Empirical Case for Abundance

[CONVICTION]

Three structural transformations are underway simultaneously, and the data on each are no longer ambiguous. Technology is producing abundance along predictable exponential curves. GDP has become a measure of extraction rather than creation. And capital is responding rationally to the coming transition -- repricing atoms above bits, physical infrastructure above software margins, energy above finance. What follows is the evidence chain.

GDP Has Been Colonized by Extraction

[EVIDENCE]

The most striking transformation in the US economy over four decades is not the rise of technology. It is the takeover by the FIRE sector.

Finance, insurance, real estate, and rental/leasing combined grew from 15.2% of GDP in 1979 to 21.7% in 2025 -- the single largest sector of the American economy. Finance and insurance alone doubled from 4.9% (1980) to 8.0% (2025). Financial sector profits, which represented 10% of all corporate profits in 1947, captured 50% by 2010. Thomas Philippon of NYU documented that total financial intermediation costs rose from 5% to ~9% of GDP between 1980 and 2010 -- "$280 billion per year in misallocated resources" -- despite information technology that should have lowered them.

Healthcare underwent a parallel expansion: from ~8-9% of GDP (1980) to 16.7% (2023), driven by administrative complexity and pricing power rather than improved outcomes. The US spends 16.5% of GDP on healthcare versus the OECD average of 9.2%, without superior health metrics.

Manufacturing moved in the opposite direction: from ~22% of GDP (1980) to 9.4% (Q2 2025). Agriculture fell from ~2-3% to under 1%.

The pattern is unambiguous: sectors that extract rents gained ~15+ percentage points of GDP share while sectors that produce tangible goods lost a comparable amount. Greenwood and Scharfstein (Journal of Economic Perspectives, 2013) attributed finance's growth to mortgage securitization and asset management fees -- neither representing new productive capacity.

The Invisible Economy

[EVIDENCE]

GDP systematically fails to measure the value that matters most. Conservative estimate of the "invisible economy": $16-22 trillion per year in human activity alone, or 15-21% of global GDP. Including natural capital, it exceeds measured GDP entirely.

• Ecosystem services: $125-145 trillion/year (Costanza et al., Global Environmental Change, 2014) -- 1.4-1.7x global GDP. Climate regulation, water purification, pollination, soil formation. Destruction of these services increases GDP through the extractive activities that cause it.
• Unpaid care and domestic work: $11 trillion/year (ILO), representing 16.4 billion hours of daily labor, 76.2% performed by women. Oxfam's 2020 estimate: women's unpaid work alone at $10.8 trillion -- "three times the size of the global tech industry."
• Open source software: $8.8 trillion demand-side replacement value (Hoffmann, Nagle, and Zhou, Harvard Business School, 2024). Firms would need to spend 3.5x more on software without OSS. Excludes operating systems like Linux.
• Digital consumer surplus: US consumers value free search engines at $17,530/year per user, email at $8,414, digital maps at $3,648 (Brynjolfsson et al., "GDP-B" research). Nordhaus estimated firms capture only 2.2% of total surplus from technological innovations -- 97.8% flows to consumers unmeasured.
• Volunteer labor: $1.3-1.5 trillion/year globally (Johns Hopkins Center for Civil Society Studies).

The BEA's Household Production Satellite Account found that including unpaid domestic work alone would expand US GDP by 25%. Lossy compression is not metaphor. It is measurable information loss at civilizational scale.

AI Reverses the Automation Order

[EVIDENCE]

Previous automation waves -- mechanization, electrification, computerization -- displaced physical and routine labor. AI does the opposite. The empirical evidence for this "reverse automation order" comes from multiple independent sources.

Exposure studies. Eloundou et al. (OpenAI/University of Pennsylvania, Science, 2024): ~80% of the US workforce could have at least 10% of tasks affected by LLMs, with higher-income jobs facing greater exposure. IMF (January 2024): ~60% of jobs in advanced economies exposed -- "AI challenges the belief that technology affects mainly middle and low-skill jobs." Goldman Sachs: 300 million full-time jobs globally face automation exposure; legal, administrative, and engineering roles most at risk.

