AI Crypto Mining: Bridging AI & Blockchain Technology
The Growing Demand for Computing Resources in AI Crypto Mining
As the influence of AI crypto mining expands across various industries, the demand for computing resources to fuel this machine intelligence grows rapidly.
Training Advanced AI Models: A Costly Endeavor
Training advanced AI models like ChatGPT requires a substantial investment, costing more than $5 million. Even the early ChatGPT demo, prior to its current level of usage, incurred a daily expense of approximately $100,000 for OpenAI. AI extends beyond text generation, as its practical application in diverse sectors necessitates large neural models trained on various data types, including medical, financial, customer information, and geospatial data. Advancing from current neural net AI limitations to systems with higher artificial general intelligence levels will undoubtedly require even greater computational power.
Repurposing Computing Infrastructures for AI Crypto Mining
A small but increasing number of crypto miners are now exploring ways to utilize their existing computing infrastructures to contribute to the advancement of AI crypto mining.
Bitcoin (BTC) mining continues to be a profitable venture, although mining other cryptocurrencies can be more volatile due to the rapidly evolving landscape. For example, Ether (ETH) miners faced significant challenges when the Ethereum network transitioned from proof-of-work to proof-of-stake.
Over the past two years, the economic and technical circumstances in the crypto space have led more mining organizations to investigate the potential of repurposing their facilities for other purposes, particularly high-performance computing (HPC) and AI.
Mining Facilities: Potential for High-Performance Computing and AI
While the computing hardware required for HPC or AI processing often differs from what’s ideal for crypto mining, acquiring servers is generally not the most challenging aspect of setting up a mining farm. The major costs and efforts lie in establishing electrical power, cooling, security, and other physical infrastructure. These factors remain relatively consistent, whether one hosts RAM-light GPUs suitable for ETH mining or RAM-heavy GPUs suitable for AI model training.
Hut 8, a mining firm, has taken the lead by repurposing its dedicated compute facilities for machine learning and other HPC applications. Hive Blockchain has also pursued a similar strategy, utilizing processor cards in their servers that can be used for cloud computing, AI applications, rendering for engineering, as well as scientific modeling.
The Convergence of Blockchain and AI in Crypto Mining
One intriguing possibility is for miners to redirect their compute resources to AI within the blockchain space itself. This can be achieved by running AI processes hosted on decentralized blockchain-based networks.
Redefining Possibilities: Running AI Processes on Decentralized Networks
Several AI projects linked to their own altcoins, including Fetch.ai (FET), Ocean (OCEAN), Matrix AI Network (MAN), Cortex (CTXC), and SingularityNET (AGIX), along with ecosystem projects like NuNet (NTX) and the ledgerless blockchain HyperCycle, provide this opportunity. In 2023, AI-related altcoins have experienced success as the market recognizes the potential of decentralized AI software.
Blockchain’s Expansion Beyond Finance: Applications in Various Industries
It has been evident since before the release of Bitcoin’s white paper that distributed computing, strong encryption, and decentralized control have applications beyond finance. Blockchain projects have emerged across a wide range of vertical markets, such as medicine, supply chain, gaming, and robotics. As AI becomes the dominant force in each of these domains, decentralizing the software and hardware supporting AI will be crucial for decentralizing the global economy. Repurposing a portion of crypto mining hardware to support AI processing, some of which operates within AI-oriented crypto networks, will increasingly become part of this narrative.
The Implications of Global AI Processing in Crypto Mining
If a significant portion of global AI processing occurs within crypto mining facilities, it could have implications extending beyond the financial realm. Crypto mining rigs are located in diverse legal jurisdictions and owned by various entities.
A globally distributed AI network spanning crypto mining rigs would present considerable challenges for governments or centralized entities to control, unlike an AI network primarily centered in Big Tech-owned server farms, which is the current norm.
Whether this decentralized AI network scenario is beneficial or raises ethical concerns depends on one’s perception of the intentions of Big Tech and big government.