SaaS Says: Revenue per Compute

Revenue per Compute

Jim Cook got a call from a client last week. "I just had one the other day with one of my clients that got surprised by a $500,000 a month bill on compute."

Cook, a former Netflix CFO, has seen this before. Cloud in 2013: engineers got the power to spend and finance had no language for it. "And the answer is no controls. We don't have controls because we don't even have the definition of talking about or this concept that compute costs money." His fix is not new. "I fundamentally believe it's applying what we already know to a different problem. We've already had to tag cloud costs." Then he moves the scorecard. "We all want our leaders thinking about the cost of their business, but we used to think about it in terms of headcount cost." His replacement: "I'm going to start talking to you about revenue per compute, not revenue per head count."

Aaron Goldsmid, who runs product at Deel, has already stopped counting heads. "I actually don't pay a lot of attention to the count anymore, right?" Headcount used to stand in for complexity and output. "in the last 18 months, those proxies are just not valid anymore, right?" Deel says agents now do the work of 600 employees and found $140 million in new ARR without hiring. The cost shows up on the other side of the ledger. "every software is less sticky than it was 12 months ago, right?"

Jad Naous, CEO of Grepr, runs a 10-person team that finished one project in "maybe about 10 days. In previous days, that would have taken maybe about 6 months." His buyers still pay for dollars, not AI. Offer them a 90% cut to their Datadog bill and they sign. "But the AI thing, they're like, no, no, no, no." Naous sells that cut, on a show from his former firm.

Sanjay Brahmawar, CEO of QAD Redzone, hears the same from manufacturers buying his software. Nobody asks him which model. The budget is a carve-out, not new money: "we are carving out 5 to 7% of our Opex IT budgets because we want to invest in AI".

George Kurdin is the holdout on the meter. The Monk CEO is "not really thinking too much about like internal inference or like our anthropic bill". He calls model costs "like really easy solve world problems in the future." His customers have also stopped buying hours saved. "I could make the case that time savings or efficiency isn't even that impressive anymore". Monk sells AI receivables software, and the show's parent fund backs it.

Last week finance took the meter. This week it wants it on the scorecard.

Sources: Interviews from Run the Numbers (Oct 15, Jim Cook), Product School (Oct 14, Aaron Goldsmid of Deel), a16z (Oct 16, Jad Naous of Grepr), Manufacturing Happy Hour (Oct 13, Sanjay Brahmawar of QAD Redzone), and GTMnow (Oct 14, George Kurdin of Monk).