The limits on AI this week came from power lines, proof of identity and public trust, and money alone will not remove them.
The AI Transformation Brief
The limits on AI this week came from power lines, proof of identity and public trust. Microsoft dropped a data center deal at a British site that may not get grid power until the 2030s. Startup Fortune World launched tools that let banks confirm a real person, or a person's approved AI assistant, is on the other end. The Next Web Anthropic pledged 5% of its shares to a public fund after its chief executive called public doubt about AI a crisis of trust. TechCrunch Money alone will not clear any of these limits.
Microsoft has pulled out as the main customer of Nscale's planned data center in Loughton, Essex, where it was set to rent 23,000 Nvidia AI chips, the Daily Telegraph reported. Grid operators had reportedly told Nscale the site might not connect to the power grid until the 2030s. Startup Fortune Neither company gave a reason. Nscale, which has filed to list its shares in New York, said its relationship with Microsoft remains strong; Microsoft and Anthropic account for about 85% of its contracts. The News
The hard limit on AI is moving from chips to electricity, and a signed contract does not build a power line. A supplier whose order book rests on two customers and an unbuilt grid connection carries risk, however large that order book looks. Before signing a multiyear AI contract, ask where the work will run, whether that site has power today, and what happens to your price and service if the site slips by years.
World, the identity network co-founded by Sam Altman, launched a version of World ID for banks and financial firms at the Money20/20 conference on Sunday. It bundles an identity wallet, a Zoom tool that confirms a meeting participant is a real person and not a deepfake, and a tool that lets people pass verified facts about themselves to an AI assistant acting for them. More than 18 million people hold its proof of being human, and Deloitte estimates US losses from AI-enabled fraud could reach $40 billion by 2027. The Next Web
Banks have spent decades proving who a customer is. Now they also need to know whether a person or software is acting, and with whose permission. Whoever supplies that proof sits between the bank and its customers. Banks should decide this year which identity checks they will build, which they will buy, and which outside identity networks they will trust, before a single supplier sets the terms.
When an AI assistant acts for a customer, the bank still answers for the money. It needs proof of who approved the action, a record of it, and a named owner for disputes.
Anthropic said that, subject to shareholder approval, it will commit 5% of its shares to a new Anthropic Public Benefit Fund, and its board may add 2% for every 50% rise in the company's value. With earlier pledges from co-founders and employees, about 20% of its shares would go to the public good, and the fund's shares will not be used for lobbying or political giving. Anthropic Chief executive Dario Amodei called negative public sentiment toward AI "fundamentally a crisis of trust." TechCrunch
AI companies now treat public trust as a business cost. Anthropic has already committed $200 million to studying AI's effect on jobs, and the fund could make direct payments to affected groups. TechCrunch Companies buying AI face the same question at smaller size: if your AI use cuts roles or changes what customers pay, what will you put back, and how will you show it? Answer it before regulators or customers answer it for you.
Palmer's October 9 post argues that OpenAI's yearly revenue rate was near $50 billion at the end of September, about $20 billion below the widely reported figure, because of how sales through cloud partners are counted. He says comparisons between AI companies need consistent accounting. Shelly Palmer The same discipline applies to large supplier order books like Nscale's.
AI growth now depends on electricity, identity checks and public trust as much as on the technology.
Name one executive to review where your AI suppliers run, how customers prove who they are, and what you owe affected staff.
Data center builders without grid connections are losing anchor customers, and identity networks are moving between banks and their customers.
Suppliers that can show power, proof and public accountability will win the larger contracts.
Read AI supplier order books with skepticism: ask how much depends on one or two customers and on sites without power.
Budget for identity checks wherever AI assistants can act for customers.
Ask management three questions this quarter: which AI suppliers depend on sites without power, who can act in customer accounts, and what the company will give back if AI cuts roles.
The shift this week is that AI's limits are moving from the technology to the world around it: power lines, proof of identity and public consent. The assumption it breaks is that more money buys more AI on schedule. The decision it forces is which suppliers, identity checks and public commitments your company will rely on before they are tested.
If your main AI supplier lost its power connection tomorrow, how long would your business keep running?
Make AI accountable. Measure the results. Redesign the business.