20VC20VC: The Future of Datacentres: What You Need to Know | Why Everyone Gets GPU Depreciation and AI's Energy Costs Wrong | Who Really Makes Money From AI & Why Most Moats Don't Exist with Chase Lochmiller
3 Oct 2026 · 61 min · Episode page ↗The bottleneck for AI infrastructure is not just hardware, but the ability to deploy data centers where energy is cheap and abundant, navigating complex local challenges from power grids to labor supply and public perception.
Summary
This conversation with Chase Lochmiller, CEO of Crusoe, delves into the overlooked complexities and misconceptions surrounding the rapid expansion of AI data centers. It highlights how factors beyond just chip availability—like energy access, vertical integration, and local community relations—are critical for scaling AI, challenging common narratives about its environmental impact and economic effects.
Interesting ideas
AI Data Centers Reduce Local Energy Costs. Contrary to popular belief, AI data centers can actually lower energy prices for local communities. They catalyze investment in new energy generation and amortize grid infrastructure over more megawatts, leading to reduced costs for residents rather than increased bills, which is a key counter-narrative the industry needs to communicate.
Vertical Integration Solves Supply Chain Bottlenecks and Reduces Costs. Being vertically integrated in building AI data centers, from electrical manufacturing to site construction, is crucial. This approach allows companies to circumvent long lead times for specialized components, navigate 'whack-a-mole' supply constraints, and gain a clearer understanding of true costs, ultimately accelerating deployment and optimizing efficiency.
Most 'Moats' in Tech are Ephemeral. In a rapidly evolving technological landscape like AI, traditional competitive 'moats' are often illusions or temporary. The focus should shift from building unassailable barriers to developing the agility to move quickly and adapt efficiently to constant changes in the market and technological capabilities, favoring responsiveness over static protection.
AI Infrastructure is a Portfolio of Monetization Engines. Operating AI infrastructure should be viewed as managing a portfolio of services across the value chain—from data centers and GPUs to managed inference and tokens. This vertical integration, similar to an 'AI super major,' provides inherent hedging against market fluctuations, as margins may shift between different layers, ensuring profitability regardless of where value accrues at a given moment.
Questions worth thinking about
If AI data centers are indeed beneficial for local communities, why is the misinformation so pervasive, and what specific strategies could the industry use to shift public perception?
Given the 'ephemeral moat' argument, how can investors and founders differentiate between genuine, long-term competitive advantages and temporary market leads in the AI space?
If power is a key constraint and global energy prices vary significantly, how will this geographic disparity influence which regions become the primary hubs for future AI development?
Written by AI from the episode's full transcript. It can get things wrong: if something here looks off, tell us at hello@podpings.com.