Bitcoin hash power shifts to AI - ForkLog: cryptocurrencies, AI, singularity, the future

# Bitcoin Hashrate Shifts to AI

The hype around cryptocurrencies has cooled, and “smart” money is rapidly flowing into AI infrastructure. Major miners are already converting data centers into neural network facilities, with construction costs skyrocketing by dozens of times.

How exactly the global economy is restructuring and whether it’s time to abandon traditional Bitcoin mining, in the “Podcast Society” interview, founder of Dataprana data center in the USA, Arseniy Grusha, shared his insights.

From Mining to AI Transition

ForkLog (FL): Are we truly witnessing the end of the crypto mining era, or is this just another transformation?

Arseniy Grusha (A. G.): Definitely not the end. Mining isn’t going anywhere, but the industry is undergoing a major transformation. Over the past five years, we’ve seen industrialization and the entry of large public companies, which caused a rally in Bitcoin network difficulty.

Today, giants in AI require enormous computing power. The profitability metrics in AI outperform mining significantly, so miners are massively converting their data centers for new tasks. This process will only accelerate.

FL: How feasible and difficult is it to retrofit mining data centers for AI needs?

A. G.: It’s very challenging and expensive because it requires completely different structures. The main advantage of miners is already dedicated capacity and legal connection to the power grid. Everything else has to be built from scratch with a different approach.

The cost difference is huge: a Bitcoin data center costs on average about $400,000 per MW, whereas an AI facility costs around $10 million per MW. The 20-25 times increase is due to strict requirements for full redundancy. AI data centers need complete backups: massive batteries and diesel generators capable of taking over load within seconds.

Design complexity also increases for liquid cooling systems for GPU capacities. All these systems require professional engineering and strict certification to ensure uninterrupted operation.

FL: Are backups necessary for continuous operation?

A. G.: Exactly, uptime is critical. It should be 99.9999%, meaning only a few minutes of downtime per year.

FL: So, miners’ main advantage is just access to electricity, and everything else they build from scratch?

A. G.: Yes, the key advantage is access to large amounts of electricity. US energy companies are slow, and bringing 100 MW online from scratch takes a couple of years.

Major IT corporations like Microsoft, Google, and Amazon pay huge sums for speed to gain an edge in the AI race. It’s more cost-effective for them to use existing mining facilities and retrofit them for their needs.

FL: Quantitatively, how much more profitable is it for miners to switch to AI?

A. G.: At Dataprana, we focus on colocation hosting and invest in infrastructure, not in ASICs, since equipment depreciates quickly. Traditional mining hosting allows recouping investments in three to four years, which is very fast.

Investing in an AI data center costs 25 times more. The payback period then extends to six or seven years.

FL: So, the focus is more on long-term?

A. G.: Yes, but the scale difference is also significant. Mining is a niche business with high volatility due to Bitcoin price dependence, with contracts typically lasting a couple of years at most.

AI is a global market with long-term contracts. A 10-year agreement with a company like Microsoft provides stability and makes it easier to attract bank financing for such projects.

FL: What happens to the Bitcoin network if miners massively shift to AI?

A. G.: I estimate that public miners control about 30% of the Bitcoin network, and they are actively moving into AI. This year, we will definitely see a drop in network difficulty and hash rate as large players start shutting down old equipment.

This will lead to an oversupply of mining hardware on the market and a natural decline in prices. The hardware will need to be stored somewhere, so demand for traditional Bitcoin data centers won’t disappear.

Eventually, the situation will stabilize, but in the next year, I expect a gradual decrease in network difficulty. Hardware manufacturers will also be forced to shift to self-mining to utilize their unsold inventory.

FL: Bitmain has already started cutting prices.

A. G.: Of course, because mining isn’t very profitable at current prices and network difficulty is at an all-time high. The market should balance out within three to six months.

The best time to invest in Bitcoin mining over the next four years will be this fall. Equipment prices will drop, entry barriers will lower, and investors could see 300–400% returns.

FL: Are chip quotas a big problem now? Is it better for providers to buy directly from Nvidia with long lead times or from resellers?

A. G.: Main buyers of GPUs are the “Big Seven” companies, controlling 80% of the market. They buy everything to stay ahead in technology or at least not fall behind.

The remaining 20% are split among small startups and brokers profiting from reselling scarce graphics cards. Chip development is accelerating rapidly, pushing the traditional Moore’s Law.

Despite new models, older cards still hold their value due to the huge demand for AI. Sooner or later, the market will saturate, and prices will fall, but the timing is uncertain. Currently, the AI segment is akin to the golden days of crypto: everything that can generate profit is working.

Energy Issues

FL: Is there a real electricity problem in the US and worldwide?

A. G.: In the US, there’s no energy shortage—Texas alone produces 85 GW, more than all of Germany’s consumption. The problem lies in outdated power grids not designed to deliver huge energy volumes to specific points.

