Mining
Bitcoin Miners Shift Focus to AI Infrastructure, Outperforming Traditional Peers
The Bitcoin mining sector is witnessing a notable shift as companies that are transitioning towards artificial intelligence (AI) and high-performance computing (HPC) are significantly outperforming their traditional Bitcoin-focused counterparts. This trend has resulted in a cohort of AI-integrated miners achieving approximately 21% growth year-to-date, while the broader Bitcoin market has remained stagnant.
Core Scientific exemplifies this transition, reporting $136.7 million in colocation revenue from AI workloads in Q2 2026, starkly contrasting with just $27.5 million from Bitcoin mining. Other companies, such as TeraWulf and RIOT, have also seen substantial gains of 73-74% and 94% respectively in the same timeframe.
As public miners redirect their resources, the realized hashrate has dropped by about 13.4% from Q4 2025 to Q2 2026. The valuation gap between miners with AI exposure and traditional Bitcoin miners is significant, with AI-focused miners trading at over 12 times expected forward sales, compared to 2 to 6 times for their Bitcoin-only peers.
The economics favor AI as well, with cloud services generating a median of $940 per megawatt-hour, while Bitcoin mining yields only between $113 and $179 per megawatt-hour. The cumulative contracts for AI and HPC in the mining sector are estimated between $70 billion and $100 billion, although actual billing is limited to around 550 megawatts of deployed capacity.
This transition comes with challenges, including a projected $50 billion funding requirement to bridge the gap between contracts and realized revenue. Some companies have exited Bitcoin mining entirely, while others maintain hybrid models. Investors in AI-exposed miners are essentially buying access to valuable power infrastructure, which is becoming increasingly scarce for large-scale AI deployments.
Latest Developments in Crypto Trading
Elon Musk's X platform has introduced an enhanced Cashtag feature for US users, allowing them to connect in-feed asset discussions directly to trading interfaces on partner exchanges such as Coinbase, Gemini, and Kraken. This feature went live on September 22, 2026, enabling users to tap on cashtags like $BTC to view live price charts and related posts, along with a "Trade" button that redirects them to partner brokers for order placement.
Unlike traditional brokers, X does not handle trade executions or custody assets; it acts as a data layer and referral engine, providing real-time pricing information and routing users to licensed exchanges for transactions.
This launch builds on X's earlier Smart Cashtags rollout in 2026, which provided real-time pricing data for cryptocurrencies. The new feature simplifies the process of buying Bitcoin, reducing the steps from multiple app interactions to just two taps.
Additionally, Musk's companies, Tesla and SpaceX, reportedly hold Bitcoin, Ethereum, and Dogecoin, and his social media activity has historically influenced crypto markets.
New Insights from the G20 Innovation Ministerial
- Nvidia CEO Jensen Huang stated that building a single gigawatt of AI computing infrastructure could cost between $50 billion and $60 billion.
- The G20 Innovation Ministerial took place on September 1-2, 2026, co-hosted by the US Department of Commerce and the White House Office of Science and Technology Policy.
- Key figures in attendance included OpenAI's Sam Altman and Anthropic's Tom Brown, highlighting the importance of cross-industry collaboration.
- Two significant outcomes were the "Carolina Principles for Emerging Technologies" and the "AI Prosperity Objectives Compact," focusing on research investment, commercialization, trusted adoption, and public-private partnerships.
- The US delegation advocated for a hands-off regulatory approach, suggesting that regulation should follow the adoption of AI technologies rather than precede it.
- Huang emphasized the need for AI infrastructure to be durable, fungible, and software-upgradeable to ensure longevity and adaptability.
- The consensus statement included a broad wish list for pro-innovation policies, workforce development, and intellectual property frameworks for AI.
New Developments in Bitcoin Advocacy
At the United Nations Digital Cooperation Day 2026, Bitcoin was featured prominently in discussions about its potential as a digital public infrastructure. This event took place on September 21, during the UN General Assembly High-Level Week.
The panel, titled "Blockchain and the Tokenized Future: Sovereign Choices and Safeguards for Next-Gen DPI," included notable speakers such as Jeremy Almond, CEO of Paystand, and Bilal Bin Saqib, Chairman of Pakistan’s Virtual Assets Regulatory Authority.
The core message emphasized Bitcoin's role in empowering developing nations to create self-sustaining economic systems, moving away from reliance on traditional aid.
Pakistan's involvement highlights its push for global regulations on digital assets, particularly in relation to sustainable development, given its large population and limited banking access in rural areas.
This event builds on previous UN engagements in 2026, including discussions on legal frameworks for digital assets by the Bitcoin Policy Institute at UNCITRAL.
While the UN did not endorse Bitcoin as a reserve asset, the positive discussion around its use cases for financial inclusion marks a significant shift in institutional attitudes toward cryptocurrency.
New Developments in AI and Cryptocurrency
- The Grayscale AI Compute ETF commenced trading on September 18 under the ticker GCPU.
- The ETF launched at a price of $37.76 per share and started with approximately $12.46 million in assets under management.
- Unlike traditional ETFs, GCPU does not hold equities but instead invests in GPU compute futures contracts.
- The fund employs an actively managed strategy, allowing Grayscale's team to select contracts and manage roll timing.
- GCPU's performance will be influenced by the prices of compute capacity and the efficiency of contract rollovers.
- As a small fund, GCPU is categorized as needing to exceed $50-100 million in assets to achieve long-term commercial viability.
New Insights from Jeff Walton on Strive's Bitcoin Strategy
- Strive has emerged as one of the best performing stocks in the Russell 2000 over the past month.
- Chief Risk Officer Jeff Walton indicates that Strive is on track to double its Bitcoin holdings approximately every 12 weeks.
- The company's balance sheet has recently surpassed $2.5 billion in total strength.
- Strive utilizes common stock and preferred ATMs to fund its Bitcoin accumulation strategy.
- There are $700 million in warrants expiring on October 13, which could significantly impact leverage and future product offerings.
- Walton highlights a systemic credit risk that he believes the market is currently underpricing.
FAQ
What is driving the shift of Bitcoin miners towards AI infrastructure?
The shift is primarily driven by the significantly higher profitability and growth potential of AI and high-performance computing (HPC) compared to traditional Bitcoin mining. Companies transitioning to AI are experiencing substantial revenue growth, while the Bitcoin market has remained stagnant.
How much revenue did Core Scientific generate from AI workloads in Q2 2026?
Core Scientific reported $136.7 million in colocation revenue from AI workloads in Q2 2026, which is notably higher than the $27.5 million generated from Bitcoin mining in the same period.
What are the growth percentages for AI-integrated miners compared to traditional Bitcoin miners?
AI-integrated miners have achieved approximately 21% growth year-to-date, while traditional Bitcoin-focused miners have not seen similar growth, with some companies like TeraWulf and RIOT reporting gains of 73-94%.
What is the valuation difference between AI-focused miners and traditional Bitcoin miners?
AI-focused miners are trading at over 12 times expected forward sales, while traditional Bitcoin miners are trading at only 2 to 6 times, indicating a significant valuation gap.
What challenges are associated with the transition to AI in the mining sector?
Challenges include a projected $50 billion funding requirement to bridge the gap between contracts and realized revenue, as well as the need for valuable power infrastructure, which is becoming increasingly scarce for large-scale AI deployments.