AI intelligence becomes abundant & broadly diffused
By 2026-2028, AI capabilities cross a threshold where they transform the entire economy—not just software—creating a positive feedback loop of falling costs, rising purchasing power, and explosive new business formation. The key variable is institutional response speed, not AI capability alone.
SaaS pricing power destroyed
Agentic coding tools enable enterprise clients to replicate mid-market SaaS products in weeks. Renewal negotiations turn brutal, annual price increases die, and the long tail of SaaS has a brutal 2026. ServiceNow-style incumbents see ugly quarters.
Savings fuel new manufacturing & hiring
The $500k enterprise clients claw back from SaaS vendors gets reinvested into physical products, new markets, and specialized talent. A procurement manager's 30% SaaS discount funds a new Ohio manufacturing line employing 200 people.
New business formation doubles
Monthly new business applications hit 700,000 by Q1 2027—double the 2019 baseline. When software build costs drop 90%, the number of viable products explodes. AI destroys old categories slower than it creates entirely new ones.
78 novel drugs approved in 2027
AI protein design tools enable tiny 3-4 person startups to tackle rare diseases and neglected infections Big Pharma wrote off. FDA approves a record 78 novel drugs; median IND-to-approval time falls from 8.1 to 4.2 years.
Housing starts surge 40%, costs fall 25%
AI agents produce full architectural plans and permit-ready drawings in hours. Combined with modular construction and 3D printing, housing starts surge 40% in 2027 while costs fall 25%. The median new home price falls for the first time in a decade.
Millennials finally buy homes; no mortgage crisis
Home prices fall 6% YOY nationally—not from distress but from supply increase. Homeownership rate ticks up for the first time in 15 years. The feared $13T mortgage market collapse never materializes as delinquencies cure faster than any prior cycle.
Ghost GDP: gains accrue only to capital
The nightmare scenario where every dollar saved on headcount flows into AI capability enabling more cuts, with no purchasing power left to circulate. Output appears in national accounts but never reaches consumers.
Core CPI falls to 0.8%—good disinflation
Disinflation is driven by falling costs, not falling demand. Consumer spending in real terms increases even as prices fall because purchasing power expands faster than any wage cut could offset. Real median household income rises 8% in 2027—equivalent to a 14% raise in purchasing power terms.
Only genuine value-providers survive
Companies whose moats were made of friction are destroyed. Incumbents that actually delivered value—not just friction with a friendly face—survive and thrive. New entrants with no legacy overhead capture market share at lower prices.
Unemployment rises to 5.8% then reverses
White-collar layoffs hit software, consulting, financial services, and middle management hard. Unemployment rises from 3.7% to 5.8% by Q1 2027. The human cost is real. But three forces prevent the doom scenario from materializing.
Median reemployment time falls to 11 weeks
AI tools make it dramatically easier to start businesses, learn skills, and match with opportunities. A displaced product manager launches a consumer health startup in six weeks with three AI agents as her team—50,000 paying customers by month four. Job openings rise to 11.2M with 2.4M in new AI-adjacent roles.
AI-era roles command premium wages
AI operations managers, human-AI team leads, and startup consultants often pay more than displaced roles because each human has vastly greater leverage. Revenue-per-employee at AI-native firms rivals Fortune 500 levels with 30 employees.
Bipartisan Transition Economy Act signed March 2027
A 68-32 Senate vote passes three-pillar legislation: an AI Dividend ($4,800/year per adult funded by 2.5% levy on enterprise AI inference), $50B in AI-powered retraining (78% completion rate vs. 30% for traditional programs), and a New Homestead Act with $15k-per-hire incentives for displaced workers.
Positive feedback loop with no natural ceiling
AI improves → costs fall → purchasing power rises → new businesses form → employment grows → spending increases → AI improves. The Gini coefficient begins its first sustained decline since the 1960s. GDP prints 5.1% for the fourth consecutive quarter above 4%. S&P gains 62% from February 2026 levels.
Right-left coalition sustains the Act
The right sees the AI Dividend as market-based without welfare expansion; the left sees broad gain-sharing; fiscal hawks see net revenue as inference spending grows 40% annually. The coalition holds because all sides get something they value.
AI becomes greatest equalizer in economic history
India's feared IT services collapse (rupee -18%, IMF discussions) instead becomes a 9.2% GDP growth story as 1.4B people gain world-class AI tools in their own languages. AI democratizes entrepreneurship globally because intelligence—not money or resources—was always the binding constraint on growth in poor countries.
Karnataka farmer accesses global-grade tools via voice
A farmer in Karnataka uses AI voice agents in Kannada to optimize yields, negotiate with buyers, access micro-insurance, and file taxes. India's 'AI for All' program distributes access through Aadhaar infrastructure. Growth shifts from IT services to agriculture, manufacturing, healthcare, and domestic consumption.
Doom scenario: every assumption fails simultaneously
Every institution fails, every policy response is too slow, every displaced worker stays displaced, and every dollar of surplus accrues only to capital. This is a coherent scenario—but requires all negative assumptions to hold simultaneously, which the authors argue is unlikely given trillion-dollar incentives to adapt.
Friction goes to zero; consumer surplus explodes
AI agents annihilate subscription inertia, information asymmetry, and bad pricing. Travel platforms, insurance renewals, real estate commissions (5-6% to under 2%), and credit card interchange all compress. Median household saves $3,200/year from agent-negotiated optimization alone.