The Great Deflation

01The approaching singularity in cost

Economics has long concerned itself with scarcity. From Adam Smith’s division of labour to the marginal revolution, the discipline’s foundational architecture assumes that production requires scarce inputs — labour, capital, land, and coordination — whose allocation determines prices, wages, and welfare. This paper argues that we are approaching a structural break in that assumption.

The convergence of artificial general intelligence and autonomous robotics is compressing the marginal cost of production across virtually every sector. We term this the Great Deflation — not a cyclical downturn in demand of the kind that haunted the 1930s or the 2010s eurozone, but a permanent, supply-side transformation in the cost structure of civilisation itself.

This is not a prediction about a distant future. The precursor dynamics are already visible: the cost of solar energy has fallen 99% since 1976; genome sequencing has fallen from $2.7 billion to under $200 in two decades; the cost of AI inference has dropped by roughly 90% year-on-year since 2020. What we describe is the generalisation and acceleration of these trajectories across the entire economy. The central question is not whether the Great Deflation will occur, but whether our institutions can manage the transition with sufficient speed and wisdom to prevent unnecessary suffering during the adjustment.

02Economic mechanisms of AI-driven deflation

The exponential blindspot

Perhaps the most consequential failure in public discourse is not scepticism about the technology, but a systematic inability to reason about exponential change. Human cognition evolved for linear, local causation; it is poorly equipped to grasp systems that double repeatedly. The majority of commentators who regard post-scarcity economics as distant are not forecasting carefully — they are extrapolating linearly from a system that is growing exponentially. What makes this moment categorically different is the recursive character of AI improvement: AI systems are now substantially involved in their own development, creating a feedback loop with no historical precedent. The slope of the capability curve is not fixed — it is itself increasing.

Supply-side versus demand-side deflation

“Deflation” carries pathological connotations because the most salient episodes — the Great Depression, Japan’s Lost Decade, the eurozone crisis — were demand-side: falling prices reflected collapsing demand, and the feedback loop pulled output and employment down. Supply-side deflation is categorically different. When prices fall because the real cost of production has declined, real living standards rise: consumers purchase more with the same nominal income. The 1870–1896 “Great Deflation” of the nineteenth century — driven by railroads, steamships, and mechanised agriculture — was a symptom of abundance, not privation. The AI-driven Great Deflation will resemble that precedent, but at vastly greater scale and speed.

The agent economy: the race to zero profit

A further mechanism is already operational: autonomous AI agents competing as independent economic actors. As of 2026, tens of millions of agents execute trades, source suppliers, write code, and deliver services on behalf of human principals. They share a characteristic that distinguishes them categorically from human labour: their reservation price — the minimum return at which they accept a transaction — is effectively zero. In markets populated by agents with near-zero operating costs and zero reservation prices, the price floor is effectively zero.

This produces what we term hyper-competitive capitalism — not a pathological distortion of market logic but its fullest expression. The natural tendency of competitive markets has always been deflationary; what historically contained it was the irreducible cost of human labour and capital. Remove those floors, and the deflationary logic runs to its conclusion. The agent economy also democratises productive capital: a citizen can deploy agents to generate income, capturing a share of the productivity dividend independently of traditional employment.

Four reinforcing cost channels

Beyond the agent economy, the deflation operates through four channels. Labour (60–70% of UK GDP) is attacked directly: AGI is a general-purpose cognitive substitute, not a narrow one, so wage pressure extends across the full spectrum of occupations. Capital costs fall because producing capital goods is itself labour-intensive — if AGI designs robots and robots assemble them, each new robot costs dramatically less. Energy and materials are dominated by labour and energy costs rather than physical scarcity; solar costs fall ~20% per doubling of capacity. Coordination costs — search, negotiation, contracting, monitoring — collapse when an AI can instantly identify suppliers, draft contracts, and enforce compliance, shrinking the wedge between production cost and consumer price.

03A ten-point policy framework

The policy challenge is the transition window — the 12–36 month period in which labour displacement precedes price relief. We propose a sequenced framework for the United Kingdom across three horizons.

The centrepiece is the AI Productivity Levy — a tax on the output of automated systems, initially 2% and rising to 5% as automation deepens, hypothecated to a National AI Productivity Fund that finances UBI, public services, and transition investment. The levy is designed to be progressive: systems that displace more labour pay more, creating an incentive for gradual rather than abrupt transition.

04The shape of the post-scarcity world

Healthcare — the NHS as a model. When the marginal cost of a specialist consultation approaches zero, the question is no longer who can afford world-class care, but how quickly institutions can scale it to universal coverage. The NHS — a single integrated delivery network — is exceptionally well-positioned: its architecture, so often criticised for rigidity, becomes a profound advantage when the underlying cost structure collapses. It becomes, in the post-scarcity era, not a rationing mechanism but an abundance distributor.

The end of poverty as a material condition. If a well-designed house costs £10,000, a year’s food £1,000, and healthcare and education are AI-delivered at near-zero marginal cost, then providing a materially comfortable life for every person becomes a trivially affordable proposition. Absolute material poverty ceases to exist as a technical constraint — it becomes purely a matter of distribution and governance.

Inequality and the state. Paradoxically, the elimination of material scarcity may not eliminate inequality. If the gains are captured by AI owners, we could see a small class of trillionaires and a large class dependent on whatever UBI governments provide; if distributed broadly, unprecedented material equality. The outcome depends almost entirely on political choices made in the next decade. The state’s role shifts from regulator of scarcity to steward of the AI commons — and its fiscal basis must shift from income and payroll tax toward productivity levies, data dividends, and resource rents.

Second-order effects. The velocity of money may fall as goods approach free; entrepreneurship may explode as the capital barriers to starting a business collapse; migration will track amenity and community rather than employment; national power will shift toward governance quality and institutional capacity; and authenticity — things made by humans, for human reasons — may become the ultimate luxury.

05Conclusion

The Great Deflation is not a prophecy; it is a projection. The technological trends driving it are visible today, the economic logic is straightforward, and the recursive dynamics are already operating in energy and computation. What remains uncertain is the speed, the sequencing, and — most critically — the institutional response.

Nations that build the infrastructure for the transition — universal basic income, AI-delivered public services, new models of community and meaning — will navigate it successfully. Nations that cling to the institutions and ideologies of scarcity economics will suffer unnecessary hardship. The United Kingdom is well-positioned to lead: the NHS, Universal Credit, a strong AI research ecosystem, and a diplomatic tradition capable of leading multilateral negotiations all provide advantages — but they are contingent, not guaranteed.

The Great Deflation will be the most significant economic transformation since the Industrial Revolution. The question is not whether it will occur, but whether we will manage it wisely. The time for institutional preparation is now.