Skip to main content

18 US companies jump on Kimi K3 after open-source launch

18 US Companies Jump on Kimi K3 After Open-Source Launch

Last night, Kimi K3 officially went open-source on Hugging Face. And almost immediately, a rather ironic scene unfolded: while the US government and companies like Anthropic spend big money lobbying to block Chinese open-source models, at least 18 US companies quickly deployed K3 and started offering commercial services.

A list compiled by a netizen named Ding exposed the growing rift between Washington's political narratives and Silicon Valley's practical realities.

Top Performance, Low Cost: Business Trumps Politics

The logic here isn't complicated. K3 is currently recognized as one of the world's three strongest models. Meanwhile, models from OpenAI and Anthropic are closed-source, expensive, and come with plenty of restrictions. Open-source deployment frees companies from relying on closed-source APIs, letting them customize and optimize on their own, which significantly cuts costs.

Of course, deploying K3 isn't as simple as "download and run." Although this 2.8-trillion-parameter giant can run with as few as 8 AMD or Nvidia AI GPUs, Moonshot AI officially recommends a 64-GPU cluster setup. External estimates suggest the minimum cost for a smooth API service is around tens of millions of yuan, and achieving a commercial-grade experience requires an investment of about 30 million yuan.

In other words, each of these 18 US companies has carefully calculated the costs. Rather than endlessly paying token-based fees to closed-source vendors, they'd rather invest in infrastructure once and truly own the model's usage rights. In the second half of large model commercialization, controlling inference costs is becoming a more important strategic decision than signing exclusive partnerships.

Key Points

  • Kimi K3 was released open-source on Hugging Face and quickly deployed by at least 18 US companies.
  • The move highlights a disconnect between US political efforts to block Chinese AI and business demand for cost-effective, high-performance models.
  • Despite requiring significant upfront investment, open-source deployment offers long-term cost savings and greater control compared to closed-source APIs.