In December 2023, Chinese executive Alice Huang met Nvidia chief executive Jensen Huang and colleagues at an Italian restaurant in Singapore, according to an account that places the encounter at the centre of growing scrutiny over access to advanced AI hardware. The report says Huang, backed by Shanghai investors, had restructured a company that would buy billions of dollars in Nvidia AI chips.

The account alleges that the company, Megaspeed, bypassed US restrictions on Chinese purchases of advanced chips by operating through a Singapore shell company. Four days after its ownership change, Megaspeed was named an Nvidia cloud-service partner, giving it access to sought-after components. The episode is framed not simply as a corporate relationship, but as a window into the high-stakes US-China contest to dominate AI — and the personal vulnerabilities that can shape it.¹

The corporate case for AI looks markedly different in Disney’s newly announced multiyear partnership with Adobe. Rather than treating generative models as a back-channel route to scarce computing power, the companies say they will build proprietary tools around Disney’s own intellectual property, from theme parks and cruise lines to studio workflows and ESPN graphics.

Adobe says Disney will get early access to its technology and work with Firefly Foundry on models described as “human-centered and artist-led.” The promise is commercially safe, production-ready output based on Disney’s characters, settings and visual styles — a pitch designed to put guardrails around a technology still provoking fierce arguments over creative labour and control. The announcement also comes as Disney explores broader efficiency measures, with reports of planned job cuts sharpening the tension between AI-enabled productivity and its workforce consequences.²

Together, the developments show AI’s split-screen reality: a geopolitical scramble over the chips that power it, and a corporate effort to make its creative uses appear managed, proprietary and safe.