The Great Global Inflation of Revenue per Employee Brought About by AI

The "Great Inflation of Revenue Per Employee (RPE)," in which AI is fundamentally transforming corporate productivity, is progressing on a global scale from startups to large enterprises. The top AI-native startups are recording an average RPE of $3.48 million, reaching 17 times the average of $200,000 for traditional SaaS companies. Midjourney, with 107–163 employees, is generating $500 million in annual revenue (RPE of up to $12.5 million), while Cursor's Anysphere, with 20–150 employees, is achieving $300 million to $1.2 billion in annual revenue (RPE of up to $15 million) — extraordinary figures by any measure. In its 2026 investment thesis, Sequoia Capital analyzed that "AI-native companies grow 4x faster with 7–8x fewer employees than traditional SaaS companies," and a16z's Marc Andreessen predicted that "the era of the AI-powered one-person billion-dollar startup is coming." Klarna reduced its workforce by 47% while achieving a record quarterly revenue of $1 billion and raised the salaries of remaining employees by 60%. Shopify CEO Tobi Lütke sent a company-wide memo stating that "headcount growth will not be approved unless you can demonstrate that the work cannot be done by AI." The McKinsey Global Institute reports that "labor productivity at AI-adopting companies is improving at 4.8 times the global average rate," while Goldman Sachs projects that "AI will boost U.S. productivity growth by 1.5 percentage points per year." At the same time, the U.S. U-6 unemployment rate has risen to 8.7%, and a "decoupling" of employment from GDP growth is becoming visible. Japan, amid a structural labor shortage characterized by a 2.6% unemployment rate and the tightest labor market in 30 years, is attracting attention as one of the rare economies in the world where AI-driven RPE gains could be "purely positive" — yet only 57% of AI-adopting companies there report measurable productivity improvements (versus the global average of 82%), and the talent acquisition gap remains the greatest challenge. The EU, anticipating the restructuring of 1.6 million jobs (in Germany alone), is moving to formulate a "European AI Social Compact." The great RPE inflation brought about by AI is fundamentally calling into question corporate management, labor markets, macroeconomic policy, and the social contract itself.

The Era of New Metrics — Why "Revenue per Employee" Has Become the Most Important KPI

In the world of venture capital, there is always "one metric" that crystallizes the investment thesis of its era. In the early 2010s, it was MAU (Monthly Active Users), symbolizing the age of growth hacking and network effects. From the late 2010s through the early 2020s, ARR (Annual Recurring Revenue) and NRR (Net Revenue Retention) dominated, measuring the health of the SaaS recurring revenue model.

Emerging between 2025 and 2026 as the metric that most sharply reflects the AI-driven paradigm shift is RPE (Revenue Per Employee).

Sequoia Capital analyzed the structural advantages of AI-native companies in its 2026 investment thesis, "AI in 2026: A Tale of Two AIs," as follows. Compared to traditional SaaS companies, AI-native companies grow 4x faster with 7–8x fewer employees per dollar of revenue. The top AI startups are achieving RPE of over $1 million, and the market is exhibiting "pull" rather than "push" sales dynamics. Sequoia calls this the emergence of the "Zero to Billion Dollar Club"—a cohort of AI companies reaching $1 billion in revenue with an extremely small headcount—placing this at the core of their investment thesis.

a16z's Marc Andreessen predicted in March 2026 "the arrival of an era where AI runs a billion-dollar startup on its own." The argument is that individuals leveraging AI tools can now produce output that once required an entire team. According to a16z's revenue benchmark analysis, the median B2B AI company reaches over $2 million ARR in its first year (double the traditional SaaS benchmark of $1 million), while consumer AI companies are recording $4.2 million ARR in their first year.

The "Lean AI Native Companies Leaderboard" (leanaileaderboard.com), operated by Henry Shi, is becoming an industry standard as a quantitative resource that systematically tracks AI companies from an RPE perspective. The listing criteria are "ARR of $5 million or more, fewer than 50 employees, founded within the last 5 years," and companies meeting these criteria show an exceptionally lean profile: headquartered in the San Francisco Bay Area (51%), with 19 employees, founded in 2021.

