Conversation with Bit Deer: Considerations Behind Mining Companies Transitioning to AI Data Centers

Author: Lin Wanwan

Everyone initially did not realize that the real bottleneck of AI is not capital, not large models, but electricity.

Long-term full-load training and AI inference running 24/7 has brought about a problem: there is not enough electricity, and chips are forced to be sidelined and gather dust. The U.S. power grid infrastructure has lagged behind in the past decade, with new large-load grid connections taking 2–4 years, making “ready-to-use electricity” a scarce resource across the entire industry.

Generative AI brings a raw and brutal truth to the forefront: what's lacking is not the model, but the electricity.

The story thus took a turn, and cryptocurrency mining enterprises, this group of people who were the first to regard electricity as a “means of production,” began to move from the margins to the center of the capital stage.

Iris Energy (IREN) is a sample of this route. This year, IREN's stock price surged nearly 600% at one point, with a 52-week range from $5.12 to $75.73. While the Bitcoin price remains attractive, it decisively withdrew power to transform its self-built AI data center.

When giants like Microsoft come forward with long-term orders worth a total of $9.7 billion, the market intuitively understands for the first time the realistic path of “from mining to AI”: first electricity and land, then GPUs and customers.

But not all mining companies, like IREN, choose to bet their entire fortune on AI. In this power-driven migration of computing power, there is also a stable force worth our attention - Bitdeer.

Bitdeer Technologies Group (NASDAQ: BTDR), a company founded by crypto legend Wu Jihan and headquartered in Singapore, holds nearly 3GW of power resources globally and has avoided the shallow trap of relying on others for “power supply” from the very beginning. As the wave of AI arrives, Bitdeer did not choose the IREN-style aggressive “All-in”, but instead retained its highly profitable Bitcoin mining as the “foundation”, while steadily upgrading some of its mining sites into AI data centers.

This strategy of “advancing to attack and retreating to defend” makes it the best sample for observing how global players think and strategize in this competition of computing power.

To this end, we interviewed Wang Wenguang, Vice President of Global Data Center Business at the mining company Bit Deer, hoping to gain insights into the current global power shortage for AI, as well as their views on mining companies transitioning to AI data centers, whether it is capital speculation or a real demand for AI. We conducted an in-depth conversation on this series of questions.

Why is the power shortage in the United States so severe?

Dingcha: First, let me ask a basic question about the overall direction, do you think electricity prices will continue to rise in the future?

Bit Deer: I think so, because this is a very important supply and demand relationship for the future.

Dongcha: Regarding the electricity shortage in the United States, there is a saying in the market that it is difficult to obtain a “power license” in the U.S.?

Bit Deer: It's not that this so-called “electricity permit” can't be approved, but rather that the physical speed of the grid expansion can't keep up. In the years following the relocation of heavy industry from the United States, the construction of the U.S. power grid has not been systematically expanded. After mining companies moved to the U.S. in 2021, much of the electricity that was “already connected to the grid and had signed PPAs” was locked by mining companies. With the influence of ChatGPT, pure AI players come in and discover that there is a large amount of electricity available for immediate use in the mining sites.

This explains why large companies are willing to collaborate with mining enterprises; rather than waiting 2 to 4 years to build 500MW from scratch, it is better to transform the existing park in 12 months.

DONGCHA: When did the industry truly realize that “inference also consumes a lot of power”?

Bit Deer: Probably after the widespread adoption of GPT-4. As companies embed models into customer service, office work, search, risk control, etc., the demand for reasoning becomes long-term and scenario-based, and the power consumption has not decreased as initially envisioned.

This brings about two types of changes.

One is the engineering upgrade: from stronger air cooling to liquid cooling / hybrid cooling, the cabinet power, distribution path, fire protection, and monitoring have all been elevated to the AI data center water level.

Another aspect is resource strategy: electricity has become the real number one bottleneck. People are no longer just talking about “buying cards”, but instead focusing on obtaining electricity and grid connection, long-term contracts (PPA), grid connection scheduling, interregional capacity backup, and when necessary, obtaining electricity upstream like mining companies (self-generation / direct procurement).

In fact, we have seen the same trend in the mining industry for a long time. Chips can be infinitely expanded (silicon comes from sand), but electricity cannot be expanded. We have done natural gas self-generation in Canada to power the mining site, following this logic. Today's AI is almost identical.

Moving Observation: What are the differences in electricity consumption scale between AI data centers and traditional internet data centers?

