One in Seven AI Deals Is Now a Compute Contract. One in Thirty Is an Acquisition.

Across 543 deals we tracked since October, acquisitions fell from 16.5% of AI dealmaking to 3.1% while compute contracts climbed from 3.9% to 14.6%. The buyers changed too: one in six recent compute deals has a government on one side.

Physical AI Is Where July’s Money Went. No Benchmark Measures It.

Japan’s trillion yen program says "physical AI", and July’s smallest rounds say the same thing. The tell is not the money, it is the partner lists. Here is how to read an applied AI deal, and the benchmark that does not exist yet.

Generating the Answer Got Cheap. Testing It Did Not.

CuspAI’s $450 million round did not lead with a model. It led with a foundry of more than 45 industrial members. When a wrong answer costs a furnace run instead of a retry, the scarce input stops being intelligence. Here is the one thing to check before you believe an AI-for-science claim.

The Hard Part Is No Longer Training the Model. It Is Serving It.

Moonshot shipped a competitive model and ran out of GPUs in 48 hours. The most interesting companies in AI right now sit between the chip and the model, and almost nobody is pricing them.