For most of modern economic history, prosperity spread because expansion required people. When companies grew, they built plants, opened regional offices, hired layers of managers and trained thousands of workers. Corporate ambition translated into mass employment, and mass employment translated into rising household income. That chain reaction defined the postwar growth model.

Today, that transmission mechanism is breaking down. The most powerful firms no longer need vast workforces to generate extraordinary value. A small team armed with scalable software, proprietary data and advanced computing infrastructure can produce output that once required entire industrial complexes. Market capitalization can double without a surge in hiring. Profits can soar even as payrolls remain flat. Economic growth is no longer tightly coupled to job creation; it is increasingly coupled to capital deepening — the increase in the capital-labor ratio.

This shift has consequences that reach far beyond corporate strategy. When value creation depends less on labor and more on intangible assets, the distribution of income changes. Gains accrue to shareholders, founders and holders of intellectual property. Wages, by contrast, rise more slowly and are often detached from the pace of productivity growth. The result is an economy capable of generating immense wealth without generating commensurate employment security. That is the defining structural transformation of our time: not simply technological change, but the weakening of the historical link between growth and broad-based labor participation.

From industrial scale to algorithmic scale

In the mid-1980s, corporate dominance required organizational breadth. IBM was emblematic of an industrial capitalism in which scale meant payroll. Its competitive advantage depended on vertically integrated production, large research teams, in-house manufacturing and long-term employment relationships. Growth translated into jobs; profits and wages expanded together. Corporate size and labor intensity were tightly correlated.

Four decades later, Nvidia illustrates a structurally different model. Its market capitalization and profitability, even when adjusted for inflation, vastly exceed IBM’s peak levels. Yet its workforce is a fraction of IBM’s. The divergence is not merely technological; it reflects a transformation in how value is produced and distributed. Modern firms scale through intellectual property, software ecosystems and platform effects rather than through proportional labor expansion. Once a chip architecture or software framework is designed, incremental output requires minimal additional employment. Revenue growth decouples from payroll growth.

This shift corresponds to a long-term decline in labor’s share of national income. Since 1980, the proportion of economic output accruing to wages and benefits has trended downward, while the share flowing to profits has risen. Multiple forces contributed: the erosion of unions, global labor competition, outsourcing and the replacement of durable industrial capital with rapidly depreciating digital capital. Expenditure shifted from factories and machinery to software, algorithms and intellectual property — assets that scale without parallel increases in employment.

Automation’s first wave targeted routine manual labor. Manufacturing productivity surged, but factory employment declined. Workers displaced from assembly lines often transitioned into services or administrative roles, albeit frequently at lower pay. The macroeconomic result was higher aggregate productivity alongside greater wage dispersion. This adjustment unfolded gradually over decades, allowing labor markets to absorb shocks incrementally.

The post-pandemic economy revealed how entrenched the capital tilt has become. Although tight labor markets temporarily boosted nominal wages, inflation diluted much of the real gain. Meanwhile, corporate profit margins reached historic highs. Equity valuations expanded not only because earnings rose but because investors priced in the durability of scalable, capital-intensive business models. When stock wealth approaches multiples of disposable income, asset performance becomes a primary driver of consumption, particularly among higher-income households. The macroeconomy becomes increasingly sensitive to capital market dynamics rather than solely to wage growth.

This structural evolution has produced a bifurcated experience. Aggregate indicators signal resilience — strong GDP, high equity valuations — yet median households perceive fragility. The explanation lies in distribution. Capital gains are concentrated, while wage growth is diffuse and comparatively modest. The economic system has become more efficient at generating returns on capital than at translating productivity gains into broad-based income growth.

Artificial intelligence as general cognitive substitution

Artificial intelligence represents not a continuation of prior automation, but a qualitative expansion. Earlier technological waves automated specific tasks within defined sectors. AI operates across domains, targeting cognitive processes that underpin professional work. Language models can draft contracts, summarize case law, construct financial models, analyze medical scans and write software. These are not peripheral functions; they are core components of white-collar employment.

Executives at leading AI firms have acknowledged the speed and breadth of this advance. Dario Amodei of Anthropic has argued that AI is progressing faster than expected and may soon replicate a wide spectrum of human cognitive abilities. Unlike factory robots, which displaced discrete physical tasks, AI systems increasingly substitute for analytical and communicative labor across multiple sectors simultaneously.

The economic implication is a compression of labor demand in high-skill occupations once considered insulated from automation. Junior legal associates, financial analysts, compliance officers and research assistants perform tasks that AI can now replicate or augment at marginal cost. Firms that integrate AI effectively may require fewer entry-level employees to generate equivalent output. Revenue per employee rises, but aggregate employment growth slows.

Consider a concrete example. Several major law firms have begun deploying AI tools to conduct document review and draft preliminary briefs. Tasks once assigned to teams of junior associates — often billing hundreds of hours — can now be completed in a fraction of the time. Hiring pipelines at the entry level are already narrowing. Revenue per partner rises, costs decline but the profession’s absorption capacity for new graduates contracts.

This dynamic extends beyond law. Investment banks use AI to construct pitch materials and valuation models. Consulting firms deploy internal language models to automate research synthesis. Customer service operations integrate AI agents capable of handling complex interactions without human escalation. The result is not mass unemployment overnight, but a compression of demand for routine cognitive labor.

