Startups & Investment

AI Is the Bubble? Anthropic’s Numbers Make Old Tech Look Bloated

Anthropic’s revenue per employee dwarfs Pinterest, DocuSign and Zoom. The bigger risk may be old software’s pricing power as AI platforms take on more work.

By Daniel Mercer Edited by Samantha Reed Published: Updated:
AI Is the Bubble? Anthropic’s Numbers Make Old Tech Look Bloated
Anthropic’s revenue per employee challenges the idea that frontier AI is all hype and no business. Photo: cottonbro studio / Pexels

Key Notes

  • Anthropic reported over $6.6 million in trailing revenue per employee with approximately 5,300 staff, far above estimates calculated for Pinterest, DocuSign and Zoom using their latest financial reports and disclosed workforce counts.
  • The comparison supports a bullish case for frontier AI’s commercial scale, while differing reporting dates, cloud costs and revenue accounting prevent it from proving individual productivity or profitability.
  • The larger competitive question is whether established software vendors can defend pricing power, since strong AI demand does not automatically justify any valuation or make lower-multiple public companies bubbles.

The easiest way to dismiss the AI boom is to point at a trillion-dollar valuation and call it a bubble. The harder task is explaining what happens when a company with roughly the workforce of Pinterest generates many times more revenue per employee, while building software that can do an expanding share of other companies’ work.

That is the challenge posed by Anthropic. Its IPO filing reports approximately 5,300 employees and more than $6.6 million in trailing 12-month revenue per employee as of September 30, 2026. This is a completed-period sales measure, rather than a projection that every dollar of future demand will materialize.

For Anthropic and OpenAI, those kinds of commercial results make the blanket dismissal increasingly unconvincing. Our view: the more interesting risk may lie in established software companies’ assumptions about how much customers will keep paying for narrower products. The question is where durable earning power will sit as AI takes on more work.

What 5,300 Employees Actually Support

Start with the numbers circulating around Anthropic. The $65 billion figure was an annual revenue run rate reached by the end of July, Reuters reported. It extrapolated the pace of sales; it was not $65 billion already booked over a completed year. The workforce figure is also closer to 5,300 than 5,000 in the latest filing.

The reported $2 trillion valuation is an IPO expectation, rather than a completed public-market price. More recent Reuters reporting described some investors considering roughly $1.5 trillion. Anthropic’s last company-announced financing, in May, valued it at $965 billion. Those distinctions change the investment calculation, but they do not erase the operating business underneath it.

The $6.6 million-per-employee disclosure is the more useful starting point for this comparison. Revenue per employee measures how much business an organization supports relative to its direct workforce. It cannot tell us how much value one individual personally created, or how much cash remains after serving customers and funding research.

Pinterest, DocuSign and Zoom: The Updated Comparison

Pinterest disclosed 5,265 employees at December 31, 2025 in its annual report. Combining its 2025 revenue with the subsequent nine-month figures in its October 29 results gives approximately $4.68 billion in trailing revenue through September 2026. Divided by that disclosed workforce, that is about $889,000 per employee.

DocuSign reported 7,044 employees at January 31, 2026 in its annual filing. Rolling that year’s sales forward using its latest quarterly results produces approximately $3.36 billion in trailing revenue through July. The same calculation yields about $477,000 per employee.

Zoom reported 7,438 full-time employees at January 31, 2026 in its annual report. Its latest earnings release brings trailing revenue through July to approximately $4.99 billion, equivalent to about $671,000 per disclosed employee.

Against those estimates, Anthropic’s reported figure is roughly 7x Pinterest’s, 14x DocuSign’s and 10x Zoom’s. The gap is large enough to challenge the instinct that a bigger workforce must support a more substantial business.

There are limits to that comparison. The public-company estimates use the latest annual workforce counts cited above, not live November payrolls or average staffing during each revenue period. Their reporting periods differ, and contractors and supplier workforces are outside the denominator. We calculated trailing sales by adding current year-to-date revenue to the last full year and subtracting the comparable prior-year period.

