AI has arrived. There’s no question about that. The stock market is near its all-time highest levels, largely thanks to FAANG, as corporations are reducing their staff by half in order to put cash into hardware. OpenClaw is a self hosted AI agent that has more stars than Linux or React on GitHub. Jack Dorsey, too, wants to restructure Block because of the digital and artificial intelligence. How much of the AI hype is real, and which companies will reap the rewards?
The United States public Bitcoin miner community has made its choice. Some have even gone so far as to rebrand themselves, making a complete break from the orange coin. There are many AI-related statements and pivots made by Bitcoin public miners, but a few stand out.
Cypher Mining, estimated to be worth around six billion dollars — placing it among the biggest in the country – announced a full rebrand away from Bitcoin and on to the AI hype train. According to their latest report, titled “Rebrands to Cipher Digital to Reflect Strategic Shift Toward HPC,” Why the Company Explained “Divested 49% Stake in Alborz, Bear, and Chief Mining Sites”. Bitfarms Ltd., another public mining company valued over one billion dollars, has also made an AI-centric pivot. Ben Gagnon went so far as to say that saying “We are no longer a Bitcoin company,” as reported by Coindesk, though they did keep the ‘Bit’ in the name.
Some companies, especially those that are public, expect higher dollar returns than Bitcoin in the short- to medium-term, while others may consider this a good diversification opportunity or one they cannot afford to miss.
Kent Halliburton — Co-Founder & CEO of SazminingBitcoin Magazine, in an exclusive Interview. “The average cost to mine a bitcoin right now is about $87,000. The spot price of bitcoin is about $70,000. So most of the industry is underwater, and the public miners are using that as their excuse to pivot.” Sazmining, a Bitcoin mining company that is private and focuses on frontier energy resources with most of its operations outside the United States.
Halliburton also noted that “$87,000 is an industry average — it includes guys running old-gen rigs on grid power in Texas. At our sites in Paraguay and Ethiopia, our clients are producing bitcoin on an energy cost basis of $50,000 to $64,000, on 100% renewable energy. That’s a 10 to 30 percent discount to spot. The profitability is right there.” The only thing that is required for American public miner to take action would be a longer time horizon, or cheaper energy.
Halliburton says that on the subject of energy costs, public U.S. mines have the opportunity to compete but failed to do so. Halliburton was blunt in his assessment of the public mining companies. “had the power contracts, the land, the infrastructure — everything you need to mine bitcoin cheaply — and they’re handing it to Microsoft and Google in exchange for lease checks. They went from securing the Bitcoin network to securing rack space for hyperscalers, and they’re calling it a strategy. Meanwhile, they’ve dumped over 15,000 bitcoin off their balance sheets to fund the transition”.
IREN Limited, one of the largest public Bitcoin miners, began its mining operations in 2009. pivot to AI cloud services in April 2025Announcement a$9.7 billion, five-year agreement with Microsoft TeraWulf has executed multiple projects using NVIDIA GPUs. TeraWulf executed several Google-backed HPC expansions Fluidstack secured 10 year contracts for more than 200 MW.
Cipher Digital, a HPC-landlord with 600 MW in contracted capacity has finished its full rebrand, which included a 15 years, 300 MW, lease with AWS, and a ten year, $300 MW, lease with Fluidstack, backed up by Google. Hut 8 also signed a Google-backed 15-year lease for 245 MW with Fluidstack. Future extensions are possible, as is a first right offer of over 1,000 MW. Core Scientific’s expanded its HPC focus to 270 MW CoreWeave is a partner that supports Microsoft OpenAI and Microsoft workloads.
Riot Platforms evaluates an expansion in AI hosting by working with AMD on an operational 10-year, 25 MW lease Assessments for 600 MW of AI/HPC at its Corsicana siteAlthough no agreements with hyperscalers have been made public, they are expected to be announced soon.
MARA Holdings, LLC is a holding company that owns and operates MARA Holdings. diversifying into AI Starwood Digital Ventures has formed a joint enterprise with Starwood Capital to target 1 GW near-term of IT capacity that can grow up to more than 2.5 GW when used for AI and hyperscale workloads. Starwood is leading the financing, and Starwood leads tenant sourcing. However, there have not been any hyperscalers contracted yet.
CleanSpark is a reputable company that offers a variety of services. pursuing a pivot to AI by acquiring Texas land and power for AI/HPC, including 447 acres in Brazoria County for 300–600 MW potential and an Austin County site contributing to 890 MW aggregate, funded by Bitcoin sales, with tenant discussions ongoing but no hyperscaler leases disclosed.
So the AI gold rush is here, there’s no doubt about it, many of these public miners apparently see an opportunity to build out the infrastructure of — what is without a doubt— a profound technological trend. History hasn’t been very kind to people who try and build infrastructure for a new age, at least not over the long-term. This is a high-risk bet with a medium reward. How many of the companies that built the railroads — for example — are still around today? Can you, even without looking back so far, name a company who built internet fiber in the 1990s or 2000s?
It is not a long list of railroad bankruptcies The late 1800s saw a severe financial crisis, known as the Panic of 1873Many people took on too much debt in order to finance constructions that were not yet needed. After the panic J.P. Morgan led a consolidation They were able to restructure the debt of railroad companies in bankruptcy, and take over their assets. The railway construction was a success for them.
Just around the corner, we are still experiencing the effects of the dotcom bubble that dominated the early 2000s. a graveyard of fiber line infrastructure companies Hyper scalers, like Google and Meta, also bought them out in the final stages of the consolidation post-crash for just pennies per dollar.
