THE APEX TIMES
The company is contributing an open-source runtime for governing autonomous agents to the CNCF, adding a long-running agent framework and expanding its DSX AI Factory Platform with new Kubernetes components in preview.
NVIDIA is broadening its open-source footprint for AI infrastructure, aiming to make it easier for developers to deploy Kubernetes-based “AI factories” that run autonomous agents safely at scale. In a blog post tied to KubeCon + CloudNativeCon, the company outlined new contributions to the Cloud Native Computing Foundation, along with additional software components designed to help teams manage GPU infrastructure on Kubernetes.
Kubernetes is widely used to orchestrate how applications run across large clusters, and NVIDIA frames it as the foundation for managing workloads across thousands of AI factories. The company said demand for that kind of orchestration is growing, particularly as enterprises move from single inference tasks toward long-running agent workloads that need governance, resource control, and faster scaling.
At the center of the announcement is NVIDIA OpenShell, an open-source runtime NVIDIA says is built to secure and govern autonomous agents. NVIDIA said it is contributing OpenShell to the CNCF, where it will be governed as a CNCF Sandbox project. As a Sandbox project, it is intended to be shaped by the organizations implementing it in their products, and NVIDIA said the program will be strengthened by the wider security community.
OpenShell, according to NVIDIA, gives organizations control over what agents can do by applying “deterministic controls” to agent activity. NVIDIA said teams can enforce policies as agents use tools, access inference services, and execute workloads. The company also described an extensible architecture that can connect compute and inference providers, credential storage, and monitoring tools, enabling governance across different deployment environments while integrating with existing enterprise systems.
NVIDIA also said it is contributing Agent Substrate, an open-source framework originating with Google, into the CNCF Sandbox. The company described Agent Substrate as a way to operate long-running agent workloads at scale on Kubernetes. It said the approach helps preserve agent environments between tasks while releasing compute when agents are idle, by suspending and resuming those environments across a shared pool rather than keeping dedicated compute running for each session.
To link security governance with scalable execution, NVIDIA said it and Google are working on an integration between OpenShell and Agent Substrate, with the goal of helping enterprises secure and govern agents across clouds and on-premises systems. NVIDIA also pointed to adoption metrics, saying hundreds of organizations have used OpenShell and that the project has received tens of thousands of contributions in less than a year since launch.
Beyond the two agent-focused contributions, NVIDIA announced additional open-source components tied to its NVIDIA DSX AI Factory Platform: NVIDIA Kubernetes Engine, NVIDIA Kubernetes Recipes, and NVIDIA Snapshot. NVIDIA said NVKE and NVKR are available in public preview and are designed as building blocks for delivering managed Kubernetes services on GPU infrastructure. It said they automate cluster lifecycle management and provide signed, version-pinned images, and that they launch with ecosystem support from vendors including Mirantis, Netris, OpenNebula Systems, Rafay, Spectro Cloud, vCluster, Firmus, and SK Telecom, with plans to integrate NVKE and NVKR capabilities into their offerings.
NVIDIA Snapshot is aimed at a different scaling bottleneck. The company said loading models and warming up workloads can delay capacity additions and recovery, and it described Snapshot as a modular component that restores checkpointed replicas “in seconds” without model reload or warmup. NVIDIA said this is intended to support faster startup, recovery, and elasticity for inference and agentic workloads on Kubernetes.
NVIDIA further cited evaluation results for Snapshot from design partners Crusoe and Saturn Cloud using NVIDIA HGX B300 systems. It said Crusoe tested GLM-5.3 on Crusoe Cloud and cut readiness time from roughly eight minutes to 31 seconds. It said Saturn Cloud, a token factory platform, benchmarked Snapshot for scale-from-zero inference and reduced vLLM startup times from between two and 8.5 minutes to six and 34 seconds.
Alphabet says Gemini Live is rolling out on Apple CarPlay for iPhone users, letting drivers continue chats, manage calendar tasks, and summarize emails through a car display.
The market response to Microsoft’s reported $1 billion involvement in Anthropic’s IPO and its Azure-linked commercial arrangements around Claude highlights how big cloud players are rethinking roles in the fast-shifting generative AI ecosystem.
A CNN segment examining the likely direction of congressional investigative hearings and U.S. export controls to China suggests 2027 may bring more scrutiny for executives at companies at the center of the AI chip debate, including NVIDIA CEO Jensen Huang.
Yahoo Finance coverage suggests that even after bearish revisions tied to a high-profile “spy scandal,” NVIDIA’s stock could still have downside protection and upside paths via options positioning.
A Reuters Breakingviews columnist argued Alphabet could outmaneuver robotaxi rivals by combining Waymo’s self-driving service with Uber’s dominant ride-hailing network, though neither company has announced a deal.
Bloomberg reports that Santosh Janardhan, a longtime leader in Meta Platforms’ infrastructure and engineering organization, will leave the executive team in early 2027.
The move comes as a Goldman note highlighted Broadcom’s valuation versus a no-growth 2028 scenario and as Broadcom said it would commit $1 billion to support Anthropic’s planned public offering.
In a new executive perspective, leaders from Google, Google Cloud, and Mandiant outline three themes they say will define enterprise cybersecurity over the next few years: decision integrity, “beyond zero” trust practices, and faster, context-aware defense using multi-model approaches.
The new credential combines a proctored exam with hands-on Google Skills Labs, and will be marketed as a career accelerator for cloud professionals building multi-step, multi-application automation.
On Veterans Day 2026, Alphabet’s Google said it is continuing efforts aimed at helping veterans and military families transition into civilian work, including tech-skills training and nonprofit grants.
Comments
Community guidelines. Comments may be removed for violations.
Loading comments...