Mark Zuckerberg, Microsoft CEO just made major AI decision

Jensen Huang has been one of the most prominent voices in technology for years, but he had never once posted on X (the former Twitter).

On July 24, he did. What he chose to say with his first-ever post on the platform says a lot about the fight now underway over who controls the future of AI.

Huang backed a letter titled “Open Weights and American AI Leadership,” signed by 25 organizations including Nvidia, Microsoft, and Meta, calling on U.S. policymakers to support open-weight AI models rather than restrict them.

Satya Nadella and Mark Zuckerberg both added their own statements on X. Elon Musk amplified the message. And two of the most visible companies in frontier AI, OpenAI and Anthropic, were not on the list.

What Zuckerberg, Nadella, Musk, and Huang actually said about open AI models

Nadella offered a clear policy argument. “Together with others across our industry, we are outlining a path for open-weight models to strengthen American competitiveness and expand economic opportunity, while protecting national security,” he posted on X.

Zuckerberg was more direct. “Open source is a positive and important force for both empowering people and preventing centralization. Proud to support this,” he posted on X. Musk reshared Zuckerberg’s post and added three words: “Overwhelming support for open source.”

Huang’s first X post made the same argument. Open models, he wrote, “strengthen safety and cybersecurity, accelerate innovation and diffusion, and enable sovereignty.” For a CEO who has never used the platform publicly before, choosing this particular moment to start says something about how seriously the industry is treating this debate.

More AI:

To understand why this fight matters, it helps to know what open-weight actually means. When a company releases a model as open-weight, it publishes the underlying numerical parameters: the billions of values that define how the model thinks and responds.

Anyone can download those parameters, run the model on their own hardware, and modify it for specific tasks without paying API fees or asking permission.

A business can deploy the model inside its own data center and keep its data off third-party servers entirely. A foreign government can take that model and run it without depending on any American company at all.

That last part is where the geopolitical argument gets interesting, and where the two sides in this debate start talking past each other.

Why OpenAI and Anthropic staying off the letter is the real story

OpenAI and Anthropic weren’t just absent. Both companies have reportedly been pushing U.S. policymakers in the opposite direction, advocating for tighter restrictions on Chinese open-source AI and arguing those models are easier to misuse once released publicly. The commercial interest behind that position is pretty clear.

OpenAI and Anthropic charge for API access to models they control. Developers who can run equally capable models on their own infrastructure don’t need to pay those fees. The open-weight push doesn’t just put these two companies on the wrong side of a policy letter. It threatens a core part of how they make money.

Critics of the letter are quick to point this out. They describe the industry push as lobbying dressed up as public interest. The argument that open models are harder to control once widely released is a legitimate one, and it’s the issue OpenAI and Anthropic are raising with policymakers.

The letter from Zuckerberg, Nadella, and 23 other organizations is a direct attempt to influence those decisions before they become law.

Sven/Getty Images

What China’s Kimi K3 launch has to do with the timing

There’s a specific reason the letter came when it did. On July 16, Beijing-based Moonshot AI put out an open-weight model called Kimi K3. It reportedly hit near-frontier performance levels. Chip stocks slid on the news.

The release made something concrete that had been theoretical for months: Chinese labs can now build open-weight models that are genuinely competitive with what U.S. labs produce, and release them publicly for anyone to download.

White House adviser Michael Kratsios accused Moonshot of using distillation to copy a U.S. model in building Kimi K3, adding another layer of tension to the debate, Fortune reported.

For the tech leaders backing the letter, that development strengthens their case rather than weakening it. Their argument is that if China can produce and globally distribute competitive open-weight systems, the U.S. response should be to accelerate its own open ecosystem, not restrict it.

Locking down open models in the U.S. would slow American innovation while doing nothing to stop Chinese models from spreading.

Why the policy fight over open AI models is just getting started

The policy stakes are real. Policymakers are currently weighing whether to restrict how openly AI models can be released, who can train on large amounts of compute, and how much access foreign entities should have to frontier systems.

The letter from Zuckerberg, Nadella, and 23 other organizations is a direct attempt to influence those decisions before they become law. It argues that regulators should target specific harms and misuse cases rather than restricting the model format itself.

The tech leaders signing it are betting that an open-weight ecosystem ultimately serves American interests better than a controlled one does.

The ones who didn’t sign are making the opposite bet. That divide is now public, and it will play out in committee rooms and policy briefings over the months ahead.

Jensen Huang waited a long time to make his first X post. The fact that he used it on this issue is the clearest signal of how seriously the biggest names in the industry are taking this fight.