The Open Weights and American AI Leadership letter underwent a massive shift immediately after its launch. While it began as a specialized group of 25 companies, it surpassed 230 signatories as major tech giants, infrastructure companies, and AI labs rushed to join the coalition The Blockbuster Additions ddition of the traditional closed-frontier model leaders
=== Major absentees are apple, amazon, anthropic - Anthropic's Amodei explains why he just doesnt get open intelligence here
Box [, 2, 3, 4]
Jensen Huang’s approach is a masterclass in regional framing, and he is not directly calling on other countries to replicate the American open-weights lobby. [1, 2]
Instead, he uses two distinct, parallel narratives customized for different markets to protect Nvidia’s bottom line. [1, 2]
┌────────────────────────────────────────┐
│ JENSEN HUANG'S DUAL STRATEGY │
└────────────────────┬───────────────────┘
│
┌────────────────────────────────┴────────────────────────────────┐
▼ ▼
[ DOMESTIC POLICY (U.S.) ] [ GLOBAL COMMERCIAL (SOVEREIGN AI) ]
• Frame: "American AI Leadership" • Frame: "Sovereign Self-Reliance"
• Goal: Keep model weights open • Goal: Build localized "AI Factories"
• Core Argument: Open weights prevent centralizing • Core Argument: Control your own data, culture,
power in just 2-3 closed labs. and intelligence on local hardware.
1. The Global Strategy: "Sovereign AI"
When dealing with individual countries (like South Korea, India, Japan, Canada, or nations across Europe), Huang rarely uses the phrase "open-weights lobby". Instead, he targets them through his Sovereign AI framework. [1, 2, 3, 4, 5]
His pitch to each nation focuses on their most distinctive regional contributions:
- Refining National Data: He explicitly tells world leaders that their country’s data, language, history, and culture are a "natural resource". [1, 2]
- The "AI Factory" Blueprint: Instead of relying on a closed American cloud provider (like OpenAI or Anthropic), he convinces countries to build their own local Nvidia-powered "AI Factories" to process that localized data themselves. [1, 2, 3]
2. Why He Doesn’t Call for Global Open-Weights Networks
Huang's recent push for open-weights is a hyper-targeted defensive maneuver explicitly for Washington lawmakers, rather than a global philosophical crusade. [1, 2, 3]
- Avoiding Geopolitical Landmines: If Huang explicitly encouraged foreign countries to build unified "open-weight networks," it could easily clash with U.S. export controls and national security directives. For example, a global open-weights coalition would inherently benefit Chinese firms like Moonshot AI (creators of the Kimi K3 open model). [1, 2, 3]
- A Subtle Commercial Reality: Open-weight models require heavy, customized physical hardware to run locally. By quietly letting open-weights proliferate globally under the guise of "Sovereign AI," Huang ensures that every country, startup, and localized cloud provider has to buy massive amounts of Nvidia chips to host those open models, all without Nvidia ever having to take a controversial, direct political stance abroad. [1, 2, 3, 4]
Would you like to look closer at an example of this strategy in action, like Nvidia's recent 200MW "AI Factory" deal with South Korea's NAVER, or see how European leaders are reacting to his Sovereign AI push? [1, 2]
Show all
Box [, 2, 3, 4]
