The release of Kimi by Moonshot—a free open-source model that caught up with proprietary engines from OpenAI and Anthropic in a heartbeat—has effectively neutralized Washington's export controls. While the US builds sanction fences, Chinese "free intelligence" is not just hopping over them; it is doing so with provocative grace. For American businesses, this isn't just another headline-grabbing achievement; it is a direct economic undermining. Why pay Silicon Valley for "premium" access when Beijing is giving away comparable power for free? According to Anton Likht of the Carnegie Endowment, these models have become an existential threat to an administration that is terrified of negative economic news.
A rift in the Trump orbit
Kimi’s arrival has sparked public infighting among Donald Trump’s inner circle. David Sacks, who oversaw AI and crypto issues until March, has launched an open attack on American market leaders. He brands Anthropic’s models as "lobotomized" and "woke," arguing that Chinese solutions are gaining popularity precisely because they lack ideological filters and censorship. In Sacks’ view, US tech giants are using the government to stifle open source under the guise of safety. This stance is opposed by a pragmatic faction at the Pentagon, which insists on total state control: since AI has become a weapon, its development must be locked inside a "digital fortress."
"Every time China releases a smart, free model like Kimi, American companies have fewer reasons to keep funneling money to Anthropic or OpenAI."
Internal friction reached a boiling point when the White House initiated a pre-release AI safety review process. Dean Ball, a former Trump advisor now at OpenAI, criticized the move, only to receive a harsh rebuke from Emil Michael. The high-ranking Pentagon official did not mince words, calling OpenAI’s head of strategic futures the "chief village idiot." While the US debates whether to save the market cap of its unicorns or turn them into state appendages, China is capturing the global market through open access.
Eroding margins and the TCO fog
For the corporate sector, the Kimi case turns tech stack selection into a minefield. Business models at OpenAI and Anthropic, built on subscription-based access to "brains," are starting to crack. The rising Total Cost of Ownership (TCO) for American solutions, contrasted with the availability of Chinese open source, is forcing CIOs to think twice before signing multi-million dollar contracts. A dangerous gap is emerging: US regulators are tightening the screws and increasing compliance costs, while Beijing subsidizes a global dumping of intelligence.
Ultimately, it looks as if the US is trying to play chess while its opponent simply hands out pieces to anyone who wants them. If American models eventually retreat into a closed circuit to serve government contracts, private business will inevitably migrate to the Chinese stack. As trust in OpenAI’s pricing fades and regulatory uncertainty grows, US technological sovereignty risks turning into isolation within a very expensive and very exclusive elite club.