The Bull Case Is Bigger Than a Payroll Ratio

Reducing the older companies to pictures, PDF signatures and video calls misses why customers pay them. Pinterest operates a visual discovery and advertising marketplace. DocuSign provides agreement workflows, while Zoom runs communications products for businesses. Trust, integrations, distribution and reliability can be expensive to build and difficult to replace.

But frontier AI platforms have a different expansion opportunity. The same underlying model can assist with coding, documents, analysis and multi-step workflows. A business buying those capabilities is potentially paying for work across several departments, rather than adding one isolated feature to one team’s software stack.

That is the strongest argument against treating OpenAI and Anthropic as empty speculation. Their opportunity is to capture a share of spending on both software and the work performed through it. If useful tasks become cheaper to complete, demand can grow beyond the budgets attached to today’s individual applications.

OpenAI’s annualized revenue was approaching $50 billion at September’s end, the Financial Times reported. That figure belongs in the run-rate category, rather than beside a completed year of sales. OpenAI and Anthropic recognize some partner sales differently, Axios reported. A larger headline revenue figure therefore does not automatically establish superior margins or economics.

For established software vendors, the pressure point is pricing power. If an agent can coordinate a workflow across several tools, customers may question how many separate subscriptions and paid seats they need. If the incumbent controls essential records, approvals or customer relationships, however, it can remain valuable and sell AI capabilities itself.

There is evidence of that response already. DocuSign says its Intelligent Agreement Management business represented 15.1% of annual recurring revenue in July. Zoom says its AI-first customer-experience portfolio is growing. These businesses are competing for the opportunity, and their existing customers give them an advantage that a headcount comparison cannot capture.

Real Demand Still Has a Price

The valuation evidence deserves the same honesty as the workforce figures. At the October 30 close, Stock Analysis put Pinterest’s market value at about $11 billion, DocuSign’s at $15 billion and Zoom’s at $28 billion. Using the trailing sales calculated above, those equity values represent approximately 2.4x, 4.6x and 5.6x revenue, respectively.

Those are equity-value-to-sales comparisons, rather than enterprise-value multiples adjusted for cash and debt. Even so, they show why a larger payroll cannot, by itself, establish that these stocks are bigger bubbles. The market is assigning very different prices to different growth expectations.

At a hypothetical $2 trillion valuation, Anthropic would be priced at roughly 31x the older $65 billion annualized sales rate. That is a scenario using a run rate, not a trailing-sales multiple or a confirmed IPO price. A compelling company can still be an expensive investment, particularly if growth slows or competition pushes down prices.

Compute also replaces some of the costs that headcount leaves out. A small direct workforce depends on chips, cloud capacity and the people who build and operate that infrastructure. Anthropic’s first-half operating loss narrowed to $693 million, but the company says compute and infrastructure will remain its largest operating expense. Serving more demand must eventually produce durable returns on those resources.

OpenAI’s latest fundraising provides another reality check: a reported reduction in its financing target, with the round still open. Investor resistance to a price or funding plan can coexist with substantial demand for the product.

The bullish case survives those qualifications. Anthropic’s workforce supports a scale of revenue that makes the familiar software staffing model look surprisingly heavy. OpenAI is pursuing a similarly broad market for useful AI work. The risk for incumbents is that their products become easier to substitute while their organizations remain costly to run.

Calling that entire shift a bubble is too easy. The harder questions are which companies can keep customers, retain pricing power and turn revenue into cash—and whether investors are paying a sensible price for that future. On the operating evidence, the AI labs deserve considerably more credit than the dismissive label allows.

Disclaimer: AIstify is an independent media brand owned and operated by NuvexMedia LLC, publishing news, research, and insights on artificial intelligence, emerging technologies, automation, and related industries. NuvexMedia LLC invests in and collaborates with companies across the AI, technology, software, and digital innovation sectors. These relationships do not influence AIstify’s editorial coverage, and the publication maintains full editorial independence to provide accurate, timely, and objective information. © 2026 NuvexMedia LLC. All rights reserved. This content is for informational purposes only and should not be considered legal, tax, investment, financial, or other professional advice.

AI & Machine Learning, News, Startups & Investment