Cloud provider stories
Utilities and cloud operators could squeeze more AI work from fixed power budgets as Nvidia's DSX software lifts throughput and cuts demand.
Electricity is becoming the main bottleneck for AI data centres as new Nvidia partnerships aim to raise output per megawatt.
Teams running AI agents at scale could save idle compute and tighten isolation as Google opens Agent Substrate on Kubernetes Engine.
The arrangement gives the Australian neocloud a single control layer to manage AI infrastructure and support up to 150,000 GPUs.
Data engineers may cut weeks of workflow setup as the open-source kit lets Airflow pipelines be built and monitored from an IDE.
Cloud builders could gain more compute headroom, as NVIDIA shifts memory control off the chip and opens NVHBM to multiple suppliers.
Researchers can claim up to USD $1 million for breaking Vercel Sandbox's isolation, as the cloud provider opens its boundary to public scrutiny.
The startup gains faster access to scarce AI compute as it prepares to scale model training across both TPU and Nvidia systems.
Thousands of cloud customers faced a potential cross-tenant database breach, though Microsoft says the Cosmos DB flaw is now fixed with no sign of abuse.
Australian firms seeking local cloud hosting will gain more options as Hyve opens live data centre capacity in Sydney and Melbourne.
European firms needing secure, low-carbon compute gain a bigger home as the deal expands cloud, AI and dedicated HPC services under one roof.
European firms can now keep sensitive data local while using cloud-style storage for AI and disaster recovery under the new platform.
The additions broaden access to mixed quantum systems as OVHcloud seeks to build Europe's quantum infrastructure around its sovereign cloud.
The hire comes as cloud providers jostle for business from customers weighing AI workloads, sovereignty and compliance in Europe.
Routine HR and finance tasks will move into Gemini Enterprise, as employees gain self-service access to Workday data without leaving Google Cloud's AI tool.
Reserved capacity for new AI workloads is now open, as Vultr prepares deployments across 2027 and 2028 with AMD's latest chips.
Regulation, resilience and AI are now shaping Chinese companies' cloud choices as hybrid models dominate and vendor risk rises.
Surrey's Longcross campus will gain more capacity for AI workloads as Ark commits GBP £807 million to meet Nebius's growing UK demand.
UK partners gain a three-tier route to Leaseweb services as the cloud provider steps up support for hybrid and AI workloads.
Tighter Indian audit rules are pushing banks and public bodies to watch database activity more closely, especially inside sensitive systems.
