AI Data Centers: A battle for energy

Artificial intelligence (AI) data centers are proving to be tough competition for the mining industry, specifically over cheap energy. AI is using more and more power. Training large language models, in particular, provides miners another obstacle to staying profitable. The BTC halving in April 2024, which cut block rewards to 3.125 BTC, shows how tight margins are for miners. This competition, mainly in energy-rich states like Texas, is changing the mining industry. Miners must develop new ideas or risk losing priority to AI, which has stable revenue and infrastructure needs.

The mining industry uses a lot of energy—about 187.9 terawatt-hours a year for BTC alone. That’s more than some countries. Similarly, AI data centers, which run on GPUs, need a lot of energy; the difference is that they offer steady income through long-term deals. Unlike ASICs, which are made for specific algorithms like BTC’s SHA-256, GPUs aren’t limited to one AI task. Miners like Hut 8 (NASDAQ: HUT), which received $150 million in 2024 to build AI infrastructure, and Iris Energy (NASDAQ: IREN), which is adding GPUs to their mining fleet, are using their current cooling and power systems to pivot into AI. This helps them generate revenue in a rapidly changing market.

The fight for energy is only increasing as more companies steer towards AI. Data centers use 3% of U.S. power, which could be 8% by 2030. This has already resulted in a strain on grids in mining areas like Texas and Wyoming. AI companies are waiting years to get grid connections and build their facilities. However, others are now looking at securing existing mining facilities. Core Scientific (NASDAQ: CORZ), for example, leased 16 MW in Austin for AI tasks, capitalizing on its infrastructure. Converting mining facilities to AI required significant investment. You need GPUs, better cooling, and skilled workers. This makes it hard for smaller miners.

In many cases, mining hardware requires complex liquid or immersion cooling as it produces a lot of heat. GPUs use simpler air-cooling. This makes conversions hard and raises costs.

The move toward AI is an opportunity for miners who can adapt their infrastructure. Those with data centers can rent out space to AI companies. This generates revenue to offset the declining mining profits. Core Scientific’s Austin lease is a good example of this. It mixes mining and AI to get the most out of its facilities. But retrofitting is expensive—often millions for GPUs and cooling upgrades. This creates a gap between big and small miners. Industry experts say miners who don’t adapt may have trouble getting energy contracts, as AI companies become better at negotiating with utilities.

The combination of digital currency mining and AI infrastructure shows that energy efficiency and versatility are essential in digital computing. Miners need to balance the short-term profits of mining with the long-term possibilities of AI. Investing in data centers that can handle mining hardware and GPUs can protect against market changes, but it requires a lot of capital.

Miners need to come up with new ideas to succeed. They can either hedge into AI or optimize their mining operations. The growing energy needs of both industries show why sustainable practices are paramount. By using renewables and flexible infrastructure, miners can reduce the threat from AI data centers. This will ensure they can survive in a market where energy is becoming increasingly competitive.

In order for artificial intelligence (AI) to work right within the law and thrive in the face of growing challenges, it needs to integrate an enterprise blockchain system that ensures data input quality and ownership—allowing it to keep data safe while also guaranteeing the immutability of data. Check out CoinGeek’s coverage on this emerging tech to learn more why Enterprise blockchain will be the backbone of AI.

Watch: How do you build a successful ecosystem? Bring blockchain to the builders!