“How many employees do you have?” That’s the question we reach for, almost by reflex, when sizing up a company. Total revenue, operating profit, headcount — for decades these three numbers have set a firm’s place in the world. Of course, the center of gravity in valuing a business still lies in growth rate and cash flow. But a metric that has long been treated as a footnote — revenue per employee — has begun to draw real attention in the AI era. A company earning $100 million with 50 people and a company earning the same with 5,000 are not the same economic creature. That difference is starting to decide who will compete and who will not.
The numbers say so. Midjourney, the AI image-generation company, reportedly earns roughly $500 million a year with about 100 employees. Estimates put revenue per employee around $4.7 million — a figure that climbs higher still when you note the company has taken essentially zero outside investment and zero marketing spend. Cursor, the AI coding tool, is approaching $1 billion in annual recurring revenue only a few years after founding, posting what is now considered the fastest B2B SaaS growth on record. Their per-employee revenue runs several multiples above traditional software companies. Nvidia, with about 30,000 employees, booked $130.5 billion in revenue, generating roughly $4.4 million per employee — an astonishing figure, multiples higher than two years ago, propelled by the global surge in AI semiconductor demand.
Big Tech has caught the same current. After cutting 21,000 jobs and reshaping its cost base, Meta — helped along by a recovery in the ad market and a few other tailwinds — grew revenue 25% and saw its share price jump 21% after earnings. Shopify’s CEO issued an internal directive: “Before requesting any new hires, prove that AI cannot do the job,” and formally folded AI fluency into performance reviews. Amazon is restructuring tens of thousands of roles and accelerating its rollout of AI agents. Across Big Tech, the strategic shift is unmistakable — toward generating more value with fewer people.
So what about Korea? SK Hynix, riding the AI-semiconductor boom, posted its strongest results ever; Naver crossed 10 trillion won in annual revenue for the first time. But set those alongside Apple’s roughly $2.38 million in revenue per employee, or Meta’s roughly $2.2 million, and the gap with leading Korean firms is still pronounced. More worrying still, the gap in AI adoption between large companies and small-and-medium enterprises is widening, and the industry is openly anxious about where that productivity divide leads.
For workers, all of this is a direct anxiety signal. Anthropic’s CEO has warned that a substantial share of entry-level office work could be significantly affected. The World Economic Forum projects that roughly 40% of core job skills will change by 2030; Goldman Sachs estimates that some 300 million jobs worldwide could be exposed to AI. Standardized knowledge work — customer support, data entry, junior analysis, paralegal tasks — will be hit first.
That said, AI-can-do-everything narratives have holes. The fintech company Klarna replaced hundreds of customer-service agents with AI, then had to hire people back as quality fell and complaints stacked up. The CEO acknowledged that the push for cost savings had outrun service quality. A joint Harvard-BCG study found that on standardized tasks, consultants using AI worked 25% faster and produced results 40% higher in quality. The lesson from that study is hard to miss. It isn’t AI as a replacement — it’s AI as a collaborator.
So what is there to do, right now? Stop fearing AI and start applying it to your own work. LinkedIn data shows that the number of profiles listing AI-related skills, and the enrollment in generative-AI courses, have both roughly doubled. The shift is already underway, and it is accelerating. The people who survive the AI era will not be those who escape replacement; they will be those who use AI best. Whether you are a company or a single worker, the yardstick of per-person productivity now leaves no one outside its measure.