An opinion piece on Crunchbase by Anurag Gurtu, CEO of Airrived, argues that a growing number of enterprise AI companies are building what he calls a "forward deployed engineer" (FDE) bubble. The idea: startups sell an "autonomous" AI platform, then hire armies of engineers to sit in customer offices and manually make the product work. Gurtu says this is not a software flywheel but "a payroll with a pitch deck."
The piece cites data showing job postings for forward deployed engineers have surged between 800% and 1,000% year over year, depending on the source. Compensation is high too: Anthropic and OpenAI list base salaries of $185,000 to $320,000, with some total packages reaching $400,000. Gurtu points to large investments in deployment vehicles, such as OpenAI's reported $1.5 billion pledge into a joint venture valued at $10 billion, as evidence that the model is being funded openly.
The core critique is that this model confuses people-market fit with product-market fit. Gurtu suggests a simple test: if the embedded engineer were pulled out tomorrow, would the customer still get what they paid for? If not, the revenue is human-powered, not software-powered. He notes that two companies with the same $20 million ARR can have very different headcount needs, and that human-powered ARR does not get cheaper to deliver as the company scales.
The piece is a guest opinion, not a news report, and it does not name specific companies beyond the cited examples. It warns that CFOs and investors will eventually separate software ARR from human-powered ARR, and that the current hype cycle may not survive that scrutiny.