Staggering Numbers — RPE Data from AI-Native Companies

Concrete data tells the story of the RPE revolution's scale. The top 10 companies on the Lean AI Leaderboard average $3.48 million in RPE with an average headcount of just 24 employees — a 17x multiplier compared to the traditional SaaS average RPE of $200,000.

Midjourney — the leader in AI image generation. Approximately $500 million in annual revenue achieved with just 107–163 employees. RPE reaches $3M–$12.5M. Notably, the company has raised no external funding and operates entirely on cash flow.

Cursor (Anysphere) — AI coding assistant. $300M–$1.2B in annual revenue with 20–150 people. RPE of $3.3M–$15M. It took just 21 months to reach $1M ARR, starting with a team of 20.

ElevenLabs — AI voice synthesis. Over $100M in annual revenue with 50 employees. RPE of $2M. Two years to reach $1M ARR.

Lovable — AI development tools. An exceptionally rare speed record: $50M in revenue achieved in just 6 months.

CHAI Research — $30M annually with 12 people. RPE of $2.5M.

Gamma — Over $50M annually with an estimated 28 people. RPE of $1.8M.

Comparing these figures to traditional SaaS leaders makes the difference strikingly clear. HubSpot has an RPE of $318,000, Salesforce $479,000, and Adobe $700,000. AI startups are recording 5.7x the RPE even relative to established SaaS leaders.

RPE Evolution Through Each Technology Revolution — Historical Context

To understand the dramatic rise in RPE within a broader historical context, here is an overview of each phase of the technology revolution.

Pre-Internet Era (1980s–1990s). RPE in the manufacturing sector was typically $100,000–$200,000. Early technology companies such as IBM and HP were also operations-intensive, keeping RPE at low levels.

Internet Era (2000s). The "digital distribution" of software fundamentally changed the scaling model. Software written once can be distributed to an unlimited number of users at near-zero marginal cost. Google/Alphabet currently achieves an RPE of approximately $1.3 million—a figure that is the product of the Internet era.

Mobile/Platform Era (2010s). Instagram reached an acquisition valuation of $1 billion with just 13 employees (approximately $77 million per employee), while WhatsApp was acquired for $19 billion with only 55 employees (approximately $345 million per employee). However, WhatsApp's actual revenue at the time of acquisition was only around $20 million, revealing a significant gap between "valuation-based" and "revenue-based" RPE. Apple ($1.9 million/employee) and Meta ($2.2 million/employee, FY2024) were the RPE leaders of this era.

AI Era (2025–2026). NVIDIA ($4.4 million/employee) reigns as the king of AI infrastructure, while top AI startups average $3.48 million (up to $12.5 million). And cases like BuiltWith——$14 million in annual revenue with a single employee——represent the ultimate expression of the "one-person company."

Each technology revolution has brought a step-function increase in RPE. However, the magnitude of the increase in the AI era—10 to 17 times the conventional SaaS benchmark—surpasses every previous revolution.

The Reach of the "One-Person Billion-Dollar Company" — Sam Altman's Prophecy and Its Reality

OpenAI CEO Sam Altman has repeatedly made some of the boldest predictions about how AI will transform the shape of companies.

"We'll soon see a 10-person company worth a billion dollars." "A billion-dollar company with 2–3 people using AI will emerge." "The future of startups might be one person and 10,000 GPUs." "Instead of hiring a designer, just use GPT-6. The need for software engineers, sales staff, and newsletter writers will be far less."

Altman has also revealed that among his circle of tech CEO friends, there is a betting pool to "guess the first year a one-person billion-dollar company appears." The predicted timeline is 2026–2028.

Marc Andreessen of a16z made a similar prediction in March 2026, foreseeing a "boom of one-person billion-dollar startups."

In a February 2025 analysis piece titled "AI Agents Might Produce the First Solo Unicorn — But What About the Social Cost?", TechCrunch argued that a solo unicorn doesn't necessarily need to build a native GenAI product, but will likely come from someone who "leverages GenAI internally at a world-class level."

