According to the Wall Street Journal, the turnings of the dark satanic rumour mill was enough to get him back on the blower with OpenAI’s Sam Altman to rescue a stalled partnership.

Earlier this year, Jensen and his crew agreed to sink a staggering $100 billion into OpenAI, providing “millions” of AI chips and up to ten gigawatts of compute power.

Talks had reportedly gone cold during the summer. Then came whispers that OpenAI was cosying up to Google’s TPU chips through a fresh cloud deal. Not long after that story hit the press, Huang rang Altman directly, asking if the rumours were true and hinting that he was ready to get talks back on track.

“Huang realised that Nvidia could help OpenAI by making a direct investment in the company,” a person who spoke to him told the WSJ.

That quick pivot worked. The final agreement not only secured OpenAI’s reliance on Nvidia’s hardware for the foreseeable future but also locked out Google’s formidable ASICs from getting too comfortable in Altman’s data centres.

In effect, the massive investment acts as a classic supplier lock-in. OpenAI’s future systems, including the forthcoming Vera Rubin supercomputers, will be running entirely on Team Green’s silicon.

For Huang, the deal ticks two boxes: it guarantees OpenAI’s dependency on Nvidia gear and shuts the door on competitors trying to muscle in with their own AI accelerators.