In December 2025, Sacra’s research on Midjourney confirmed a figure that still seems implausible: $500 million in annual recurring revenue, generated by approximately 107 employees. That is $4.7 million in revenue per employee — roughly ten times what Salesforce produces per head, and more than any venture-scale SaaS company in history had achieved at comparable revenue.

Then in February 2026, TechCrunch reported that Cursor had surpassed $2 billion in annualised revenue. Its team is estimated at 50 to 150 people. At the midpoint, that is $20 million in revenue per employee. The history of software has no precedent.

These are not outliers. They are the leading edge of a structural pattern.

The Data

The conventional productivity benchmark for enterprise SaaS is the Rule of 40: a company’s revenue growth rate plus profit margin should exceed 40%. By that standard, a company growing at 25% with 15% margins is healthy. Revenue per employee benchmarks for scaled SaaS businesses sit at a median of ~$130,000 for private companies, rising toward $250,000 to $500,000 for IPO-stage businesses, according to SaaS Capital. Salesforce generates roughly $480,000 per employee; HubSpot, around $343,000. These are the benchmarks that AI-native companies are shattering.

AI-native companies are operating in a different register entirely. Midjourney reached profitability within two months of its July 2022 launch without any external venture funding or paid marketing. The product grew through Discord, word of mouth, and the inherent virality of generated images. No sales team. No outbound. No traditional marketing budget. The company crossed $300 million ARR in 2024 and $500 million in May 2025, according to Sacra’s analysis.

Cursor’s trajectory is steeper. The AI coding tool crossed $100 million ARR in January 2025, $500 million by June, $1 billion in November, and $2 billion by February 2026 — roughly doubling every two to three months. The company raised $2.3 billion in its most recent round at a $29.3 billion valuation. That is a 14.5x ARR multiple, which looks expensive by traditional SaaS metrics and reasonable by AI-native ones.

The structural reasons for this density are not complicated:

Distribution is ambient. Midjourney users discovered the product inside Discord. Cursor users discovered it inside their IDE. Neither company built a traditional acquisition funnel. The product is where users already are.

Marginal delivery costs approach zero. An additional user adds near-zero incremental cost. The infrastructure scales; the headcount does not need to.

AI handles functions that previously required headcount. Support, QA, documentation, parts of product development — all partially or substantially automated. The headcount-to-revenue ratio breaks because the denominator no longer grows proportionally.

Why It Matters

For investors, revenue per employee is becoming a first-order diligence metric for AI-native companies — alongside ARR, NRR, and burn multiple. A company generating $5 million per employee at Series A is structurally different from one generating $500,000, and pricing those companies identically is a category error.

The valuation implication is real. Cursor at $29 billion on $2 billion ARR looks expensive at 14.5x. Reframe it as capital efficiency — a company extracting $2 billion in value from a team of fewer than 150 people — and the multiple starts to make sense. The comparable is not Salesforce at scale. It is the earliest SaaS companies before enterprise sales headcount normalised.

For founders, the benchmark has shifted. The target is not “build a great product and then add a sales team.” The target is “build a product that grows without one.” Midjourney and Cursor are the proof points that this is achievable at nine-figure ARR. The constraint is no longer capital — it is whether the product has the inherent virality and daily utility to drive adoption without a traditional GTM machine.

The Charaka View

Manthan Intelligence’s portfolio evaluation framework now includes an explicit operational efficiency score — revenue per employee indexed against company stage. What the Midjourney and Cursor data is teaching us is that AI-native companies with sub-150 headcount and $100 million or more in ARR represent a genuinely new category of investment thesis. The moat is not patents, not proprietary data, not exclusive distribution. It is the product loop itself: tools that become indispensable inside the workflows where users already spend their time, with economics that scale without proportional headcount growth. That is a pattern worth paying a premium for.

This analysis draws on Sacra’s Midjourney research (December 2025), TechCrunch’s report on Cursor’s $2B revenue milestone (March 2026), Sacra’s Cursor research, and SaaS Capital’s revenue-per-employee benchmarks. Human editorial oversight applied.

This analysis is informational and does not constitute investment advice, a research report, or a recommendation to buy, sell, or hold any security.

Charaka Notes by Manthan Intelligence. Subscribe