Bit Deer: It's not just a change in quantity, but a change in scale. In the past, a traditional internet data center of 20-30 MW was considerable, but now AI data centers demand 500 MW or even 1 GW. AI has transformed data centers from “rack business” to “power engineering,” requiring everything to be recalibrated: circuits, substations, cooling, fire protection, redundancy, PUE… The experience from traditional internet data centers is still useful, but it's no longer sufficient.

Insight: Why has “electricity” become the most scarce element in upstream?

Bit Deer: Chips can be expanded because they come from silicon and capacity management; power is difficult to expand because it comes from power generation and grid upgrades. In the past, mining has tried to “look upstream for energy,” including self-generated power projects in Canada; the path of AI is similar to this—whoever can seize the power first will have the advantage in deployment time.

AI New Battlefield: From “Grabbing GPUs” to “Grabbing the Power Grid”

Action Insight: Mining companies are transforming into AI data centers. What exactly needs to be changed? Previously, it was said that “Bitcoin mining power can be used for AI”, but mining chips (ASICs) are not compatible with the GPUs required for AI. So why are mining companies now able to “provide AI computing power”?

Bit Deer: Global mining was once divided into two. Bitcoin relies on mining chips ASIC, which are efficient but have a single use; Ethereum relies on NVIDIA GPUs, which are versatile but have exited the mining stage after transitioning to PoS.

So, the so-called “mining farms turning to AI” in the market today almost all refer to Bitcoin mining farms undergoing transformation. The core point is that the mining farms no longer “compute hashes”, but upgrade themselves into AI data centers.

This is an upgrade of infrastructure, replacing ASIC racks with GPU servers; upgrading the “just enough” power system to a professional-grade power distribution system with N+1/2N redundancy; upgrading traditional air cooling to a cooling system capable of supporting high-density GPUs; and standardizing and auditing the facilities for sealing, dust-proofing, fire prevention, etc. in the data center.

Complete these four steps, and the crypto mining farm will transform from a “mining workshop” into an “AI server room.”

Why can mining companies build faster than AI giants? Power.

AI is a business of “electricity and heat”, and the time frame for building AI data centers is 3-4 years, with time cost being the biggest barrier. Mining companies happen to hold these “hard assets”, thus they are at an advantage in the transformation process.

Insight: In recent days, Microsoft and Amazon have consecutively signed long-term AI contracts with cryptocurrency mining companies. Iris Energy (IREN) signed a contract with Microsoft worth a total of $9.7 billion over 5 years; another company, Cipher, signed with Amazon Web Services for $5.5 billion over 15 years. This is seen as one of the first cases of collaboration between mining farms and major companies. What is your opinion on this?

Bit Deer: Iris Energy is a forward-looking Australian company that has been mining in the United States for a long time.

Iris Energy's shift towards AI resembles a flare, as Bitcoin prices reach highs and peers continue to expand mining, it diverts some of its power to invest in its own AI data centers. Consequently, AI companies are actively approaching them.

The real breakthrough comes from the actual capital of Hyperscalers - for example, Microsoft's commitment of about $9.7 billion - based on which the market has clearly seen for the first time that the relationship between mining companies and hyperscalers is not just about “technology integration,” but rather about “the exchange of electricity and time.”

The popularity of AI has amplified the demand for infrastructure, opening up collaborative opportunities.

Dongcha: Why are leading mining companies more likely to be chosen by American AI giants at this stage?

Bit Deer: Because of “available electricity + engineering delivery speed.” The site selection and grid connection of mining companies in the previous cycle have now become the upfront capital for AI data centers. Time is the biggest discount factor, and it directly determines who can go live within the window period, acquire customers, and generate rolling cash flow.

Dongcha: So, is it difficult to select land for AI data centers?

Bit Deer: Overall not large. In the United States and most countries, what is truly scarce is electricity, not land.

The reason is simple: places that can connect to large electricity sources are mostly energy-rich areas (natural gas fields, coal mine belts, near hydroelectric power stations, etc.), which are sparsely populated and have low land prices.

For example, Bitdeer's large data centers in Norway and Bhutan are located away from population centers, where power resources are concentrated and land costs are low. The same applies to the United States; such parks are not located in urban cores but rather in more peripheral areas where it is easier to find land and the prices are lower. The “first principle” of site selection is power and grid connection, and land usually follows power, rather than being the main bottleneck.

Dongcha: AI is now referred to as an upstream business like “steel, electricity, and land”, even resembling another form of real estate. What are your thoughts on this?

Bit Deer: After the release of large models, the power consumption of AI far exceeds most people's expectations.

