Key takeaways
- The demand for energy from the AI sector is immediate and high, reflecting the urgency in energy provisioning.
- Hyperscalers prioritize bringing capacity online quickly, often over cost considerations, to maintain competitive advantage.
- International markets offer more economically viable opportunities for bitcoin mining compared to the US
- Excess energy in countries like Saudi Arabia and France is underutilized due to inadequate grid connections.
- The cost of bitcoin mining capacity is at an all-time low due to oversupply and reduced demand.
- Public miners are shifting focus towards high-performance computing (HPC) and AI, impacting the demand for new mining equipment.
- Modularizing data centers can cut infrastructure costs and boost efficiency in AI inference.
- Revenue in the AI sector is predominantly driven by inference processes.
- Data center providers are increasingly becoming energy developers to meet future operational needs.
- AI technologies are eroding the value traditionally held by software systems by automating insights.
- The strategic focus of hyperscalers on capacity wars underscores the importance of operational readiness.
- The shift of public miners towards HPC and AI reflects a broader industry trend impacting bitcoin mining dynamics.
- The economic viability of international markets for bitcoin mining is influenced by lower AI demand and energy availability.
- Modular data centers offer a solution to high infrastructure costs, particularly in AI applications.
- The erosion of software value by AI highlights the transformative impact of technology on traditional business models.
Guest intro
Fred Thiel is the Chairman and Chief Executive Officer of Marathon Digital Holdings (NASDAQ: MARA), one of the largest publicly traded bitcoin mining companies. He joined the company in 2018 as a director and became CEO in 2021, overseeing its growth from a market cap under $30 million to over $5 billion. Under his leadership, Marathon has expanded bitcoin mining operations across four continents while emphasizing energy and infrastructure strategies.
The urgent energy demands of AI
-
The big demand right now for energy is from obviously the hpc ai sector and they need energy that is on now and not on in 2029
— Fred Thiel
- Immediate energy needs are driven by the high-performance computing (HPC) requirements of AI.
- Hyperscalers are in a capacity war, focusing on who can run more GPUs.
-
It’s all about a capacity war who has gpu’s running… the more stuff you have running the smarter your model is
— Fred Thiel
- The timing of bringing capacity online is prioritized over cost by hyperscalers.
- The AI sector’s energy demands are shaping market dynamics and strategic decisions.
- The focus on operational readiness is critical for maintaining competitive advantage.
- Energy provisioning for AI is a key factor influencing infrastructure development.
- The urgency in energy demand reflects the rapid growth and importance of AI technologies.
International opportunities in bitcoin mining
-
A lot of our growth for bitcoin mining is actually targeted internationally… internationally there is much less demand for ai in hyper scalar capacity.
— Fred Thiel
- International markets are more economically viable for bitcoin mining than the US
- Lower demand for AI and hyperscale capacity abroad creates opportunities for mining.
- Excess energy in countries like Saudi Arabia and France is underutilized.
-
You have all this renewable energy that’s available but isn’t connected to a grid to be sold… 66 of the energy is nuclear… and the capacity at the nuclear power plants is running below 70%.
— Fred Thiel
- The lack of grid connections limits the effective use of available energy resources.
- Energy market dynamics vary significantly between the US and international regions.
- Strategic shifts in mining operations are driven by regional energy availability.
- International growth in mining is influenced by economic and infrastructural factors.
The current state of bitcoin mining capacity
-
The cost of a terahash of bitcoin mining capacity is at its lowest point most probably ever… it’s a combination of oversupply and no demand.
— Fred Thiel
- Bitcoin mining capacity costs are at an all-time low due to oversupply.
- Reduced demand for mining rigs is linked to strategic shifts by public miners.
-
40% of miners in the us are the public miners… they’re not buying any more miners.
— Fred Thiel
- Public miners are increasingly focusing on high-performance computing (HPC) and AI.
- The shift in focus impacts future demand and investment in mining infrastructure.
- Current market dynamics are shaped by economic factors affecting pricing.
- The oversupply of mining capacity reflects broader industry trends.
- Strategic changes by public miners highlight the evolving landscape of bitcoin mining.
Modularizing data centers for AI efficiency
-
Take the infrastructure down to a million dollars a megawatt and solve that problem through modularizing a data center and in the world of inference you don’t need to have all these huge clusters that are interconnected.
— Fred Thiel
- Modular data centers can significantly reduce infrastructure costs.
- Efficiency improvements in AI inference are achieved through modularization.
- The cost of data centers is a critical factor in AI infrastructure development.
- Modularization offers a solution to the high costs associated with traditional data centers.
- Revenue generation in AI is primarily driven by inference processes.
-
Revenue generation in the ai world is only done through inference that’s what you pay for.
— Fred Thiel
- The focus on inference highlights its importance in the AI business model.
- Modular data centers align with the strategic needs of AI applications.
Data centers as energy developers
-
It’s all about time to time to energization of the data center… as you get beyond ’29 there you can look at i’m going to go build my own power generation because that’s going to take a couple of years to three years just to get permitting etcetera.
— Fred Thiel
- Data center providers are increasingly becoming energy developers.
- The shift towards energy generation addresses operational capacity needs.
- Building energy infrastructure involves significant timelines and challenges.
- The strategic focus on energy development reflects broader industry trends.
- Operational challenges in energy provisioning influence data center strategies.
- The relationship between data centers and energy production is evolving.
- Energy development is a critical component of future data center operations.
- The shift highlights the importance of energy infrastructure in data center planning.
AI’s impact on traditional software value
-
All the value accretes at the end… the moat that sap has is being eroded now by ai because while you won’t necessarily rip sap out you will through agentic technologies essentially generate the insights that the humans who would be operating sap would have.
— Fred Thiel
- AI technologies are eroding the value traditionally held by software systems.
- The automation of insights by AI impacts traditional software business models.
- The competitive landscape in the tech industry is shifting due to AI advancements.
- AI’s impact on software highlights the transformative nature of technology.
- The erosion of software value underscores the need for adaptation in the industry.
- AI technologies offer new ways to generate insights traditionally provided by humans.
- The strategic implications of AI on software systems are significant.
- The shift reflects broader industry trends towards automation and efficiency.
