Special Edition: The week AI became infrastructure

This week, in a single seven-day stretch, six different stories made it crystal clear that AI just stopped being a product cycle and started being an infrastructure cycle. And every SaaS board agenda needs to reset around that shift.

Here's what happened. In a single week — Anthropic closed a thirty billion dollar funding round at a nine hundred billion plus valuation, overtaking OpenAI by market cap. Anthropic also projected ten point nine billion dollars in second-quarter revenue and its first-ever operating profit. OpenAI began preparing an S one filing for an IPO. NextEra Energy announced a sixty-seven billion dollar acquisition of Dominion Energy — the largest US utility merger in history — driven explicitly by AI-driven grid demand. Pope Leo the Fourteenth issued the first papal encyclical on AI, co-presented with an Anthropic co-founder. And the White House scrapped a planned AI safety executive order after tech CEO pressure.

Six different events. One structural shift. Capital, regulation, and physical infrastructure are all reorganizing around AI as a base layer — not a product.

Here's why that matters for you. If AI is infrastructure, your strategy needs to reset around three questions your board has probably not asked yet.

First — where does AI sit in your financial model? Most SaaS P and Ls bury AI cost in SG and A or R and D. If AI is infrastructure, it belongs in your unit economics — next to hosting and bandwidth. Visible. Modeled. Defensible to investors.

Second — power availability is now a strategic constraint. The NextEra-Dominion merger exists because utilities cannot build capacity fast enough for AI workloads. Your largest model spend sits inside a capacity-rationed pipeline. Do you actually know which cloud regions your providers will defund or deprioritize first when power costs spike?

Third — public markets are about to see frontier-lab unit economics for the first time. An OpenAI S one will expose gross margins, infrastructure spend, and revenue concentration. Your vendor pricing assumptions for the next twenty-four months are about to be testable against real numbers. If your model assumes prices stay flat — you are about to be wrong.

One other beat worth your attention. The papal encyclical and the White House executive order scrap arrived in the same week, pulling AI governance in opposite directions. The Catholic Church is now part of the AI ethics conversation. The US executive branch has stepped back from regulating it. The companies waiting for regulatory clarity before setting an AI governance posture just lost their excuse. Clarity is not coming. Pick your posture proactively.

Connect the dots. The companies treating AI as infrastructure — same level of financial modeling as your cloud spend, same level of board attention as liquidity — are the ones that will not be caught flat-footed when the next capacity squeeze or vendor repricing arrives.

Your action item from this one. Get infrastructure on your next governance review as the frame. Not AI as a product strategy. AI as an infrastructure strategy.

Until next week — stay sharp. Just Keen A.I. dot com.