The distinctive feature of AI is that it narrows the traditional refuge of retraining. When manufacturing was automated in the late 20th century, displaced workers could shift toward clerical and managerial roles. Today, retraining into screen-based occupations offers less insulation if AI can perform similar tasks at marginal cost.

At the same time, AI development itself is highly capital-intensive. Training frontier models requires advanced semiconductors, vast data centers and enormous energy capacity. Only firms with substantial financial and technological resources can operate at the cutting edge. This reinforces concentration. If productivity gains accrue primarily to shareholders and intellectual property holders, labor’s share of income may decline further.

Recent military applications further illustrate this structural shift. Artificial intelligence is increasingly deployed in intelligence analysis, target selection, logistics coordination and operational planning in modern conflicts. In contemporary warfare, AI enhances the capacity to process vast streams of data, accelerating decision cycles and improving precision. This evolution reflects the broader economic logic of algorithmic scale: Complex outcomes once requiring large human organizations can now be achieved through capital-intensive computational systems. The strategic implications extend beyond the battlefield. As military effectiveness becomes tied to access to advanced computing infrastructure and proprietary algorithms, technological concentration reinforces both geopolitical asymmetries and the declining centrality of labor in high-stakes institutional decision-making.

Yet AI also creates tension within labor markets. Highly skilled engineers and AI specialists often receive equity-based compensation, aligning their income with capital performance. They are not purely wage earners; they are hybrid participants in capital gains. Meanwhile, mid-level professionals without equity exposure face substitution pressure without participation in upside. The labor market bifurcates between those augmented by AI and those displaced by it.

History suggests the pattern could resemble manufacturing automation: productivity rises, consumer costs fall but wage growth becomes uneven. The difference is scope. Manufacturing affected a segment of the workforce. AI touches the cognitive foundation of modern economies.

Macroeconomic and policy consequences

If AI accelerates the capital-deepening trend, the macroeconomic framework itself will evolve. A lower labor share implies that aggregate demand depends more heavily on asset values. Wealth effects become central. When equity markets rise, consumption expands among asset-owning households. When markets contract, spending retrenches. Economic volatility increasingly mirrors financial volatility.

In such a regime, monetary policy faces a dual sensitivity. Interest rate changes influence not only borrowing costs but also equity valuations. Policymakers must weigh labor market conditions against asset-price stability. A tightening cycle that depresses markets may suppress consumption disproportionately relative to its impact on wages. Conversely, accommodative policy may inflate asset bubbles, reinforcing inequality.

Distributional tensions are likely to intensify. If profit shares continue to climb while wage growth moderates, demands for redistribution will increase. Policy responses could include capital gains taxation reforms, expanded social insurance, public investment in AI infrastructure or new frameworks for worker ownership. Alternatively, governments may prioritize national competitiveness, subsidizing domestic AI champions and reinforcing capital concentration.

The trajectory will depend partly on productivity diffusion. If AI tools become widely accessible and enable small firms to compete effectively, competitive pressures could compress margins over time, moderating capital’s dominance. Conversely, if network effects and data advantages entrench a handful of firms, profit concentration may persist. The balance between diffusion and concentration will shape labor outcomes.

Several plausible scenarios emerge. In a balanced diffusion scenario, AI boosts productivity broadly, reduces service costs and creates complementary occupations, stabilizing labor’s share near current levels. In a concentration scenario, AI-driven firms maintain high margins, employment growth slows and labor’s share falls below half of national income. In a policy-mediated scenario, governments intervene to redistribute gains or foster broader ownership of AI infrastructure, partially offsetting capital’s ascendancy.

The most probable near-term outcome is continued capital deepening. Equity markets have already priced in sustained profitability for leading AI firms. Labor market adjustments, by contrast, occur gradually. Early evidence of professional layoffs alongside record corporate earnings suggests that the distributional shift is underway.

The central economic challenge is not productivity itself. AI promises substantial efficiency gains. The challenge is institutional adaptation. Education systems must prepare workers for hybrid human-machine roles. Regulatory frameworks must address concentration without stifling innovation. Fiscal policy must reconcile revenue needs with incentives for investment.

The transition from industrial scale to algorithmic scale marks a structural reordering of capitalism. In the industrial era, growth required mobilizing large labor forces. In the AI era, growth increasingly depends on capital-intensive intelligence systems that scale with limited incremental labor. Unless mechanisms emerge to align productivity gains with broad income growth, the divergence between capital and labor will widen.

Modern capitalism is entering a phase in which ownership structure may matter more than employment structure. If access to capital remains concentrated, inequality will widen structurally. If ownership broadens — through retirement systems, public investment vehicles or employee equity participation — the gains of intelligence could be shared more widely.

The transition from industrial scale to algorithmic scale is not simply technological. It is a redefinition of how prosperity circulates. The coming decade will determine whether AI becomes an engine of inclusive productivity or a mechanism that further decouples growth from labor participation. That choice will shape not only economic performance, but the political legitimacy of the system itself.

[Kaitlyn Diana edited this piece.]

The views expressed in this article are the author’s own and do not necessarily reflect Fair Observer’s editorial policy.