Part VII · The Distribution Question
VII.A — The Shape of the Transition
8 min read · 1,470 words
Every technological transition redistributes. The steam engine enriched factory owners and impoverished handloom weavers. The automobile created Detroit and hollowed out the livery trade. The personal computer made knowledge workers more productive and eliminated typing pools. The distribution of gains and losses was never uniform, never immediate, and never fully anticipated.
The Factor Prime transition will follow this pattern. The question is what shape the redistribution takes. The framework developed in prior sections suggests a specific answer—not a prediction of which firms will win or which workers will suffer, but an ordering principle that determines the sequence of change.
The expected path has four components, each following from the production function.
First, concentration at the actuation layer. Cognitive capability commoditizes while actuation bottlenecks persist. Returns flow to whoever controls physical throughput, trusted interfaces, verification infrastructure, and liability capacity. The pattern has precedent: the electricity transition enriched owners of transmission and distribution networks, not dynamo manufacturers; the railroad transition enriched owners of rights-of-way and terminal facilities, not locomotive builders; the internet transition enriched platform owners, not content creators or connectivity providers. In each case, the scarce layer captured value from the abundant layers.
The Factor Prime transition points toward similar concentration. Utilities with generation and transmission assets. Industrial conglomerates with manufacturing capacity. Financial institutions with regulatory licenses and balance-sheet capacity. The low-glamour segments of the technology sector—datacenter operators, chip foundries, cloud providers with physical infrastructure rather than pure software margins. The pattern is visible in current capital allocation: hyperscalers are building power plants; frontier labs are securing long-term power purchase agreements; the constraint on model deployment is increasingly megawatts rather than parameters.
Second, displacement ordered by verification cost. The V/C ratio determines which tasks cross the substitution threshold and when. High value-to-verification-cost tasks automate first because selection can operate rapidly: the output is observable, feedback is immediate, errors are caught and corrected, deployment proceeds. Low V/C tasks automate slowly because selection is constrained: the output is difficult to evaluate, consequences are delayed, errors compound before detection, deployment stalls regardless of capability.
The ordering is not a ranking of task difficulty or model capability. It is a ranking of how quickly the selection gradient can operate. Code generation automates before medical diagnosis not because models are better at coding than at medicine—they may not be—but because code compiles or it does not, while diagnostic accuracy requires months of patient outcomes to assess. The verification infrastructure, not the cognitive capability, determines the sequence.
Third, coordination scaling gated by the term structure. If the Bitcoin yield curve emerges—if a benchmark rate for multi-period commitments consolidates—then agent-mediated markets can support forward contracts, escrow arrangements, and complex coordination across time. The O(N²) problem collapses to O(N). Liquidity concentrates around the reference rate. The infrastructure for agentic commerce scales.
If the term structure does not emerge, coordination remains bilateral. Multi-period contracts require bespoke negotiation. Agent-mediated markets remain shallow. The transition proceeds more slowly because the coordination infrastructure does not exist. The term structure is not a technical detail; it is a gating factor. The difference between a world where autonomous agents can make 90-day commitments at standardized rates and a world where every commitment requires bilateral negotiation is the difference between a functioning market and a set of isolated transactions.
Fourth, adjustment windows determined by bottleneck durability. The speed of the transition depends on the slower of two rates: the rate at which AI capability improves and the rate at which actuation infrastructure expands to support deployment. Part II established that infrastructure is currently the binding constraint. Interconnection queues, transformer lead times, fab capacity, and permitting timelines all expand more slowly than model capabilities improve.
This creates adjustment windows. Workers and regions positioned in low-V/C domains—where verification is expensive and deployment is slow—have time to adapt. Workers and regions positioned in high-V/C domains face pressure that arrives faster than institutional adjustment can accommodate. The window is not infinite; it depends on how long the actuation bottlenecks persist. If bottlenecks loosen faster than expected—through regulatory reform, technological breakthrough, or institutional adaptation—the windows compress.
The monitoring signals indicate whether the transition is proceeding as the framework predicts.