This prediction is already approaching reality. BuiltWith achieves $14 million in annual revenue with one person, and Testimonial.to and Seats.aero have reached $1.5 million ARR as solo founders. The gap from $14 million to $1 billion is large, but with AI capabilities improving exponentially, the prevailing view in the investment community is that it's a matter of "when," not "if."

AI Efficiency at Large Corporations — Klarna, Shopify, Meta, and Duolingo

The great RPE inflation is not limited to startups. Examples from large corporations demonstrate an even greater impact.

Klarna — The Symbol of "AI Halving the Organization"

Swedish fintech company Klarna has become the most frequently cited example of AI-driven organizational transformation. Its headcount was reduced from 5,527 at the end of 2022 to 3,422 at the end of 2024, and further down to approximately 2,907 in 2025 — a 47% reduction. CEO Sebastian Siemiatkowski revealed that AI is handling the equivalent workload of 853 full-time employees.

Alongside these massive layoffs, Klarna posted record quarterly revenue of $1 billion. RPE reached $1.24 million, a 152% increase compared to Q1 2023. And the salaries of remaining employees were raised by 60%. Siemiatkowski has set a goal of reducing headcount to 2,000 by the second half of 2026.

The structure Klarna's case reveals is clear: AI replaces routine tasks (customer support, back-office processing), a portion of the saved labor costs is redirected toward compensation increases for remaining employees, and another portion is returned to shareholders as profit.

Shopify — "Prove That AI Can't Do It"

An internal memo sent by Shopify CEO Tobi Lutke to the entire company in April 2025 sent ripples through the tech industry. "Managers cannot request additional headcount unless they can prove the work cannot be done by AI." AI utilization is now part of performance reviews, and all employees are expected to "handle 100 times the workload."

Shopify carried out a 30% headcount reduction from 11,600 in 2022 to 8,100 at the end of 2024, while revenue has continued to grow at an annual rate of 20–40%.

Meta — $115–135 Billion in AI Investment and Consideration of 20% Cuts

Meta had 78,865 employees as of end-2025, but reports emerged that the company was considering cutting approximately 20% (around 15,800 employees) to offset its $115–135 billion AI capital investment planned for 2026. CEO Mark Zuckerberg stated that "AI tools are enabling smaller teams to do work that used to require much larger teams." Meta is planning $600 billion in AI data centers by 2028, and the efficiency pressures to support that investment will only intensify going forward.

Duolingo — 4–5x Productivity Gains Without Layoffs

Language learning platform Duolingo took an approach of gradually reducing external contractor engagements without laying off full-time employees. CEO Luis von Ahn stated that AI is making employees "4–5 times more productive." Daily active users surged to 47 million (up 51% year-over-year), and 2025 revenue is projected at $1.02 billion.

Chegg — Lessons from the "Disrupted" Side of AI

Not every company is a winner in the AI revolution. Ed-tech company Chegg saw its entire business model (an answer database) replaced wholesale by the emergence of ChatGPT, with its stock price falling 96% from its peak and revenue declining 24%. Chegg's case serves as a cautionary tale showing the flip side of AI-driven RPE gains — the complete obsolescence of traditional businesses.

McKinsey, BCG, Goldman Sachs — Perspectives from Consulting Firms and Analysts

What about macro-level analysis?

McKinsey Global Institute (December 2025) concluded that "the greatest economic impact from AI is likely to come from a subset of companies adopting it 'all in.'" Industries adopting AI are seeing labor productivity improve at 4.8 times the global average rate, and high AI-exposure sectors show 3 times higher revenue growth per employee. However, despite 88% of companies using AI in some capacity, only 39% report any impact on EBIT (earnings before interest and taxes), and the vast majority of those see an impact of less than 5%. The world holds $600 trillion in wealth yet faces a severe productivity shortfall, and AI is the most promising tool for closing that gap.

BCG (2025), in a joint study with Harvard Business School, ran an experiment with 758 BCG consultants and demonstrated that AI users completed 12.2% more tasks, 25.1% faster, and with over 40% higher quality output. However, a separate BCG report sounded the alarm that 60% of companies are failing to generate meaningful value from their AI investments, with only 5% succeeding in creating value at scale. AI agents are projected to grow from 17% (2025) to 29% (2028) of total AI value.

