Artificial Intelligence

Nvidia and SK Group’s plan links a 2-gigawatt Vera Rubin AI factory with HBM4 supply, raising capacity, power, networking, and operations questions.

Fifty technology organizations are urging U.S. policymakers to preserve open-weight AI, a push that could shape model access, self-hosting and platform strategy.

Ray now treats Google Cloud TPUs as a first-class accelerator, giving platform teams supported paths for serving, data pipelines and distributed JAX training on GKE.

OpenAI is rolling out connected health records in ChatGPT. For platform teams, the real story is consent, retention, read-only access and the boundary between consumer AI and regulated clinical systems.

AMD’s ROCm.ai brings vendor-authored agent skills, a unified CLI and automated inference tuning to AMD platforms as Helios systems enter production.

Google’s latest Flash models put token efficiency, latency and specialized cyber work ahead of a simple benchmark race. Fresh Alphabet usage figures show why production economics now matter to engineering teams.

OpenAI Presence packages permissions, simulations, evaluations and controlled updates into a managed service for production voice and chat agents. Here is what platform and DevOps teams need to know.

OpenAI proposes measuring enterprise AI by the cost of successfully completed work, pushing platform teams to connect model spend with verification, retries, latency and human review.