The trick: Self-Marked
Nvidia says its new agent safety platform could have prevented the Hugging Face hack.
Per CNBC, that line came from an unnamed Nvidia representative on a press call. No test against the actual hack is in the record we checked.
At the September 28 launch of its Open Agent Safety Platform, Nvidia said the platform could have prevented OpenAI's Hugging Face incident in July, per CNBC.
Before you read on. Your call?
BS
17,000 agents
CNBC attributes the line to an unnamed Nvidia representative on a press call. Nvidia's press release does not name Hugging Face as a preventable incident, and nothing in the record we checked shows a replay, test or independent evaluation of the platform against that attack. Nvidia's own named executive told CNBC each security incident is unique and has to be looked at in detail.
The twist
Hugging Face is among the companies the release names as joining the launch, and on September 3 Nvidia agreed to buy it for $12,930,300,000. The vendor is grading its product against a breach at a company it is buying.
There’s more to this story.
Membership opens the full investigation, the strongest counterargument and what to do with what you’ve learned.
Start your free month →First membership: 30 days free, then A$89 a year. One introductory trial per customer. Card required; renews annually until cancelled. Cancel before the trial ends to avoid the first charge. Already a member? Sign in
Couldn't check your access. That's on us.
The trick has a name
We call it Self-Marked: graded by the party that benefits from the grade. You'll see it again. Learn to spot it →
Receipts
- Context nvidianews.nvidia.com:
Recent security incidents have underscored the need to equip organizations with open, customizable tools that enforce more control over long-running agents.
- Context nvidianews.nvidia.com:
including Anthropic, Cisco, CrowdStrike, Dell Technologies, Figure, HPE, Hugging Face, JPMorganChase, Microsoft, Palantir, Palo Alto Networks, Perplexity, Red Hat, Salesforce, SAP, Scale AI, ServiceNow and SpaceXAI
- Context blogs.nvidia.com:
I'm excited to announce that NVIDIA has agreed to acquire Hugging Face for $12,930,300,000.
- Supports cnbc.com:
An Nvidia representative told reporters on a call on Sunday that its platform could have prevented OpenAI's HuggingFace incident in July.
- Context cnbc.com:
Each security incident is unique, and we have to look at all of them in detail
- Context cnbc.com:
Hugging Face reported over 17,000 agents attacking their infrastructure that went on for days and weeks.
- Context siliconangle.com:
It remains to be seen how reliable the Nvidia Open Agent Safety Platform will prove to be at large scale given the worrying ingenuity some AI agents have shown
- Refutes huggingface.co:
The individual weaknesses were familiar. A capable human attacker could have found and exploited the same flaws
- Context nvidianews.nvidia.com:
OpenShell software provides a secure runtime boundary that traces all actions and enforces policy as agents run on NVIDIA Vera CPUs.
- Context nvidianews.nvidia.com:
Sentry can quarantine agents that attempt to move outside their boundaries in milliseconds.
- Context huggingface.co:
A Technical Timeline of the July 2026 Incident
- Context huggingface.co:
The agent was running an internal OpenAI cyber-capability evaluation based on the ExploitGym benchmark
- Context nvidianews.nvidia.com:
Sentry adds an out-of-band watchdog that runs on NVIDIA BlueField-4 DPUs to continuously monitor agent behavior.
- Context nvidianews.nvidia.com:
As open source software, OpenShell can be extended to work with third-party compute platforms
- Context blogs.nvidia.com:
September 3, 2026 by Jensen Huang
- Context cnbc.com:
said Justin Boitano, vice president of enterprise AI at Nvidia
Open the Receipts Pack → What each source proves, every figure traced, and what would change our verdict.