Goldman Sachs offers the most cautious outlook. As of March 2026, it concluded that "there is still no significant correlation between AI and economy-wide productivity." That said, at the level of specific tasks, a median 30% productivity improvement has been confirmed, and Goldman projects AI will boost U.S. productivity growth by 1.5 percentage points per year. However, the firm notes that GDP impact exceeding 0.1 percentage points will not materialize until 2027 or later for the U.S., and 2028 or later for other major economies.

Moody's Analytics (February 2026) forecast that AI will contribute 0.50 percentage points to U.S. real GDP growth in 2026, but warned that the economic benefits are highly concentrated among shareholders, exacerbating widening inequality. The five most aggressive AI infrastructure investors are projected to deploy more than $700 billion in 2026.

The "Decoupling" of Employment and GDP — Macroeconomic Consequences

The most significant macroeconomic consequence of the great RPE inflation is the decoupling of employment growth from GDP growth.

Historically, GDP growth has been strongly correlated with employment growth. When the economy grows, jobs increase; when jobs increase, consumption rises; when consumption rises, GDP grows further. This virtuous cycle was the foundation of post-war advanced economies.

AI has the potential to structurally alter this cycle. If companies can generate more revenue with fewer employees through AI, GDP growth can proceed independently of employment growth.

US labor market data is showing early signs of this shift. The unemployment rate has risen to its highest level in four years, and the U-6 unemployment rate (including part-time and underemployed workers) has reached 8.7%. The three-month average of new jobs created has fallen to 22,000, below the 40,000 needed to stabilize the unemployment rate. A new labor market pattern of "low-hire, low-fire" — where companies neither lay off workers nor hire new ones — has become entrenched since the summer of 2025.

The ainvest analyst captured this situation most aptly: "A polarized economy scenario where GDP growth of 4–5% is maintained despite job losses. Capitalists thrive while task-based workers struggle."

There is also a deeper structural problem. The US tax code allows companies to immediately expense AI servers (bonus depreciation), while deductions for employee retraining costs are constrained by six Internal Revenue Code provisions. In other words, the tax system structurally incentivizes the replacement of humans with AI.

Japan — The Only Economy in the World That Can Be "Purely Positive"

Within global analysis, Japan occupies an exceptionally unique position.

With an unemployment rate of 2.6%, labor shortages at a 30-year high, a shortage of 220,000 IT professionals (estimated 2025–2026), and 1.3 million vacant technology positions overall, Japan faces structural and irreversible demographic pressures from aging and population decline. In this context, AI-driven RPE improvement can function not as a replacement for labor, but as a supplement to fill the labor shortage.

OECD and IMF analyses place Japan in a different category from other advanced economies. Japanese AI users are more likely to expect AI to "create" jobs rather than take them away, and hold highly positive views on AI's impact on workplace performance and wages. AI-induced job losses are less likely to be severe in Japan than in other countries, due to long-term employment practices (lifetime employment) and chronic labor shortages.

However, the challenges are serious. Only 57% of Japanese companies report experiencing ROI or productivity gains from AI, significantly below the global average of 82% — a 25-point gap. There is a severe shortage of talent capable of bridging AI knowledge with on-the-ground operational experience, and this gap has become the single largest bottleneck to AI adoption.

From an investment thesis perspective, Japan is one of the few major economies in the world where AI adoption can be purely accretive (value-additive). It fills vacancies rather than displacing jobs. The constraining factor is not the risk of job loss, but the speed of adoption. This structural characteristic points to investment opportunities in AI productivity tools tailored to the Japanese market and in the development of Japanese-language-specific enterprise AI agents.

EU — The Cost of Regulation-First and the "European AI Social Compact"

The EU faces a triple disadvantage.

First, a regulatory gap. The EU AI Act covers safety, transparency, and ethics, but does not directly address socioeconomic impacts or effects on employment. This structural void has been widely criticized.

Second, a competitiveness lag. Europe is home to only 4 of the world's top 50 technology companies, and there is an annual R&D investment gap of $700 billion with the United States. European companies' returns trail U.S. performance by approximately 25%.