Initially, everyone thought that “training consumes a lot of power, while inference would be light,” but the opposite is true. After inference becomes more mainstream, it also continues to consume a lot of power for a long time. As ChatGPT and DeepSeek become part of daily life and more terminals connect, the baseline noise of inference keeps rising.

From an engineering perspective, AI is essentially a resource-consuming industry:

• Chip side: The accelerator card runs at almost 100% load during training, which naturally results in high power consumption;
• Data center side: The thermal density is significantly higher than traditional servers, the PUE is clearly elevated, and cooling itself also consumes a large amount of electricity;
• Scale Side: The electricity demand of AI data centers has jumped from 20–30MW of traditional internet data centers to 500MW, or even 1GW level, which was almost unimaginable during the era of traditional internet data centers.

So comparing it to “real estate” is only half right; it indeed requires land, factories, and long cycles (construction cycles often take 3–4 years), but what determines life and death is electricity and heat, whether it can obtain large capacity grid connection on time and achieve N+1/2N redundancy and efficient heat dissipation. In this regard, it is very similar to the strong dependence on steel, electricity, and land.

What are the characteristics of AI data centers?

Insight: What are the characteristics of the data center construction model in the United States?

Bitdeer: Due to power constraints and historical paths in the United States, Hyperscalers often need to get involved personally and collaborate with mining companies to obtain available electricity.

Motion Insight: Is it possible for foreign companies to establish AI data centers in the United States?

Bit Deer: In simple terms, AI data centers are a strong regional business. The real implementation of hundreds of megawatts and thousands of kilowatts is still led by major companies in the United States. We are only discussing AI data centers and not traditional internet data centers.

![]###https://img-cdn.gateio.im/webp-social/moments-f42f3ee7bf68519e6fb58c80cb825c52.webp(

Insight: Will AI Data Centers evolve into tools of geopolitical influence? Will this affect your decision-making?

Bit Deer: I agree with this judgment.

The foundation of AI is data, which is inherently constrained by sovereignty and security. To prevent data leakage and security risks, various regions are tightening related policies: even though the United States allows foreign investment to build data centers, as the amount of data controlled by AI increases, countries are likely to move towards “local deployment, local compliance, and data not leaving the country.”

In simple terms, AI in the US is in the US, in the Middle East is in the Middle East, and in Europe is in Europe; regionalization will be a long-term trend.

) Industry Landscape and Potential

Insight: Besides IREN and Bitdeer, which mining companies have more potential to transition to AI data centers?

Bit Deer: To see who has a chance, first check if they have big electricity, and then see if they can quickly transform the mining site into a GPU room. A type with grid connection + land + substation, and can also manage N+1/2N redundancy, liquid cooling / high density, this type is the easiest to receive AI orders.

Another type of pure custody / light asset, where you do not control the electricity and the park, and transition to an AI data center becomes passive.

In the US, resources like Riot, CleanSpark, Core Scientific, TeraWulf, and Cipher that are reliable for expansion and in their own hands are more likely to be targeted by big companies.

So the conclusion is straightforward: electricity is the ticket, and transformative power is the speed; only when both are in place can you be ahead of the race.

Overall, the key is to see who controls “high-quality, sustainable large-load available electricity.” For example, companies with more self-owned grid-connected resources have greater potential; those that primarily rely on hosting, lacking self-owned energy and parks, do not have an advantage in this round of structural transformation.

What is Bitdeer thinking?

Dongcha: What are Bit Deer’s strategies and paths in “Mining to AI”?

Bit Deer: Wu Jihan's approach has always been to create a full industry chain. Bit Deer holds approximately 3GW of power and park resources, which is our biggest underlying advantage.

When we first entered AI, we did not anticipate that “electricity” would become a core bottleneck, so initially, we adopted a self-built and self-operated approach: we established a partnership with NVIDIA and became an NVIDIA PCSP, deploying a small-scale H100 cluster in Singapore, launching our own AI Cloud, and undertaking training services for external clients. This project has been successfully implemented.

Subsequently, we also established a second data center in Malaysia. As Hyperscalers enter this track and begin collaborating with mining companies, we are simultaneously upgrading the high-load parks to AI data centers: it has been announced that the site in Norway with approximately 180MW will be completely transformed into an AI DC, and the site in Washington State, USA with approximately 13MW will also be converted.

Ultimately, the essence of AI is quite similar to Crypto mining - both are businesses of “electricity + infrastructure”; we have the capability to operate the entire chain from electricity, parks, to computing power, so the transition to AI is relatively smooth.

Dongcha: What are the core differences between Bit Deer and other mining companies like IREN?

