Nvidia details AI security platform, buyback and insurance talks

NVIDIA said its board authorized an additional $150 billion for share repurchases, bringing the remaining authorization to $235 billion. The company expects to use that authorization through fiscal 2028.[1] NVIDIA described the increase as the largest share-repurchase authorization increase in history; The Hill noted that Apple’s $110 billion authorization in 2024 had been a record at the time.[1][2] NVIDIA said cash generation enables investment in technologies as well as returns of capital to shareholders.[1] The Hill reported the stock was up about 2.1% to nearly $230 by mid-afternoon on the announcement day.[2]

NVIDIA announced an open software platform and reference system design intended to strengthen AI security from agent testing through deployment.[3] It combines OpenShell software with the Sentry reference system design, and NVIDIA says it is designed to provide governance and control across agent software, computing hardware and robotics systems.[3] OpenShell sets runtime boundaries, enforces policies and traces agent actions; NVIDIA says it is broadly available and can control autonomous agents running across open and closed models.[3] The open-source software can also be extended to third-party computing platforms, including those from Arm and Intel.[3]

NVIDIA says Sentry runs on BlueField-4 data-processing units to monitor agent behavior continuously, and can quarantine and stop an agent within milliseconds if it attempts to leave its software boundary.[3] NVIDIA named Anthropic, Cisco, CrowdStrike, Microsoft, Salesforce, SAP and SpaceXAI among participants in the effort.[3] Salesforce and NVIDIA integrated OpenShell with Slack so teams can review agent activity, audit events, and approve or reject permission requests.[3] SAP is embedding OpenShell with the Joule Studio runtime, while SpaceXAI is using the platform for Cursor coding agents and Grok models.[3]

Bisnow reported that NVIDIA is working with insurance broker Howden Re on a possible insurance structure for neocloud lending.[4] Under the reported concept, insurers would compensate lenders if a neocloud defaults and its collateral GPUs cannot be resold for enough to repay the debt.[4] The talks are at an early stage and no agreements are in place.[4] Bisnow also reported that NVIDIA shared information with at least one insurer about chip depreciation and expected future computing-capacity values, and is exploring whether insurers could syndicate risk to hedge funds and other alternative investors.[4] It reported that NVIDIA has also considered joining potential financing consortia itself.[4]