While both the railway and fiber line build-outs overall helped scale commerce for the world in incredible ways — demonstrating the overall wisdom of the markets — most individual companies involved did not survive the process, and venture capitalists looking at the AI boom today are aware of this dynamic.
The Capex to Revenue AI Gap
Investor groups have begun to wonder where this huge infrastructure investment will be repaid. In an October 2025 report titled “AI: In a bubble?”GoldmanSachs has taken a position that valuations for some of the companies are starting to fall. “frothy”.
David Chan of Sequoia pointed out the increasing gap between AI driven revenue and capital expenses (Capex) starting in 2023. This led to an widely reported $600 Billion gap. Capex spending commitments in 2026 are north of $700 What are the benefits for hyperscalers but what about the return?
OpenAI’s annual recurring revenues (ARR) of $20 billion are impressive numbers for a brand new company. “roughly 3% of the projected 2026 hyperscaler capex total” According to FuturumGroupThe author noted: “Anthropic’s $9 billion run rate, while showing 9x year-over-year growth, occupies a similar position. The entire cohort of pure-play AI vendors – including Cohere ($150 million ARR), Mistral (~$400 million), Perplexity ($148 million annualized), and others – likely accounts for less than $35 billion in projected combined 2026 revenue.”
A number of people have also voiced their skepticism regarding the actual value that AI can bring. VC’s like Chamath Palihapitiya. He was an investor who was prominent in Groq. This company was building custom silicon to meet the needs of AI. licensed by NVIDIA in a $20 billion deal Last year I was an insider of Facebook, witnessing its rise from a small company to a global player. Perhaps if he is unsure about whether building railroads for artificial intelligence will be profitable, it’s worthwhile to take a closer look.
Palihapitiya argued, in a podcast by All In Podcast recently, that companies might start realizing this soon. they are exposing their trade secrets to cloud AIInstead, they prefer to host themselves. Can you risk your business secrets by allowing AI companies to train using user data when building out an in-house GPU farm? The data used to train new models will be exposed in the knowledge base of AI providers. It is possible that corporate agreements to refrain from training on corporate data will become the standard. However, this would create a relationship of high trust which poses a significant risk for certain companies.
Also, there are questions as to whether cloud AI is fundamentally desired by the market for the same reason. You wouldn’t hire a personal assistant knowing that the information you shared with them was going to be on the Internet. AI is a good example. It’s true that AI is a growing field. U.S. Southern District of New York recently ruled It is important to note that AI chatbots do not offer the same level of protection as traditional lawyers. As a result, any sensitive information exchanged with AI may be used in court against clients. Some believe that new types of agreements and terms will be needed to accommodate this application. The legal case highlights a key element in the AI demand: the desire for humanoid intelligent systems, whether digital or not.
AI Loyalty and Trust
Ah, “Trust”The word “trust” is ubiquitous and almost superstitious. It does so much to support the weight of our world. What exactly is trust? It is fundamentally about predictability. One person’s trust that another AI, human or system will act in a predictable and reliable manner towards their interests. AI in the cloud can’t give these assurances because the data will be processing by someone else. “the cloud,” It is almost beyond mortal imagination what we can do. It is true. “the cloud” It may not be loyal in some scenarios due to legal issues. OpenClaw is an intriguing product, which may explain the fascination of many people.
Recently, an open source AI project that has been gaining momentum in recent weeks has taken the industry by surprise. 289,000 stars are more on GitHub than Linux, despite it supporting software infrastructure around the globe, and more than React, the most popular language for web development. How long has it been available? How is this possible? Why do people like it so much?
There are two main reasons. It is more human-like than a bot; it remembers your interests, keeps a journal, and adapts to suit you. It can be hosted on your computer, which is the most important thing. OpenClaw users bought Mac Minis by the thousands, paired up with Claude Max API plans for about $200 a year. Others claim this is the future of self-hosting despite the fact that the setup above is still cloud dependent. What’s happening is OpenClaw is loyal and remembers your name. “in your home” In your PC. The chat will not die a slow death because the context window is too large. Instead, a brand new tab for chats replaces it. OpenClaw does not work like a typical chatbot. Instead, it is an AI-based entity with which users build relationships. Trust is the foundation of any good relationship.
What does this all have to do public Bitcoin miners? Perhaps self-hosted AI will be the next big thing. Chinese AIs are becoming leaner. They can also run on less advanced machines. This is largely due to sanctions imposed on expensive AI chips like Nvidia’s. All kinds of open-source tools are constantly being launched to manage and host local models. And if the history of AI is anything to go by, then the mass production and distribution of AI hardware may lead to the commoditization and standardization of powerful machines that can run AI.
Apple, FAANG’s worst AI product provider to date, is likely to be one of AI’s biggest winners. Why? Their user hardware is outstanding. Macs of recent years don’t distinguish RAM from VRAM. All other computers that depend on GPUs (such as Nvidia) do. It limits the speed and size of the models which can be hosted by themselves. In the new Macs, RAM is all unified, so users can run models that are difficult to run on non Apple hardware. In the future, self-hosted AI will be the norm.
In their pursuit of short-term gains in fiat currency, it is possible that public Bitcoin miners have fallen into a trap. It’s the same trap that the titans of dotcom fell into. This is the same trap the giants who constructed the railroads in the Industrial Era fell into. Infrastructure is not always the key to the future.
“DagelijksCrypto is not responsible for any activities you perform outside DagelijksCrypto.”