Third, disproportionate impact. Women are nearly twice as likely as men to hold jobs with high AI exposure, and in Germany alone, an estimated 1.6 million jobs are projected to be restructured or lost over the next 15 years.

In response, the ETUC (European Trade Union Confederation, representing more than 45 million workers) has warned that AI's positive benefits will be nullified "if monopolized by a handful of technology companies," and the Carnegie Endowment for International Peace published "How Europe Can Survive the AI Labor Transition" in February 2026. Discussions are underway around developing a "European AI Social Compact" — a comprehensive framework of worker protections and retraining anchored to a fiscal framework.

The EU's fundamental dilemma is that it wants to reap the productivity gains of AI, yet fears becoming a consumption market for American and Chinese AI. A regulation-first approach risks slowing adoption rates and widening the gap further.

"AI Leverage" — The Core Concept of the VC Investment Thesis

In the context of VC investment, the core concept theorizing the great inflation of RPE is "AI leverage."

Internal AI Leverage. AI-native companies utilize AI agents across their own operations (legal, recruiting, sales, engineering) to achieve compounding efficiencies. Sequoia Capital calls these "self-improving companies." The more a company grows, the more data AI processes, the better AI performs, and the more efficiencies are generated — a flywheel effect.

The Capital Efficiency Revolution. The market capitalization of private AI companies grew 130% in 12 months, from $283 billion to $658 billion. AI companies' net retention rates average 132% (compared to 108% for traditional SaaS). Because the marginal cost of serving additional customers approaches zero for AI companies, rising RPE at scale is structurally guaranteed.

Valuation Premium. The median EV/Revenue multiple for AI startups reaches 25–30x. Total VC investment in AI in 2025 was $258.7 billion, with AI accounting for 61% of all VC investment.

The "Burn Multiple" Shift. The "growth at all costs" era of 2021–2022 is over, and capital efficiency has become the central criterion for AI investment. Companies where AI has structurally elevated RPE grow faster with less capital, improving VC investment returns.

Wage Inflation vs. Employment Deflation——The Central Paradox of the AI Economy

The great inflation of RPE creates a fundamental paradox in labor economics: the simultaneous progression of wage inflation for "survivors" and employment deflation for "exiters."

The Klarna case is the most striking example. A 47% workforce reduction and a 60% wage increase are happening at the same time. Companies use AI to accomplish more work with fewer people, redirecting a portion of the reduced labor costs to the compensation of remaining employees (wage inflation), while returning the rest to shareholders as profit (increased capital returns). Meanwhile, those who have exited face a shrinking job market (employment deflation).

A December 2025 analysis by CNBC noted that "AI may suppress wage inflation." As AI replaces routine tasks, workers' bargaining power shifts from labor to capital. The Federal Reserve faces a policy dilemma: weakness in employment suggests monetary easing, yet AI-driven supply-side efficiency gains could make that easing inflationary.

Signs of structural AI-driven disinflation (a slowdown in the rate of price increases) are also emerging. AI-optimized logistics are cutting costs by 5–12%, exerting an estimated downward pressure of 0.5–0.7 percentage points on annual CPI (Consumer Price Index). Central banks face an unprecedented challenge: distinguishing healthy, technology-driven disinflation from harmful deflation caused by insufficient demand.

Timeline——From the Corporate Level to the Macro Level Ripple Effects

Organizing the ripple timeline of the RPE revolution based on collected data.

2025 (already reality). AI startups routinely achieving over $1M RPE. Klarna cuts headcount in half. "Low hiring, low firing" labor market pattern emerges. Corporate-level revolution already underway.

2026. AI contributes 0.50 percentage points to US GDP growth. Top 5 companies' AI capital investment exceeds $700 billion. Employment-GDP growth decoupling becomes visible.

2027. Goldman Sachs projects a measurable impact on US GDP (exceeding 0.1 percentage points).

2027–2028. Emergence of the first billion-dollar one-person companies (Altman prediction).

2028. Measurable GDP impact begins in other major economies. AI agents account for 29% of total AI value.

2028–2030. AI compute costs decline as data centers reach full operation. Marginal margin improvement accelerates.

