The Data Centre Boom Makes No Sense - Until You Look at Government Debt

Summary

Miles Harris argues that the data-centre boom is connected to rising government debt through the financial system’s need for productive assets, dependable cash flows, and collateral—not because data centres will directly repay sovereign debt.

• Debt backdrop: US national debt is cited at about $40 trillion, or roughly 120% of GDP. Harris says many other Western countries face similar debt burdens and ongoing interest costs.
• Core thesis: In his view, the constraint on credit creation is not simply the supply of money, but the availability of reliable borrowers, acceptable collateral, and strong cash flows. Banks create much of the new money supply through lending, so a larger base of productive assets could support further private credit.
• Why data centres fit: They combine land and commercial real estate with energy infrastructure, expensive equipment, computing capacity, long-term contracts, and potentially observable operating data. Harris argues these features can make them financeable assets whose cash flows can be owned, lent against, securitized, or funded through corporate debt.
• Broader economic effects: Data-centre construction can attract investment in power grids, utilities, water, transformers, energy storage, cooling, cables, and related real estate. Harris describes this as an ecosystem of assets and contractual cash flows that may support credit creation, economic activity, and taxable income.
• Government and private-sector incentives: The state can encourage buildout without owning most of the infrastructure, while private companies and financial institutions supply capital. The video cites UK AI development zones, US federal sites identified for data-centre and energy projects, and the EU’s InvestAI initiative, which aims to mobilize €200 billion in AI investment through the European Investment Bank.
• Credit quality distinction: Harris characterizes loans financing productive infrastructure as potentially different from mortgage lending that mainly boosts asset prices. Data-centre investment may generate income over decades and improve borrowers’ ability to repay, though the buildout can also affect land prices.

Framework Presented

Harris’s argument can be followed as a sequence:

1. Rising sovereign debt increases the financial system’s need for buyers, refinancing, and collateral.
2. Assess the economy’s broader productive asset base—not government debt in isolation.
3. Identify assets that combine tangible collateral with predictable, contract-backed cash flows.
4. Consider how infrastructure investment may create additional lending, related assets, and taxable activity.
5. Test the thesis against what happens when the investment cycle turns: whether assets remain useful and operating, even if their owners or valuations change.

Risks, Cautions, and What Could Disprove the Thesis

• Harris says low profitability, defaults, and declining data-centre valuations are plausible; investors can lose money.
• He expects some market clearing or a change in ownership to be possible, while arguing that this would not necessarily mean the infrastructure itself fails.
• A stronger challenge to his thesis, he says, would be mass abandonment of data-centre infrastructure because it has little value outside AI.
• He presents this as a structural interpretation of the boom, not as a guarantee of investment returns. No specific investment recommendation or ticker is given.

Assets, Sectors, and Instruments Mentioned

• Assets and infrastructure: Data centres, land, commercial real estate, energy infrastructure, computing equipment, power networks, water systems, transformers, storage, cooling, cables, and related real estate.
• Sectors: AI and technology, utilities, energy, banking, private lending, and infrastructure.
• Financial instruments and concepts: Government debt, corporate debt, bonds, bank loans, collateral, securitization, and private credit.
• Other markets and assets referenced: Silver and critical minerals, as examples in a discussion of strategic competition.
• Tickers: None mentioned.

Presenter/source: Miles Harris; he refers viewers to his Substack, The Collateral Age.