Airrived Co-Founder Anurag Gurtu: Enterprise AI Companies Disguise Service Costs as Software Valuations with Forward-Deployed Engineers
Anurag Gurtu, co-founder and CEO of Airrived, argues in a guest commentary that multiple enterprise AI companies are hiring large armies of forward-deployed engineers (FDEs) to hand-build workflows at customer sites that their products were supposed to automate, then booking the labor cost as software ARR to earn software valuation multiples. He poses a testable question: if you removed the embedded engineer, would the customer still get what they paid for? He cites OpenAI's pledge of up to $1.5 billion into DeployCo, FDE job postings growing 800% to 1,000% year over year, and Anthropic base salaries of $280,000 to $320,000 to illustrate the model's scale. Gurtu contends this is not operating leverage but 'people-market fit,' whose ceiling is that you can never hire humans cheaply or quickly enough to sustain a software multiple.
ImportanceMaterialEvidenceE3 inspectableWrite-upStandard
An enterprise AI company just closed another nine-figure round to solve a problem software is supposed to solve by itself: getting humans to use the product. In a guest commentary published on Crunchbase News on October 19, 2026, Anurag Gurtu, co-founder and CEO of Airrived, calls this pattern 'Silicon Valley's most expensive game of pretend.'
His breakdown runs through seven steps: sell the customer an 'AI platform'; discover the platform doesn't actually do the thing on its own; send in forward-deployed engineers who spend six months hand-building the workflow the website implied already existed; bill the customer for 'software'; book it as ARR; raise a bigger round because 'ARR go up'; repeat forever, with an increasingly large and increasingly expensive engineering org standing between the product and the promise it made. Gurtu's judgment: this is not a flywheel. It's a payroll with a pitch deck.
The data supporting this assessment point to scale. Gurtu cites multiple sources: Lightcast data reported by Fortune put 'forward deployed engineer' job postings up more than 1,000% from January to August 2026; Financial Times data carried by PYMNTS showed listings up more than 800% from January to September 2025; Indeed data showed postings rising from 643 in April 2025 to 5,330 in April 2026. On compensation, Anthropic lists base pay of $280,000 to $320,000 for the role and OpenAI $185,000 to $300,000 base plus equity; Fortune reported some Anthropic roles reaching $400,000 in total compensation.
More structurally telling is OpenAI's own move. Gurtu notes that OpenAI pledged up to $1.5 billion into a deployment joint venture valued at $10 billion, with another $4 billion from TPG, Bain Capital, Advent, Brookfield and Goanna. The vehicle, DeployCo, has been reported with more than $4 billion in investment. Gurtu specifically mentions that OpenAI guaranteed the private-equity backers of its deployment vehicle an annual 17.5% return over five years — a different structure from a growth thesis.
Gurtu's core test question is a single one: if you pulled the embedded engineer out tomorrow, does the customer still get what they paid for? If the honest answer is 'well, no, but the ARR is real' — that's not product-market fit. That's people-market fit. He reasons that people-market fit has a brutally unglamorous ceiling: you can only hire humans so fast, and you can never hire them cheap enough to earn a software multiple. Every new logo doesn't get more profitable to serve. It gets another headcount request.
He illustrates the cost-structure difference with a comparison: two companies, same $20 million ARR. One needs 10 engineers to keep the lights on. The other needs 100. Both slides say 'AI software company.' The difference shows up in gross margin, cost to onboard the next customer, and whether that cost falls as the company scales. Gurtu argues that software ARR should get cheaper to deliver as it scales. Human-powered ARR does not.
He acknowledges the model feels great in a hype cycle — companies funding their FDE armies with fresh nine-figure rounds are betting nobody asks the unfashionable questions until after the round closes and the confetti is swept up: What does it actually cost, fully loaded, to onboard the next customer? Does that number go down as the company scales, or does it just grow the org chart? What happens to gross margin if first-year growth slows and the headcount can't hide behind hypergrowth anymore?
Gurtu's conclusion: your AI isn't autonomous. But your business model might be — autonomously transferring the cost of a services company onto a software valuation, for as long as the music plays.
It should be noted that Gurtu does not name specific companies (other than OpenAI's DeployCo), and his central thesis — that embedded engineering is a permanent cost structure rather than a transitional phase — is an analytical judgment, not a proven fact. Complex enterprise sales genuinely require deep integration in their early stages, and some forward-deployed effort may diminish as products improve. Gurtu provides no quantified transition benchmark and does not discuss under what conditions the FDE model would converge. Readers should treat this as an actionable test framework, not a blanket verdict on all enterprise AI companies.