Agentic inference capacity grew 4.6× in a year, versus 2.6× for compute capacity
Why it mattersIf you buy AI agent services, capacity is outgrowing raw computing power, which gives you grounds to push providers on price.
Why it mattersIf you buy AI agent services, capacity is outgrowing raw computing power, which gives you grounds to push providers on price.
Why it mattersOpenAI now has an emergency stop across its research work, a control you can ask any AI supplier about.
What happenedOpenAI built Pando, a backend that works with Git and is designed to handle the volume of code development that its Codex tool produces.
Why it mattersCompanies planning large data centres now have an OpenAI template for working with utilities to protect local power supply and prices.
Why it mattersBusinesses in Iowa and Wisconsin get 2.6 GW of added power capacity and a projected $1 billion in savings on electricity costs.
Why it mattersOpenAI says ChatGPT reads got 12× faster on Azure Cosmos DB with global caching, a design you can copy for high-traffic products.
Why it mattersFederal money is heading to 2.5 GW of new grid capacity, and it is aimed at power costs.
Why it mattersChatGPT's storage layer now handles 22M requests per second, the scale bar your own infrastructure gets compared against.
Why it mattersYour model host is now a chip vendor
Why it mattersGPU utilization up 33 points cuts compute cost per workload without new hardware.