Bit Deer: Three points. First, it will not be 100% converted into an AI enterprise; based on calculations, the current profits from Crypto Mining are still superior to those from AI data centers, and mining has a stable cash flow and better returns.

Our second advantage is our international engineering organization capability. The engineering organization and execution ability of the Bit Deer team is unparalleled in the world. For the same AI data center, the usual timeline in the U.S. is two years, but we can typically achieve it in one and a half years. This is achieved through parallel advancement and supply chain collaboration, synchronizing key aspects such as civil engineering, electromechanical, power distribution, and heat dissipation, compressing the usual cycle of about 24 months to approximately 18 months, thus enabling faster usable capacity.

The third company's strategy remains robust: the AI industry is very young, even younger than Crypto, and does not adopt an “all-in” approach, pursuing a more sustainable development pace.

Movement Observation: Where is the current distribution of Bitcoin electricity infrastructure?

Bitdeer: Bitdeer is currently focusing on a global layout of approximately 3 GW of power and related infrastructure, covering the United States, Canada, Norway, Ethiopia, and Bhutan, to support the construction and operation of mining and AI data centers.

Cost and Financing

Observation: I saw a report from Goldman Sachs mentioning that an AI data center could cost 12 billion USD. Is it really that expensive?

Bit Deer: Indeed large, in magnitude it's “dozens of times”. Let me give you an intuitive comparison in “human-friendly numbers”: Bitcoin mining farm (USA): building 1 MW costs about 350,000 to 400,000 USD. However, constructing 1 MW for an AI data center costs about 11 million USD. This is because the investment in an AI data center is a combination of “heavy machinery + heavy standards”: plus grid connection queuing, environmental assessment/energy assessment, and regional compliance, the cycle usually takes 18 to 36 months.

You will find that the essence of an AI data center is not just “buying a few more cards,” but rather connecting a piece of land to create an “electric city” that can consume 500MW–1GW, ensuring proper electricity connections, dissipating heat, managing redundancy adequately, and navigating compliance, all of which are very costly.

Dongcha: Where does the money come from? Is financing needed?

Bitdeer: To be honest, everyone needs financing.

Here are a few common financing strategies in the industry:

1. Project financing / infrastructure loans: Use the park + equipment as collateral, relying on long-term leases or computing power offtake (customers commit to purchasing your computing power for many years) to reassure the bank.

2. Equipment leasing / sale-leaseback: Lease GPUs and some electromechanical equipment to spread out the long-term costs, avoiding the need to expend so much cash all at once.

3. Long-term PPA: First lock in the electricity price and available capacity, only then will the debt side be willing to offer low interest rates.

4. Binding with large companies: Major clients / large companies provide minimum consumption, prepayment, guarantees, or even joint ventures (JV), allowing you to obtain cheaper funding.

In the collaboration between IREN, CoreWeave, and Google/Microsoft, these terms can be seen.

Dongcha: Will Bitdeer also seek financing? Will it soon announce its collaboration with major companies?

Bitdeer: This cannot be publicly discussed in detail at the moment.

Conclusion

Not long after the interview ended, Bit Deer presented its next answer in the capital market.

On November 13, Bitdeer announced that it will raise $400 million by issuing convertible preferred notes and grant initial purchasers the option to subscribe for up to $60 million in additional notes within 13 days, bringing the total fundraising amount to a maximum of $460 million. The new funds will be used for data center expansion, ASIC miner research and development, AI and HPC cloud business expansion, as well as general corporate purposes.

As electricity has become the most scarce upstream resource in the AI industry, where this $460 million will ultimately be invested and how many megawatts of new load will be connected will largely determine the position of Bit Deer in the next round of computing power competition.

For Bitdeer, this money is more like writing the judgments from the interview into the balance sheet: one end connects to the cash flow foundation of the mining sector, and the other end connects to the long-term and substantial business line of AI data centers. It may not immediately reflect in the next quarterly report's revenue and profit, but it will slowly rewrite the power structure of the computing power business in the coming years—who is qualified to sit at the negotiation table and who can only line up on the grid connection list waiting for electricity.

Looking back from the results, the story of this round of AI infrastructure is not complicated: electricity has become the real upstream, time has become the new currency, and the parks and grid connection indicators in the hands of mining companies have turned into “old assets” that others cannot buy even with money.

As the noise around models and applications gradually recedes, the market will likely have to revisit the ledger: it is no longer important whose narrative sounds the loudest; only those companies that can connect every megawatt of electricity and run steadily in a world of power shortages will qualify to remain at the table in the next phase.