Observed displacement. The St. Louis Federal Reserve (Ozkan and Sullivan, August 2025) found a correlation coefficient of 0.57 between AI adoption intensity and unemployment increases (2022-2025). Computer and mathematical occupations (~80% AI exposure) showed "some of the steepest unemployment rises." Computer programmer employment fell 27.5% over two years (2023-2025), from ~166,000 to ~121,200. Indeed software engineer postings dropped 35% from January 2020 levels. ADP Research confirmed fewer US software developers in January 2024 than six years prior. Microsoft's CEO stated 30% of company code is now AI-written. Over 50,000 layoffs in 2025 explicitly attributed to AI. Tech unemployment hit 5.7% in February 2025 -- above the national average. CS graduates' unemployment exceeded philosophy graduates' for the first time.

Physical labor resilience. Construction added 190,000 jobs in 2024 and 33,000 in January 2026. Healthcare support is projected as the fastest-growing category through 2034. WEF projects farmworkers and delivery drivers among the largest absolute job-growth categories globally by 2030.

The pattern is consistent: AI deflates knowledge work while physical work remains resilient. This is the reverse of every previous automation wave and accelerates the deflationary-cascade into the service sectors that now dominate GDP.

Wealth Has Decoupled from Value

[EVIDENCE]

McKinsey's 2025 "Out of Balance" report provides the definitive accounting. The global balance sheet quadrupled from 2000 to 2024, reaching $1.7 quadrillion in total assets. Households gained $400 trillion in wealth -- but only ~$100 trillion (25%) reflected cumulative net investment. A full 36%, or $146 trillion, was paper wealth: asset price appreciation with no corresponding increase in productive capacity. For every $1 of net investment, $3.50 in new household wealth was created. For every $1 of net investment, $4 in financial liabilities were generated.

The stock market: Buffett Indicator reached ~220% in early 2026, nearly 3x the historical average of ~75%. Shiller CAPE at ~40, versus historical median of 16. Real estate: Knoll, Schularick, and Steger found up to 80% of house price increases (1950-2012) attributable to land price appreciation alone -- not construction costs, which flat-lined. US median home prices rose ~207% from 2000-2024; per-capita income rose only ~155%.

Labor market extraction: CEO compensation at top 350 US firms rose 1,094% from 1978 to 2024. Typical worker compensation rose 26%. Productivity grew 80.5%. Since 1979, productivity has grown 3.5x as much as pay for the typical worker. The EPI's "jaws chart" documents the divergence precisely. Piketty's r > g is empirically confirmed: a 2025 Cambridge Journal of Economics study found a 1 percentage point increase in the r - g gap is associated with a 3.7% increase in the top 1% wealth share. The global top 1% controls 37% of all wealth and captured 41% of all new wealth generated between 2000 and 2024.

The derivatives market: ~$699 trillion in notional outstanding OTC derivatives (year-end 2024), or 6.4x global GDP. Global debt reached $318 trillion (328% of GDP). Financial claims on future value that dwarf the real economy's capacity to honor them -- the textbook definition of fictitious capital expanding faster than productive capacity. The empire-collapse-pattern has seen this before.

Capital Is Migrating to Atoms

[EVIDENCE]

Corporate capital expenditure tells the decisive story. Hyperscaler capex (Amazon, Microsoft, Google, Meta) grew from ~$24 billion (2015) to $211 billion (2024), with projections of $315-443 billion for 2025 and $602 billion for 2026. 75% of this spending flows to physical AI infrastructure -- data centers, GPUs, power systems. NVIDIA's data center revenue surged from $2.9 billion (FY2019) to $115 billion (FY2025), a 40x increase in six years. Capital intensity reached 45-57% of revenue for major hyperscalers -- historically unprecedented.