2030–2035. If broadly adopted, AI boosts US productivity growth by 1.5 percentage points per year (Goldman Sachs 10-year outlook).

The most critical insight here is that there is a time lag between the corporate-level revolution (2025–2026) and macro-level statistical manifestation (2027–2028). This gap creates a "policy blind spot" — large-scale structural changes in the labor market proceed before policymakers have the data to respond.

Impact on the Industry

The great inflation of RPE driven by AI brings about the following irreversible structural changes.

First, the design principles of corporate organizations change at their foundation. We have entered an era where the answer to "how many employees are needed?" varies by orders of magnitude depending on the level of AI adoption. Shopify CEO Lutke's directive — "prove that this is a job AI cannot do" — will become the default hiring policy for all companies going forward. Organizations will be redesigned as "human + AI hybrid teams," and RPE will become the CFO's most important KPI.

Second, the evaluation criteria for VC investment will transform. In addition to "ARR growth rate," "RPE" and "AI leverage ratio" will become core metrics for investment decisions. Even with the same $100 million in ARR, a company achieving it with 20 employees is far more attractive to investors than one requiring 2,000. The former demonstrates structurally higher operating margins and significant room for scaling.

Third, the polarization of the labor market will accelerate. Demand and compensation for high-skilled talent capable of leveraging AI will surge, while demand for routine work susceptible to AI replacement will shrink. Klarna's "47% reduction, 60% pay raise" is a microcosm of this polarization. Middle-tier employment — roles that are routine yet require a certain level of expertise — will face the greatest pressure.

Fourth, the premises of macroeconomic policy will be overturned. The decoupling of GDP growth from employment growth challenges the policy frameworks that have prevailed since Keynesian economics. Whether traditional tools of monetary policy (interest rates) and fiscal policy (spending and taxation) can function effectively in an AI-driven economic structure remains an open question. The pace at which discussions of Universal Basic Income (UBI) and an "AI robot tax" transition from theory to policy options will accelerate.

Fifth, Japan has the opportunity to establish a unique position as "the most AI-friendly major economy." Against the structural tailwind of labor shortages, Japan's situation — where AI can be introduced as a labor supplement rather than a replacement — faces the least friction in AI adoption, both politically and socially. Closing the gap in Japan's AI adoption rate (25 percentage points behind the global average) will become the core of the nation's productivity challenge.

References: Sequoia Capital, "AI in 2026: A Tale of Two AIs" (2026); a16z, "Revenue Benchmarks for AI Apps" (2025); Marc Andreessen, "Solo Billion Dollar Startups Prediction" (March 2026); Sam Altman, Various Interviews on One-Person Companies (2024–2025); Henry Shi, "Lean AI Native Companies Leaderboard" (leanaileaderboard.com); Jeremiah Owyang, "AI Startups Are Dominating Traditional Software in One Key Metric" (May 2025); Klarna CEO Sebastian Siemiatkowski Interviews (CNBC, 2025); Shopify CEO Tobi Lutke Internal Memo (April 2025); McKinsey Global Institute, "AI Productivity Outlook 2026" (December 2025); BCG, "AI at Work 2025"; BCG / Harvard Business School Consultant AI Experiment; Goldman Sachs, "AI May Start to Boost US GDP in 2027"; Goldman Sachs, "No Meaningful Economy-Wide AI Productivity Relationship Yet" (March 2026); Moody's Analytics, "Macroeconomic Consequences of AI" (February 2026); OECD, "AI and the Labour Market in Japan"; IMF, "Impact of Aging and AI on Japan's Labor Market"; Carnegie Endowment, "How Europe Can Survive the AI Labor Transition" (February 2026); European Policy Centre, "AI's Impact on Europe's Job Market"; ETUC AI Position Paper; Crunchbase, "Global VC AI Investment Report 2025"; TechCrunch, "AI Agents Could Birth the First One-Person Unicorn" (February 2025); CNBC, "AI Could Weigh on Wage Inflation" (December 2025); PwC Japan, "Survey on the State of Generative AI, Spring 2025"