
NVIDIA's next CEO needs to keep private AI practical
Todd Johnson on NVIDIA after Jensen Huang, the case for open-weight AI, private compute and sovereign control alongside cloud services.
Published · 2 minute read
Open models can support AI on private infrastructure alongside cloud services. AI-generated illustration.

Open models can support AI on private infrastructure alongside cloud services. AI-generated illustration.
Jensen Huang’s departure leaves me concerned about whether NVIDIA will keep investing in AI models people can run themselves.
Colette Kress is interim CEO while NVIDIA searches for a permanent successor. I’d like the next CEO to be as vocal about the need for open models as Huang was.

Jensen Huang. Photo: NVIDIA.
Huang publicly supported cloud services alongside open models that companies could use to build proprietary AI. From a US buyer’s perspective, that was a useful position from the leader of a major compute supplier. I see a commercial reason for NVIDIA to continue: privately operated AI still needs computing infrastructure.
Cloud investment and competition will remain important. Buyers also need the choice to keep sensitive information on infrastructure they control. NVIDIA’s Nemotron models support private deployment, and I want to see that option continue to improve.
The recent Microsoft 365 outage interrupted access to services including Copilot. Organisations building AI into important processes need a way to keep operating when their provider goes down.
For defence, sovereign compute and control over sensitive data should be requirements. Governments need people who can keep critical systems operating when an external service is unavailable. Sensitive businesses need to know who can access customer information and what their AI software sends outside the organisation.
Open weights give them the option to run models within their own systems. With the right access controls and software configuration, that gives businesses a practical way to keep sensitive information private.
I’d like NVIDIA’s next CEO to keep funding open models and the software needed to run them privately. The next releases and their licences will give buyers evidence of whether that support continues.

Written by Todd Johnson
Founder of InnovateX Solutions. Todd provides fractional CIO, CTO and Chief AI Officer leadership for Australian businesses and government, and wrote the XAF Connected Architecture framework.
Todd on LinkedIn