Three weeks ago, Etched was a four-year-old chip startup that had just come out of stealth with $800 million raised.
The number that ended the debate about whether private markets could still fund artificial intelligence at scale is $122 billion.
In seventy-two hours at the end of June, two of the largest technology companies on earth committed roughly $3.5 billion.
Somewhere between the Prime Day promos and the warehouse robots, Amazon quietly became one of the three largest chipmakers.
Nine months ago, CuspAI was a $520 million bet on an idea: that a search engine for molecules could work.
For ten months, Huawei's Atlas 950 SuperPoD existed only as a slide deck and a promise. On July 16, it appeared.
Every leaderboard has a champion. Almost none of them come with an expiration date printed on the box.
The most consequential agentic AI rollout of the year is not in a coding IDE or a customer service queue.
For two years, the pitch for diffusion language models has been the same: they can generate tokens in any order.
Every image a diffusion model generates is the end of a countdown. Twenty steps, fifty steps, a thousand steps.
The most credible AI-in-pharma number of the week did not come from a discovery claim. It came from the loading dock.
A protein structure predicted by AlphaFold is a photograph. Biology is a film. The gap between the two matters.
For a decade of antitrust fights, the thing regulators tried to pry out of Google's grip was the browser.
American law has never quite decided what you own about yourself. Your name and image are protected in some states.
There is no New Jersey AI statute for investment advisers. There is no rule proposal, no notice-and-comment docket.