Building data centers requires extensive upgrades of substations and transmission lines, which takes years. To balance loads, authorities are compelling AI companies to invest in their own generation stations.

There won’t be a global collapse; grid adaptation is happening gradually. But uninterrupted server operation is critical: if AI data centers go down, global payments and millions of workflows could halt.

FL: If there’s no global problem and Earth resources are sufficient, why are Elon Musk and Sam Altman interested in placing data centers in space?

A. G.: Space-based deployment offers two main advantages: automatic cooling and infinite solar energy without atmospheric losses. On Earth, about 30% of data center electricity is used just for cooling.

The main obstacle now is complex logistics and high costs of launching equipment into orbit. To build gigawatt-scale clusters there, rockets would need to launch as frequently as airplanes fly today.

I believe we’ll reach that point in 20–30 years. For now, humanity needs to learn how to efficiently build such unprecedented structures here on Earth.

FL: There were fears that AI queries consume tons of water. How true is that?

A. G.: Algorithms are becoming more efficient, but half a year ago, a standard ChatGPT query consumed about 3 W. So, 300 queries cost the data center roughly 10–30 cents, including electricity and infrastructure depreciation.

It’s not cheap, but costs will decrease as scaling continues. Meanwhile, demand for computations will grow thousands of times, since today we use AI capabilities at just a fraction of their potential.

Regarding water, data centers use closed-loop cooling systems. Water circulates constantly through pipes, cooling chips, and doesn’t evaporate, so actual water consumption is minimal.

FL: Has AI already hit a development ceiling?

A. G.: Not at all; we’re just at the beginning. Everyone is waiting for AGI (Artificial General Intelligence), which will surpass humans in all tasks.

Elon Musk predicts AGI within this year; other experts suggest 2027–2028. This superintelligence will radically change the world and take over many human functions.

Economy and Market

FL: Is the US a safe haven for infrastructure builders or a field of intense regulatory battles?

A. G.: The main issue in the US is bureaucracy: permits for construction take a long time, and connecting to the grid can take years. Neighbors in remote areas often oppose industrial projects, requiring land buyouts.

However, regulatory policies are very favorable. The government encourages construction through tax incentives because leadership in AI is a matter of national security.

In the global tech race with China, America is currently leading significantly. I hope this trend continues.

FL: Do you see investors turning more toward AI?

A. G.: Honestly, Bitcoin mining is now of little interest to investors. The industry resembles traditional oil extraction with high entry barriers, where quick 100x returns are impossible.

All capital is flowing into AI, as it’s the new industrial revolution. In 20 years, routine household tasks will be performed by robots, and we will delegate most daily functions to cloud-based agents.

FL: Do you think AI is already in a bubble?

A. G.: There is definitely an economic bubble, and it will burst—that’s a typical market cycle. It resembles the dot-com boom: Amazon’s stock fell from hundreds of dollars to $6, but the internet forever changed the world.

About 80% of companies will be wiped out because they produce clearly useless products. The main players will survive the crisis and continue aggressive growth.

I expect the decline to last about a year, not decades. AI and robotics are developing so rapidly that they will quickly pull the economy out of recession.

FL: A recent Citrini Research report predicted an economic collapse due to AI. What’s your take?

A. G.: That’s a plausible scenario. Consulting firms will lose relevance, and routine white-collar jobs will be replaced by automation. But AI won’t replace humans entirely—it will replace those who can’t use it effectively.

To mitigate social crises, there’s a 99.9% chance we’ll see universal basic income introduced within the next five years. To fund these payments, governments will need to impose heavy taxes on robots and corporations.

The economy will undergo painful restructuring over 10–20 years, with depressions along the way. But ultimately, the system should stabilize, hopefully in a way beneficial to humanity.

The Future

FL: What’s your view on sovereign AI? Will states build their own closed data centers, and will there be room for private enterprise?

A. G.: Sovereign AI will emerge with a 99.9% probability, similar to how sovereign internet appeared in China. Today, user data is the main currency and a powerful tool for political influence.

All countries will want to keep citizen data strictly within their borders. Governments will develop their own chips, AI tools, and search engines.

Creating such secure systems will take from five to twenty years. Government agencies always operate much slower than private businesses.

FL: Top 3 technologies that will change the world in the next few years?

A. G.: First, AGI. It will replace many employees and have a huge impact. Second, longevity. Medicine is advancing so rapidly that life expectancy could reach 120 years within decades. Third, space. In 5–10 years, space travel will become much more common, radically changing civilization’s consciousness.

FL: Any advice for entrepreneurs aiming to build infrastructure rather than just code?

A. G.: Infrastructure business is more complex than online services because it requires huge investments in “hardware” that can’t simply be moved if problems arise.

Main advice: build infrastructure only within a reliable legal framework. Avoid unstable countries chasing cheap electricity if there’s no guarantee of asset security.

Build where business can be legally scaled, where laws are predictable, and where investors feel safe putting their money.

The interview is summarized. More insights are in the full episode.