🚀 OpenAI picks AWS in $38B cloud deal
Seven-year commitment, hundreds of thousands of Nvidia GPUs, and a clear signal that OpenAI wants multi-cloud flexibility-Microsoft’s grip loosens, a bit.
Amazon gets bragging rights and likely a hefty utilization bump for AWS. Wildly, OpenAI keeps talking about truly colossal compute plans… which, honestly, read like sci-fi translated into a spreadsheet.
🔗 Read more
🤝 Microsoft inks $9.7B compute pact with IREN
Five years of access to Nvidia GB300 gear via IREN-a clever way to add capacity without pouring more concrete.
There’s a parallel Dell supply deal, talk of liquid cooling, and a Texas campus with serious megawatts. Efficient, if slightly baroque.
🔗 Read more
🌍 Microsoft wins green light to ship Nvidia chips to UAE
Export licenses secured, plus a pledge to invest over $15 billion in UAE infrastructure. The geopolitics here are… delicate, but the chips are moving.
This slots neatly into Microsoft’s wider Middle East compute push-the timing feels very not accidental.
🔗 Read more
🏗️ Lambda lands multibillion AI infra deal with Microsoft
Nvidia-backed Lambda says it will deploy tens of thousands of GPUs for Azure-including those hulking NVL72 systems.
No sticker price disclosed, but yeah, it’s big. It also neatly piggybacks on Microsoft’s other capacity grabs.
🔗 Read more
🎬 Studio Ghibli and friends tell OpenAI to stop training on their work
Japan’s CODA-repping Studio Ghibli, Bandai Namco, and others-fired off a letter asking OpenAI to stop training Sora 2 on member content.
Creators are drawing a louder line in the sand. Fair-use debates just got a fresh splash of anime ink.
🔗 Read more