Battery Ventures

The @AnthropicAI S-1 is out and most of the takes I’ve seen are just copy/paste of metrics without much analysis. For me, the most interesting stat is revenue per dollar of compute. In Q1'25, Anthropic brought in $0.40 of revenue for every dollar it spent on compute. By Q2'26 that was $2.08. Quarterly revenue hit $11.6B, up 14.7x Y/Y, while compute and infrastructure costs grew 4.0x to $5.6B. Put another way, compute went from 177% of revenue to 48% in a year! This efficiency helps explain the huge bet Anthropic seems to be making: the company has signed roughly $830B in compute commitments running out to 2036. Meanwhile RPO is $10.8B, with about 75% due within 12 months, and nearly half of revenue comes through cloud marketplaces. The commitments stretch out a decade, but most of the revenue Anthropic can see is due within a year. They’re betting revenue keeps outrunning compute costs, and so far it is. It also shows where the margin in AI infrastructure is headed. The hyperscalers turn megawatts into GPU hours and sell them on long-term contracts. That locks in revenue, but it also caps it. The labs take on the commitments and keep the value they create turning those hours into high-value tokens. As the cost of a megawatt evens out across providers, the money will be in how much value you can squeeze out of each one. The same logic should apply to inference providers like @baseten and @fal. And the next step is bigger - moving up the stack. If frontier labs, inference providers and other neoclouds go from tokens to business output, abstracting tokens from customers while mixing/matching the best tokens to accomplish a task, they can capture even more revenue. This will happen while their costs plateau, giving them even more leverage. That’s the part I’m watching most closely.

The most contrarian thing a VC can do right now is stay on the sidelines! Here's the dirty secret of late-stage AI investing: Just like nobody got fired for buying IBM years ago, nobody gets fired today for overpaying next to the top Silicon Valley VC firms. You get fired for missing the round everyone else got into. So everyone knows the price is crazy, and everyone wires anyway – sometimes multiple times a year or even in multiple tranches with different prices happening simultaneously. Imagine buying the same public stock the same day at 3 different prices! Textbook prisoner's dilemma: every investor makes the rational call, and the industry ends up overpaying. We're paid to generate alpha, but if everyone is piling into the same rounds at the same prices, we aren't generating alpha. What most of the industry is doing right now is buying beta with a FOMO premium. Though I’m extremely bullish on the overall AI opportunity, I think a lot of this activity with 2026 high priced early rounds ends up out of the money. Real conviction sometimes means sitting out the round everyone else is fighting to get into. Founders who are fortunate enough to raise rounds before the music stops, should totally do it if they can manage the dilution. But check if your new investors can truly help you with customers, exec hiring, or strategic decisions if they are just following others and building an investment portfolio including you and your competitors!

VCs inclined to generate alpha will have to pick themes and companies with real conviction, not just party rounds. Hot consensus rounds also ignore the fundamental unit economics that the HBS grad is supposed to be focusing on as Harry says! - Reselling land, power and shell as co-location and GPU hours with no software differentiation will end up in zero or negative gross margins, and won't sustain no matter how large the top-line is or the growth rate is. Remember what happened in the dot-com era! - Locking up you H100 cluster with hundreds of millions in equity raise, is an asset only if you know how to use it to productively generate high margin revenue from your AI product. Time to get real!