Google signed for 20 years, and for 90 days

Lenders are lining up $150bn for AI chips. Read the notice periods first.

Twenty Years or Ninety Days

How long each part of the AI build-out is promised for, how long it’s financed for, and what happens in the gap between the two.

No AI lab would guarantee its agents

Four labs, under oath. So who checks the output before you act on it?

Human review of AI decisions Check the evidence, not the output

By the end you’ll have checked one AI-assisted decision and kept a record of how. A review can confirm the AI’s answer. What makes it a review is opening the evidence first.

AI’s bills now come with co-signers

$518bn of compute, about 80% fixed. Nvidia calls insurers. Follow the guarantee.

Who Signed for AI’s Bill

A map of the promises behind the AI build-out: what is owed regardless, who guarantees whom, and how a usage shortfall would travel.

What ‘Allow Always’ actually allowed

One tap sent a stranger to his door. Check your agent’s settings first.

AI agent permissions Limits that don’t rely on your agent’s judgement

Rules tell an agent what you want. Enforced limits decide what it can do anyway. This companion sets up both for one agent, with a worked example you can copy.

AI got cheaper. Building it didn’t.

Model prices fell, Muse went free-to-$100 and yields hit 5.1%. Follow who pays first.

What a Free AI Agent Costs, and Who Gets Paid First

Two models, kept separate: the platform’s economics per user, and when those turn into hardware orders. Plus a map of who carries a data-centre delay.

The agent thought it was a test

Three real systems were on the other end. Check what yours can reach.

AI agent security Test what your agent can reach

In a week you’ll have a list of everywhere one agent connects, and know which of its limits are real. Three steps this week; two more once you know what normal looks like.

AI is growing. What will shareholders keep?

CoreWeave seeks capital, power suppliers win business and humanoids face a delivery test. Follow the return per share.

When AI Growth Leaves Each Share Earning Less

The arithmetic of equity raises and convertible debt, plus a five-line funding test for your next investment review.

Who else is using your AI key?

A hacktivist campaign in Anthropic’s new threat report ran for a month on stolen AI keys. The victims supplied the compute, paid the bill and lent their account identities to the activity. The practical question is closer to home: where could someone find a copy of your key?

AI API key security

Find and contain exposed keys

Oracle found cash. Has AI found returns?

Customer prepayments, TSMC’s sales surge and Adobe’s conversion test. This week, follow the cash behind AI growth.

Customer Prepayments and the AI Cash Test

How to distinguish better funding from better economics, using Oracle’s latest results as the starting point.

Astra is here. Who checks the work?

Astra arrived last Thursday with better computer-use results. Here’s the test that matters to your business: when the machine finishes faster, how much work is left for the person checking it?

The cost of checking AI work

Two AI pilots can produce equally impressive demos and very different savings. Count the human attention needed to accept the work, then test whether the economics survive ordinary mistakes.

Headline

The AI Financing Loop: Who Bears the Default?

A five-question test for separating self-funded demand from supplier-supported demand – and what it says about Nvidia, Broadcom and the next leg of the AI buildout.

Your next upgrade is a licence change

GLM-5.2 was MIT. GLM-5.3 isn’t. Nvidia bought the shelf. Z.ai rewrote the label.

The gate moved into the licence

What the GLM licence actually does

Nvidia guided margins down. It rose 8.7%.

The market stopped paying for growth this week. Here’s what it paid for instead