G42 is putting Abu Dhabi's AI ambitions behind Nvidia's Open Secure AI Alliance at a moment when open-weight models are being pitched as both a security tool and a sovereignty choice.

The National reported that the Abu Dhabi-based AI company joined the Nvidia-led group on Monday.

The alliance gives G42 a place in a campaign that links open AI access with cyber defence.

Nvidia formed the group after chief executive Jensen Huang used his first X post on Friday to support open-source software and open-weight AI models.

The National reported that Huang's earlier industry letter carried 25 signatures from companies such as Microsoft, Palantir, Perplexity, CrowdStrike and IBM.

Nvidia Alliance Brings G42 Into An Open-Model Security Push

The Open Secure AI Alliance is built around a distinction that matters for companies and governments deciding where sensitive AI systems should run.

Open-source models disclose software and development processes and allow modification and reproduction, while open-weight models can be downloaded, run and adapted even when the underlying code and development process remain proprietary.

G42 chief executive Peng Xiao posted on X that frontier open models should sit beside closed systems, with security, trust and sovereign control treated as design requirements from the start.

The UAE company's role therefore sits inside a debate over whether nations and enterprises should rely only on closed systems for high-risk AI workloads.

The membership list now includes at least 37 other companies, The National reported.

The named participants include IBM, SpaceX, Hugging Face, Doordash, Cloudflare, Salesforce, Adobe and Cisco.

Hugging Face Incident Frames The Security Argument

Nvidia's public case for the alliance points to a disputed security episode involving Hugging Face and OpenAI models.

In Nvidia's statement, the company linked the incident to a need for open frontier agentic systems that cyber defenders can operate on infrastructure they control.

The same statement said Hugging Face used the open-weight GLM 5.2 model on its own infrastructure to analyse more than 17,000 actions and contain the intrusion after closed AI tools blocked forensic analysis.

OpenAI has called the incident unprecedented, and the episode remains controversial and under scrutiny.

The security case gives the alliance a narrower purpose than a simple endorsement of open AI.

The debate over proprietary, open-source and open-weight systems also includes recent rumours that US regulators may seek restrictions on certain open AI models.

Sovereign Control Is The UAE Angle

For G42, the source-backed operating question is not only whether open models are more transparent.

Peng's comments put sovereign control beside security and trust, which matters for an Abu Dhabi AI company selling into governments and enterprises that may want to keep model operations, forensic review and sensitive workloads closer to their own infrastructure.

Nvidia's position is narrower than an open-model-only claim.

The company framed closed and open frontier models as complementary tools, with transparency, adaptation and sovereign control available where security demands them.

The alliance record does not identify G42-specific deployments, UAE government customers or a timetable for turning membership into a security product.