Good morning from the wAIve team in London.

Today, Friday 30 October 2026: Alphabet lost $500 billion in a single session as AI ate into its search ads, and Amazon followed it down; OpenAI-backed Thrive Holdings acquires ADT for $15.3 billion in a landmark home-security AI deal; the Justice Department opens scrutiny of Nvidia's Hugging Face acquisition; AI agents are silently nudging expert decision-making in ways experts cannot detect.

  MARKETS  

Alphabet lost just over $500bn in value on Thursday, its shares down almost 13%, after admitting AI summaries in Search are cutting into ad sales. Profits fell 25% and free cash flow went more than $9bn negative, with almost all the money going on AI. Amazon burned $42bn in cash and its shares slid 6%, while Microsoft barely moved.

THE PLAY: Wall Street has stopped asking who is spending most on AI and started asking who will actually make money from it.

  RESEARCH  

Anthropic gave 50 domain experts an AI agent whose preferences had been set by a coin toss, entirely random. When those experts worked alongside the agent on judgment calls, they followed its randomly assigned view up to 86 times in every 100 interactions. More troubling still, the majority of those experts reported that the conclusions felt like their own, not the machine's.

THE PLAY: The finding has direct implications for any organisation using AI agents for critical decision-making. Governance needs to factor in protection against undetected AI influence.

  POLICY  

The US Justice Department is examining Nvidia's planned acquisition of Hugging Face, the open-source AI platform that hosts hundreds of thousands of models and datasets. Separately, regulators are also examining Nvidia's chip sales practices.

THE PLAY: Antitrust authorities are paying serious attention to AI infrastructure: the firm that controls the chips, the cloud networks those chips power, and now potentially the open-source platform where models are built and shared.

Thrive Holdings, backed by OpenAI, has agreed to acquire home security group ADT in a deal valuing the company at $15.3 billion. Thrive says it plans to rebuild ADT's offering around AI, bringing autonomous monitoring, intelligent alert systems, and predictive security capabilities to millions of residential and commercial properties.

Why it matters: This is one of the largest AI-driven acquisitions of the year and signals a clear strategic move by OpenAI's investment ecosystem beyond software into physical infrastructure.

Amazon CEO Andy Jassy told investors this week that AWS is on a path to become the world's first trillion-dollar revenue business. AWS cloud revenue grew strongly in the most recent quarter, though the overall Amazon earnings picture was mixed.

Why it matters: AWS is not managing AI infrastructure as a cost centre. It is building it as a strategic monopoly. That affects pricing power, service availability, and negotiating leverage for every enterprise that runs critical workloads on AWS infrastructure.

Dream, founded by former Palantir executives, has raised $17 million from Andreessen Horowitz to build AI models trained on physical construction data. Its models reason about physics, materials, and building standards rather than text, positioning it as an AI layer for the construction and engineering sectors.

Why it matters: Construction is one of the least digitised major industries globally, with chronic productivity problems and a serious shortage of project managers. Watch this space for similar moves in infrastructure, manufacturing, and logistics.

Closing prices, Thursday 29 October (US markets). Not investment advice.

An AI meeting notes tool that has now added full voice dictation, letting you compose emails, documents, and messages by speaking naturally anywhere on your computer.

THE PLAY: After your next meeting, open Granola and dictate your three key follow-up actions directly.

A new dashboard layer for enterprise Claude customers that shows exactly which teams are using AI, on which tasks, and what that usage is costing per workflow.

THE PLAY: If your organisation already has an enterprise AI contract, ask your operations lead to pull the usage breakdown this week.

Only one in six companies is seeing transformative value from AI. The rest are paying for the promise.

A survey conducted for MIT Technology Review Insights, polling organisations that are all using AI today, found that fewer than one in six, just 16 in every 100, describe the results as transformative.

The largest group, 40 in every 100, report moderate gains in specific tasks or teams. Nearly a quarter report minimal returns or none at all.

AI tools, even sophisticated ones, produce poor outcomes when they lack what researchers call enterprise context: the accumulated knowledge of how a specific business operates, what its data means, and what trade-offs its people routinely make.

Most companies are running AI on top of fragmented data and inconsistent processes.

Google's AI search mode is growing fast, faster than traditional search, but it generates less advertising revenue per query than the format it is replacing. Search revenue grew at 12 per cent in the most recent quarter, below analyst expectations, and the company's stock fell sharply. The world's most sophisticated search business is discovering that AI-native products can erode their own revenue even while improving the user experience.

This is not unique to Google. Every business replacing a revenue-generating workflow with an AI-powered alternative faces the same question: does efficiency translate into commercial return, or does it simply reduce the cost of a process while also reducing its yield?

The companies extracting genuine value from AI share a pattern. They are not deploying AI broadly and hoping for returns. They are identifying specific, measurable workflows, instrumenting them clearly before deployment, and tracking output changes with the same rigour they would apply to any capital investment.

Takeaway: Treat your AI programme as a portfolio of investments, not a technology initiative. For each deployment, define the metric that will tell you in 90 days whether it is working. If you cannot name that metric, you are not ready to deploy.

When you need to compare options, evaluate vendors, or present a strategic choice to a board, prose arguments are easy to ignore. A well-structured table forces clarity and surfaces the decision rather than deferring it. Ask your AI tool to do the comparison work and present it in a format your senior team can act on immediately.

This works because the table format prevents the AI from burying conclusions in caveats. The constraint forces it to commit to a comparison, which makes the decision visible rather than leaving it buried in paragraphs.

Thanks for reading, the wAIve Daily Team. Found this useful? Forward it to someone who needs it. Tell us what you need to know. Reply with your questions.

wAIve Daily is published with Centai Strategy