Enterprise switching behavior. If customers readily switch cognitive providers when superior alternatives appear, commoditization proceeds. If customers remain despite superior alternatives, something else generates lock-in—data, integration, workflow, or trust—and the margin pool may stay upstream longer than the framework predicts. Enterprise renewal rates and price-performance spreads between frontier and open-weight alternatives provide signal.
Permitting and interconnection timelines. ERCOT’s interconnection queue has surged past 100 GW, with projects spending years in study phases. If these timelines compress—if permitting reform accelerates construction—the physical bottleneck loosens. The spread between actuation margins and cognitive margins narrows. Concentration at the actuation layer may be less pronounced than expected.
Liability rulings and insurance development. Early disputes involving autonomous agents will establish precedents that either enable or constrain deployment. If courts hold principals strictly liable for agent actions, deployment slows in high-stakes domains. If liability frameworks accommodate agent participation with manageable exposure, deployment accelerates. Insurance product development—the emergence of standardized coverage for autonomous operations at reasonable pricing—indicates whether the institutional infrastructure is keeping pace.
Term structure emergence. The emergence of observable market prices for Bitcoin-denominated obligations at multiple tenors, the publication of standardized rate feeds, the development of oracle infrastructure connecting on-chain contracts to off-chain attestation—each indicates whether the coordination infrastructure is forming. If liquidity concentrates around a benchmark, the framework’s predictions about coordination scaling are on track. If liquidity remains fragmented, the term structure may not emerge, and agent-mediated markets may remain shallow.
Current readings favor the thesis: interconnection queues are lengthening, liability doctrine has not adapted, insurance products remain nascent. But the assessment requires continuous updating. The signals that would indicate the framework is wrong are as important as the signals that confirm it.
Three contingencies could alter the expected path.
Contingency one: cognitive moats persist. If frontier model companies achieve escape velocity through workflow lock-in before commoditization completes, the margin pool stays upstream. The mechanism: usage generates evaluation data; evaluation data improves reliability; reliability enables compliance deployment; compliance deployment generates more usage. If this loop compounds faster than capability diffuses, integrated platform companies capture more value than the actuation layer. The bet on infrastructure underperforms.
The tell: whether frontier labs pursue vertical integration into actuation. If OpenAI and Anthropic build their own liability coverage, regulatory infrastructure, and verification systems, they concede that the model layer alone does not capture sufficient value. That concession would validate the framework even as it redirects the margin pool.
Contingency two: institutional adaptation accelerates. If permitting reform, liability frameworks for autonomous agents, and standardized insurance products emerge faster than assumed, the bottlenecks clear. Scarcity rents dissipate. Returns distribute more broadly across the value chain. The actuation layer remains important but commands smaller premiums.
The mechanism: institutions adapt through crisis. The turning point in the Perez framework often involves a financial crisis that resets valuations and forces institutional adaptation. If the Factor Prime transition triggers a crisis that accelerates regulatory and institutional response, the bottlenecks may loosen faster than the gradual-adjustment framework assumes.
Contingency three: capability advances discontinuously. If a breakthrough creates agents that bypass current actuation constraints—that can establish identity, settle transactions, verify physical states, and bear liability without the infrastructure the framework assumes they require—then positioning strategies built on gradual adjustment become obsolete. This is not an error within the framework; it is a failure of the framework itself. Concentration occurs, but at different points than the framework predicts.
The signal: deployment lag compression. GPT-4 became available via API in March 2023; broad enterprise deployment with compliance certification required roughly a year. If deployment lag for subsequent frontier models compresses below six months, or if agentic deployment expands to high-stakes domains faster than institutional constraints would predict, the transition is accelerating beyond the speed at which capital and labor can reposition.
The first two contingencies are recoverable errors for liquid capital. Positions can be adjusted; lessons transfer forward. The third is regime change. The framework itself requires revision, and positions built on it require unwinding.
The transition is underway. Energy structured through computation and disciplined by selection is becoming a primary input to economic production. The uncertainty is not whether this is happening but how fast, and who will be positioned to capture the gains.
The following sections examine who bears the costs of this transition, and what institutions will be required to navigate it.