Daily Briefing — June 3, 2026

Google's parent company Alphabet just announced plans to raise up to $80 billion in equity to fund its AI infrastructure push. That includes a $10 billion share sale directly to Berkshire Hathaway, which is a notable vote of confidence from one of the most conservative institutional investors around. Analysts are calling it the largest equity fundraising in history, bigger than the three largest IPOs ever combined. The stock dropped over 4% when markets opened, which tells you investors have some feelings about it.

A raise this size could look like distress, but it isn't. Alphabet is so convinced the AI infrastructure race requires massive capital right now that it's willing to dilute existing shareholders to fund it. When a company at Alphabet's scale does that, it signals something about where the whole industry is headed. The costs of staying competitive in AI are enormous, and the gap between companies that can absorb those costs and those that cannot is about to get a lot wider.

For anyone working in tech or finance, this is a flashing signal about where budget priorities are going inside large organisations. AI compute and infrastructure spending is not slowing down. It is accelerating, and someone has to pay for it.

Microsoft just held Build 2026 and, predictably, the AI announcements kept coming. Satya Nadella and crew covered everything from new Surface hardware to an always-on personal assistant to updates on Microsoft's own AI models. It was a lot. A few things stood out as actual signal rather than keynote theatre.

The Surface RTX Spark Dev Box is the most interesting piece of hardware for anyone working in or adjacent to development. It is built around Nvidia's new Arm-based Spark RTX chip, packs 128GB of unified memory, and comes preloaded with Visual Studio Code and GitHub Copilot. The pitch is simple: run local AI models on your own machine without depending on cloud infrastructure. That is a real shift in how developers can work, and it matters beyond just the device itself.

The software story is arguably more consequential for more people. Microsoft is adding Linux-like command line utilities that run natively on Windows 11, expanding what you can do through Windows Subsystem for Linux, and shipping a new Intelligent Terminal that connects directly to AI agents. Basically, Microsoft is collapsing the gap between Windows and the Linux-first developer workflow that most serious engineers have preferred for years.

Trump signed an executive order this week creating a voluntary framework for AI companies to share their frontier models with the federal government before public release. The idea is to let federal agencies assess the cybersecurity risks of advanced AI before it hits the market. Key word there: voluntary. Companies can share or not share. They just get confidentiality protections if they do.

This order is a watered down version of something that almost happened. An earlier draft would have required companies to share models 14 to 90 days ahead of release. That version got shelved because Trump was worried it would slow down the US race against China. What we got instead is essentially a handshake agreement with no teeth. The government is asking nicely.

That tells you something important about where AI regulation actually stands right now. The administration wants to look like it is taking safety seriously without actually slowing anyone down. Whether that is responsible governance or a liability waiting to happen depends on who you ask. What it definitely means is that the companies building these systems, and the people working inside them, are still largely setting their own rules.