just bought an additional 200 acres. The capacity to supply 25 MW for the first quarter of 2026 (and maybe 75 MW based on the agreement), convinced me that RIOT has the potential for rapid, phased deliveries. As I said at the beginning, there's no need to invest from scratch in a new data center, and the bulk of the investment, as you can see in the image above, it's still crypto mining. The Strategic Location Of Texas RIOT is the owner of almost 1,100 acres and most of it located in Texas (Corsicana and Rockdale), according to the statement of the agreement with AMD. Besides, it has a direct interconnection to an electrical network (which also belongs to RIOT, an operational 700 MW substation). This decreases the dependence on local energy. The locations of both facilities are also strategic because they are in the Texas Triangle, a very close position to technology hubs such as Austin and Dallas. Besides, Texas is the state with the most data centers of the country, only behind Virginia. Texas is home to several large data center projects, including the establishment of OpenAI and Oracle's Project Stargate in the city of Abilene, 180 miles from Dallas. For its part, Texas also has some cost challenges for its warm weather, but liquid cooling is playing a key role to avoid extreme heat generated by high-intensity AI servers, combined with high ambient temperatures. I believe RIOT has resolved this challenge thanks to its solid infrastructure and because it has a specialized cooling operator like Kelvion , which manages the Rockdale and Corsicana facilities. What I Took Away From The Q3 Results And What I Expect From BTC In 2026 The analysis of Q3 2025 results was a little late for me, but it helped me understand the company's revenue model so far. Bitcoin mining remains a strong total source of income. In Q3 it was $160 million out of a total of $180 million, while in the first 9 months of 2025 it was $444 million out of a total of $494 million. The improvement in EBIT and net income is an important factor in convincing me of my investment, but not necessarily the most crucial. Both metrics, along with revenue, you can see marked in yellow in the image of the 10-Q. RIOT Why did I say that profitability is important, but not fundamental? Because I'm thinking about RIOT as a long-term investment. Historically, RIOT suffers from the volatility of the bitcoin price, and profitability is relatively new to its balance sheets. In fact, just in 2024 its earnings were $109 million . In the first 9 months of 2025, net income was $27 million. I'm not expecting this metric to grow too much in 2026, and it could even fall due to investments destined to facilities improvements, and even due to the volatility of the price of bitcoin. For 2026, BTC projections vary widely between lows and highs, with an approximate floor of $75,000 and a ceiling of $225,000 . Costs per bitcoin in Q3 were $89,074 (see slide 45 of the 10-Q), representing an approximate gross margin of 21-22%. BTC experienced some sharp declines in the last quarter of 2025 from its highs of $121,000 and for now, seems to be far from that figure. If BTC trends downwards in 2026, to a minimum of $75,000, margins will shrink sharply if the company does not lower its mining costs. RIOT Anyway, I took into account the cost of depreciation of machines, which represents almost 50% of the cost per bitcoin. If RIOT extends the useful life of its machines, the margins could remain strong. Energy efficiency could also be improved through new investments, so even at the $75,000 minimum, I don't think it's a serious problem for my investment thesis. Long-Term Upside Potential In The Stock Price The company operates in a volatile sector that relies heavily on the price of Bitcoin to determine its margins. By comparing its ratios with those of the sector, the company might seem to be a little bit expensive, especially considering share dilution and the volatility of its returns. But compared to its ratios over the last few years, I think RIOT is cheap. The stock is trading at 10 times its sales, compared to the 5-year average of 18 times. The EV/EBITDA ratio is 21.89% below the industry median and extremely cheap compared to its historical average. The Price-to-Book of the last 12 months is also 47.32% below the sector average. Meanwhile, the stock is trading at 47 times its earnings compared to 37 times for the sector. I consider this could be due to the recent increase from the rise in BTC and even the recent agreement with AMD. Anyway, I believe its Price to Earnings ratio is not the ideal metric for a company that is betting for huge future income. As I explained throughout this article, I believe the stock is also cheap regarding its potential: trading in a range of $19-$22. Below that range, I consider it a good long-term buying opportunity. Furthermore, revenue projections, based solely on the AMD announcement, already estimate constant revenues of $1 billion. Other analyses also project total revenues of $1.6 billion by 2028. These projections fit very well with my bullish perspective due to the additional revenue from AI. Even more: in the following quarters the company will probably incorporate a revenue line in its earnings presentations related to AI infrastructure. Although AMD revenue already has a very good projection, I'm betting on seeing new contracts in the future. Based on the agreement with AMD, the supply capacity could be expanded up to 200 MW at Rockdale. But even in that case, there's still a remaining capacity just in Rockdale of 500 MW, and another 1 GW in Corsicana. If a contract similar to AMD's is signed in 2026 alone, it could reach the $1.6 billion estimate before 2028. I took into account that the increased use of installed capacity for AI data centers will reduce a little bit the energy required to min Bitcoin, but I believe that's precisely what the market will reward. In fact, that's something that has me very excited. I think the news about a new contract could have a very positive impact on the stock price and could lead to an upside of at least $24 or $25. My estimate is even below the analysts' consensus . Seeking Alpha The Risks Of Financing Through Share Issuance Issuing shares is a common practice for the company to finance its operations without resorting to excessive debt or the volatility of bitcoin. Given the maintenance of mining machines is expensive, the company needs a lot of cash flow. Now, with the adaptation to a first-rate energy infrastructure for AI data centers, the need for financing will continue to exist. That's why the past shared issuance rate caught my attention, and the trend I saw was deceleration. The year-on-year growth in issuance over the last 3 years is lower and more stable than the 2019-2021 period. MacroTrends I think this is a risk because the future perspective points out an issuance of more than $500 million, so 2026 could be another year of stock dilution if mining margins do not improve. Anyway, I don't think this is an issuance to cover operating losses or risks of failure, ergo I don't think it would bring problems to the business. The dilution is solely to create more value in the future. And I believe that the future value from AI data centers is worthwhile. Conclusions I am a fan of crypto, and I'm very passionate about the companies behind the phenomenon. RIOT Platforms is one of them, and I believe is one of the more efficient miners of the market. However, my article wasn't specifically about bitcoin, but about the step the company is taking to become a key player in the AI data center market. What caught my attention, then, was the idea of investing in the factories, not only in the final product. I think RIOT has the capacity to provide key services to AI players, and the agreement with AMD is a demonstration of my thesis. I hope that, over time, there will be more news about big technological companies requiring RIOT's services, especially because its infrastructure is already installed, it has a robust electrical supply network and intelligent energy procurement management, and liquid cooling to prevent high-temperature stress. I believe RIOT could be a risky investment both in the short and medium term, but since I'm investing for the long term, I consider the risks to be lower. AI is a reality, and the infrastructure that will keep it alive will come, to a large extent, from companies like RIOT, which will complement their volatile revenues with stable income thanks to their high-intensity data centers. I think a big part of that strategy is already underway, and that's why I'm investing in its potential.