Global clean energy investment crossed $2.2 trillion in 2025, doubling from ~$1.0 trillion in 2015, with a 2:1 ratio over fossil fuel investment. China invested $800 billion (4.5% of GDP) -- more than the US, EU, and UK combined. Private equity infrastructure AUM quadrupled to $1.3 trillion over the past decade. Data center PE deals jumped from $11 billion (2020) to $50 billion (2024). Infrastructure fundraising surpassed real estate fundraising in 2024 for the first time in history.

Traditional software is being repriced. AI's share of US VC exploded from 16% (2021) to 71% (Q1 2025). Public SaaS valuation multiples collapsed from 18-19x EV/Revenue (2021 peak) to 5.1x (December 2025) -- ~70% compression. Private software M&A multiples fell from 6.7x to 2.9x. The IEA noted AI VC reached 3x the level of energy VC in 2024.

Goldman Sachs projects $1.15 trillion in cumulative hyperscaler capex for 2025-2027 alone. The era of asset-light software generating outsized returns is narrowing to AI or ending altogether. The market is pricing in what economic theory has not yet absorbed: the future is physical.

National Exposure Is Asymmetric

[EVIDENCE]

The deflationary-cascade creates radically different vulnerability across economies. The paradox: the most AI-capable nations are the most structurally exposed.

Most exposed (service-heavy, financialized): US (77.8% services, $109 billion private AI investment in 2024, 73% household debt-to-GDP), UK (73% services), Australia (66% services + 110% household debt-to-GDP).

Best industrial buffers: China (36.5% industry, $800 billion clean energy investment, 46% household debt-to-GDP), South Korea (32% industry, 5.3% R&D-to-GDP), Germany (27% industry, universal healthcare, 51% household debt-to-GDP).

Highest financialization risk: Switzerland (126% household debt-to-GDP), Australia (110%), Canada (102%). These economies face severe balance-sheet crises if AI-driven deflation erodes asset values.

Weakest AI readiness among large economies: India (0.6% R&D/GDP, ranked 46th in government AI readiness), Brazil (~30th+ on most indices). India's large IT services sector is among the first exposed to AI deflation -- a significant vulnerability.

China surpassed the US in absolute R&D spending for the first time in 2024 ($785.9 billion vs. $781.8 billion, PPP). The US maintains higher R&D intensity (3.4-3.6% vs. 2.6% of GDP). The US is simultaneously the primary creator and primary target of AI disruption.

What the Data Mandate

[CONVICTION]

The empirical record points to a structural transition that conventional economics cannot adequately explain. Technology produces abundance along predictable curves, but GDP registers this as stagnation. GDP has become a measure of extraction -- the sectors gaining share are predominantly rent-seeking, while the invisible economy likely exceeds $150 trillion annually. Capital responds rationally: hundreds of billions flowing to physical infrastructure, SaaS valuations collapsing, the market pricing in deflation of knowledge work and the premium on atoms over bits.

The wealth-to-value divergence provides the clearest signal that the current system is nearing a structural limit. When 36% of household wealth gains are paper appreciation, when financial claims exceed the real economy by 6-7x, when CEO compensation rises 42x faster than worker pay despite only 80% productivity gains, the system extracts value rather than creating it.

The countries best positioned -- China, Germany, South Korea -- maintain substantial industrial bases, aggressive energy investment, and lower financialization. The most capable AI nations, particularly the United States, face a paradox: building the tools of abundance while maintaining economic structures optimized for scarcity and extraction. The deflationary-cascade guarantees this contradiction cannot hold. The empire-collapse-pattern shows what happens when it breaks. The question is whether the break produces another scarcity regime or the first transition through scarcity into abundance coordination.

Related

• deflationary-cascade -- the simultaneous cost collapse this evidence documents
• lossy-compression -- why GDP cannot measure what matters
• money-as-scarcity-tool -- why the coordination mechanism breaks under abundance
• empire-collapse-pattern -- historical precedent for scarcity tools failing
• abundance-distribution-problem -- the governance challenge that follows production abundance
• economics -- domain overview